add missing release notes to HISTORY.md

Summary:
Added unit tests that have paranoid_check = true that perform range deletions.  At the moment, the deleted ranges do not appear to be checked as part of the paranoid checks.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/7521

Reviewed By: zhichao-cao

Differential Revision: D24262175

Pulled By: ajkr

fbshipit-source-id: 1035e968f7ab8ccaa7af086b835a4e72c7e56743

.circleci/cat_ignore_eagain

@@ -0,0 +1,54 @@
+#! /bin/bash
+
+# Work around issue with parallel make output causing random error, as in
+# make[1]: write error: stdout
+# Probably due to a kernel bug:
+#   https://bugs.launchpad.net/ubuntu/+source/linux-signed/+bug/1814393
+# Seems to affect image ubuntu-1604:201903-01 and ubuntu-1604:202004-01
+
+cd "$(dirname $0)"
+
+if [ ! -x cat_ignore_eagain.out ]; then
+    cc -x c -o cat_ignore_eagain.out - << EOF
+#include <unistd.h>
+#include <errno.h>
+#include <stdio.h>
+int main() {
+    int n, m, p;
+    char buf[1024];
+    for (;;) {
+        n = read(STDIN_FILENO, buf, 1024);
+        if (n > 0 && n <= 1024) {
+            for (m = 0; m < n;) {
+                p = write(STDOUT_FILENO, buf + m, n - m);
+                if (p < 0) {
+                    if (errno == EAGAIN) {
+                        // ignore but pause a bit
+                        usleep(100);
+                    } else {
+                        perror("write failed");
+                        return 42;
+                    }
+                } else {
+                    m += p;
+                }
+            }
+        } else if (n < 0) {
+            if (errno == EAGAIN) {
+                // ignore but pause a bit
+                usleep(100);
+            } else {
+                // Some non-ignorable error
+                perror("read failed");
+                return 43;
+            }
+        } else {
+            // EOF
+            return 0;
+        }
+    }
+}
+EOF
+fi
+
+exec ./cat_ignore_eagain.out

.circleci/config.yml

@@ -2,95 +2,287 @@ version: 2.1
 
 orbs:
   win: circleci/windows@2.4.0
+  slack: circleci/slack@3.4.2
+
+aliases:
+  - &notify-on-master-failure
+    fail_only: true
+    only_for_branches: master
+
+commands:
+  install-pyenv-on-macos:
+    steps:
+      - run:
+          name: Install pyenv on macos
+          command: |
+            HOMEBREW_NO_AUTO_UPDATE=1 brew install pyenv
+
+  increase-max-open-files-on-macos:
+    steps:
+      - run:
+          name: Increase max open files
+          command: |
+            sudo sysctl -w kern.maxfiles=1048576
+            sudo sysctl -w kern.maxfilesperproc=1048576
+            sudo launchctl limit maxfiles 1048576
+
+  pre-steps:
+    steps:
+      - checkout
+      - run: pyenv install --skip-existing 3.5.9
+      - run: pyenv global 3.5.9
+      - run:
+          name: Setup Environment Variables
+          command: |
+            echo "export GTEST_THROW_ON_FAILURE=0" >> $BASH_ENV
+            echo "export GTEST_OUTPUT=\"xml:/tmp/test-results/\"" >> $BASH_ENV
+            echo "export SKIP_FORMAT_BUCK_CHECKS=1" >> $BASH_ENV
+            echo "export PRINT_PARALLEL_OUTPUTS=1" >> $BASH_ENV
+
+  post-steps:
+    steps:
+      - slack/status: *notify-on-master-failure
+      - store_test_results: # store test result if there's any
+          path: /tmp/test-results
+      - store_artifacts: # store LOG for debugging if there's any
+          path: LOG
+
+  install-clang-10:
+    steps:
+      - run:
+          name: Install Clang 10
+          command: |
+            echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
+            echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
+            echo "APT::Acquire::Retries \"10\";" | sudo tee -a /etc/apt/apt.conf.d/80-retries # llvm.org unreliable
+            sudo apt-get update -y && sudo apt-get install -y clang-10
+
+  install-gflags:
+    steps:
+      - run:
+          name: Install gflags
+          command: |
+            sudo apt-get update -y && sudo apt-get install -y libgflags-dev
+  
+  install-gflags-on-macos:
+    steps:
+      - run:
+          name: Install gflags on macos
+          command: |
+            HOMEBREW_NO_AUTO_UPDATE=1 brew install gflags
+
+  install-gtest-parallel:
+    steps:
+      - run:
+          name: Install gtest-parallel
+          command: |
+            git clone --single-branch --branch master --depth 1 https://github.com/google/gtest-parallel.git ~/gtest-parallel
+            echo "export PATH=$HOME/gtest-parallel:$PATH" >> $BASH_ENV
 
 executors:
   windows-2xlarge:
     machine:
-      image: 'windows-server-2019-vs2019:201908-06'
+      image: 'windows-server-2019-vs2019:stable'
       resource_class: windows.2xlarge
       shell: bash.exe
 
 jobs:
+  build-macos:
+    macos:
+      xcode: 11.3.0
+    steps:
+      - increase-max-open-files-on-macos
+      - install-pyenv-on-macos
+      - pre-steps
+      - install-gflags-on-macos
+      - run: ulimit -S -n 1048576 && OPT=-DCIRCLECI make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
+
   build-linux:
     machine:
-      image: ubuntu-1604:201903-01
+      image: ubuntu-1604:202007-01
     resource_class: 2xlarge
     steps:
-      - checkout # check out the code in the project directory
-      - run: pyenv global 3.5.2
-      - run: sudo apt-get update -y
-      - run: sudo apt-get install -y libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 PRINT_PARALLEL_OUTPUTS=1 make J=32 all check -j32
+      - pre-steps
+      - install-gflags
+      - run: make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-shared_lib-alt_namespace-status_checked:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - run: ASSERT_STATUS_CHECKED=1 TEST_UINT128_COMPAT=1 LIB_MODE=shared OPT="-DROCKSDB_NAMESPACE=alternative_rocksdb_ns" make V=1 -j32 all check_some | .circleci/cat_ignore_eagain
+      - post-steps
 
   build-linux-release:
     machine:
-      image: ubuntu-1604:201903-01
+      image: ubuntu-1604:202007-01
     resource_class: 2xlarge
     steps:
       - checkout # check out the code in the project directory
-      - run: make release -j32
+      - run: make V=1 -j32 release | .circleci/cat_ignore_eagain
+      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags
+      - install-gflags
+      - run: make V=1 -j32 release | .circleci/cat_ignore_eagain
+      - run: ./db_stress --version # ensure with gflags
+      - post-steps
 
-  build-linux-lite:
-    machine:
-      image: ubuntu-1604:201903-01 
-    resource_class: 2xlarge
-    steps:
-      - checkout # check out the code in the project directory
-      - run: pyenv global 3.5.2
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 PRINT_PARALLEL_OUTPUTS=1 LITE=1 make J=32 all check -j32
-
-  build-linux-lite-release:
+  build-linux-release-rtti:
     machine:
       image: ubuntu-1604:201903-01
     resource_class: large
     steps:
       - checkout # check out the code in the project directory
-      - run: make release -j32
+      - run: make clean
+      - run: USE_RTTI=1 DEBUG_LEVEL=0 make V=1 -j16 static_lib tools db_bench | .circleci/cat_ignore_eagain
+      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags
+      - run: sudo apt-get update -y && sudo apt-get install -y libgflags-dev
+      - run: make clean
+      - run: USE_RTTI=1 DEBUG_LEVEL=0 make V=1 -j16 static_lib tools db_bench | .circleci/cat_ignore_eagain
+      - run: ./db_stress --version # ensure with gflags
+
+  build-linux-lite:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - run: LITE=1 make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-lite-release:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: large
+    steps:
+      - checkout # check out the code in the project directory
+      - run: LITE=1 make V=1 -j32 release | .circleci/cat_ignore_eagain
+      - run: if ./db_stress --version; then false; else true; fi # ensure without gflags
+      - install-gflags
+      - run: LITE=1 make V=1 -j32 release | .circleci/cat_ignore_eagain
+      - run: ./db_stress --version # ensure with gflags
+      - post-steps
 
   build-linux-clang-no-test:
     machine:
-      image: ubuntu-1604:201903-01
+      image: ubuntu-1604:202007-01
     resource_class: 2xlarge
     steps:
       - checkout # check out the code in the project directory
-      - run: sudo apt-get update -y
-      - run: sudo apt-get install -y clang
-      - run: CC=clang CXX=clang++ USE_CLANG=1 PORTABLE=1 make all -j32
+      - run: sudo apt-get update -y && sudo apt-get install -y clang libgflags-dev
+      - run: CC=clang CXX=clang++ USE_CLANG=1 PORTABLE=1 make V=1 -j32 all | .circleci/cat_ignore_eagain
+      - post-steps
 
-  build-linux-clang10-no-test:
+  build-linux-clang10-asan:
     machine:
-      image: ubuntu-1604:201903-01
+      image: ubuntu-1604:202007-01
     resource_class: 2xlarge
     steps:
-      - checkout # check out the code in the project directory
-      - run: sudo sh -c 'echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" >> /etc/apt/sources.list'
-      - run: sudo sh -c 'echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" >> /etc/apt/sources.list'
-      - run: sudo apt-get update -y
-      - run: sudo apt-get install -y clang-10
-      - run: CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 make all -j32 # aligned new doesn't work for reason we haven't figured out
+      - pre-steps
+      - install-gflags
+      - install-clang-10
+      - run: COMPILE_WITH_ASAN=1 CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 make V=1 -j32 check | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out
+      - post-steps
+
+  build-linux-clang10-mini-tsan:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - install-clang-10
+      - run: COMPILE_WITH_TSAN=1 CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 make V=1 -j32 check | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out.
+      - post-steps
+
+  build-linux-clang10-ubsan:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - install-clang-10
+      - run: COMPILE_WITH_UBSAN=1 OPT="-fsanitize-blacklist=.circleci/ubsan_suppression_list.txt" CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 make V=1 -j32 ubsan_check | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out
+      - post-steps
+
+  build-linux-clang10-clang-analyze:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - install-clang-10
+      - run: sudo apt-get update -y && sudo apt-get install -y clang-tools-10
+      - run: CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 CLANG_ANALYZER="/usr/bin/clang++-10" CLANG_SCAN_BUILD=scan-build-10 USE_CLANG=1 make V=1 -j32 analyze | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out. For unknown, reason passing "clang++-10" as CLANG_ANALYZER doesn't work, and we need a full path.
+      - post-steps
 
   build-linux-cmake:
     machine:
-      image: ubuntu-1604:201903-01
+      image: ubuntu-1604:202007-01
     resource_class: 2xlarge
     steps:
       - checkout # check out the code in the project directory
-      - run: mkdir build && cd build && cmake -DWITH_GFLAGS=0 .. && make -j32
+      - run: (mkdir build && cd build && cmake -DWITH_GFLAGS=0 .. && make V=1 -j32) | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-unity:
+    docker: # executor type
+      - image: gcc:latest
+    resource_class: xlarge
+    steps:
+      - checkout # check out the code in the project directory
+      - run: apt-get update -y && apt-get install -y libgflags-dev
+      - run: TEST_TMPDIR=/dev/shm && make V=1 -j16 unity_test | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-gcc-4-8:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: large
+    steps:
+      - pre-steps
+      - run: sudo apt-get update -y && sudo apt-get install gcc-4.8 g++-4.8 libgflags-dev
+      - run: CC=gcc-4.8 CXX=g++-4.8 V=1 SKIP_LINK=1 make -j4 all | .circleci/cat_ignore_eagain # Linking broken because libgflags compiled with newer ABI
+      - post-steps
 
   build-windows:
     executor: windows-2xlarge
-    
+    parameters:
+      extra_cmake_opt:
+        default: ""
+        type: string
+      vs_year:
+        default: "2019"
+        type: string
+      cmake_generator:
+        default: "Visual Studio 16 2019"
+        type: string
     environment:
       THIRDPARTY_HOME: C:/Users/circleci/thirdparty
       CMAKE_HOME: C:/Users/circleci/thirdparty/cmake-3.16.4-win64-x64
       CMAKE_BIN: C:/Users/circleci/thirdparty/cmake-3.16.4-win64-x64/bin/cmake.exe
-      CMAKE_GENERATOR: Visual Studio 16 2019
       SNAPPY_HOME: C:/Users/circleci/thirdparty/snappy-1.1.7
       SNAPPY_INCLUDE: C:/Users/circleci/thirdparty/snappy-1.1.7;C:/Users/circleci/thirdparty/snappy-1.1.7/build
       SNAPPY_LIB_DEBUG: C:/Users/circleci/thirdparty/snappy-1.1.7/build/Debug/snappy.lib
-
+      VS_YEAR: <<parameters.vs_year>>
+      CMAKE_GENERATOR: <<parameters.cmake_generator>>
     steps:
       - checkout
+      - run:
+          name: "Setup VS"
+          command: |
+            if [[ "${VS_YEAR}" == "2017" ]]; then
+              powershell .circleci/vs2017_install.ps1
+            elif [[ "${VS_YEAR}" == "2015" ]]; then
+              powershell .circleci/vs2015_install.ps1
+            fi
       - run:
           name: "Install thirdparty dependencies"
           command: |
@@ -112,37 +304,156 @@ jobs:
           command: |
             mkdir build
             cd build
-            ${CMAKE_BIN} -G "${CMAKE_GENERATOR}" -DCMAKE_BUILD_TYPE=Debug -DOPTDBG=1 -DPORTABLE=1 -DSNAPPY=1 -DJNI=1 ..
+            ${CMAKE_BIN} -G "${CMAKE_GENERATOR}" -DCMAKE_BUILD_TYPE=Debug -DOPTDBG=1 -DPORTABLE=1 -DSNAPPY=1 -DJNI=1 << parameters.extra_cmake_opt >> ..
             cd ..
             msbuild.exe build/rocksdb.sln -maxCpuCount -property:Configuration=Debug -property:Platform=x64
       - run:
           name: "Test RocksDB"
           shell: powershell.exe
           command: |
-            build_tools\run_ci_db_test.ps1 -SuiteRun db_basic_test,db_test,db_test2,env_basic_test,env_test,db_merge_operand_test -Concurrency 16
+            build_tools\run_ci_db_test.ps1 -SuiteRun db_basic_test,db_test,db_test2,db_merge_operand_test,bloom_test,c_test,coding_test,crc32c_test,dynamic_bloom_test,env_basic_test,env_test,hash_test,random_test -Concurrency 16
+
+  build-linux-java:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - run:
+          name: "Build RocksDBJava"
+          command: |
+            export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-amd64
+            export PATH=$JAVA_HOME/bin:$PATH
+            echo "JAVA_HOME=${JAVA_HOME}"
+            which java && java -version
+            which javac && javac -version
+            make V=1 J=32 -j32 rocksdbjava jtest | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-examples:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: medium
+    steps:
+      - pre-steps
+      - install-gflags
+      - run:
+          name: "Build examples"
+          command: |
+            OPT=-DTRAVIS V=1 make -j4 static_lib && cd examples && make -j4 | ../.circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-non-shm:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    parameters:
+      start_test:
+        default: ""
+        type: string
+      end_test:
+        default: ""
+        type: string
+    steps:
+      - pre-steps
+      - install-gflags
+      - install-gtest-parallel
+      - run:
+          name: "Build unit tests"
+          command: |
+            echo "env: $(env)"
+            echo "** done env"
+            ROCKSDBTESTS_START=<<parameters.start_test>> ROCKSDBTESTS_END=<<parameters.end_test>> ROCKSDBTESTS_SUBSET_TESTS_TO_FILE=/tmp/test_list make V=1 -j32 --output-sync=target build_subset_tests
+      - run:
+          name: "Run unit tests in parallel"
+          command: |
+            sed -i 's/[[:space:]]*$//; s/ / \.\//g; s/.*/.\/&/' /tmp/test_list
+            cat /tmp/test_list
+            export TEST_TMPDIR=/tmp/rocksdb_test_tmp
+            /usr/bin/python ../gtest-parallel/gtest-parallel $(</tmp/test_list) --output_dir=/tmp | cat  # pipe to cat to continuously output status on circleci UI. Otherwise, no status will be printed while the job is running.
+      - post-steps
 
 workflows:
   build-linux:
     jobs:
       - build-linux
+  build-linux-shared_lib-alt_namespace-status_checked:
+    jobs:
+      - build-linux-shared_lib-alt_namespace-status_checked
   build-linux-lite:
     jobs:
       - build-linux-lite
   build-linux-release:
     jobs:
       - build-linux-release
+  build-linux-release-rtti:
+    jobs:
+      - build-linux-release-rtti
   build-linux-lite-release:
     jobs:
       - build-linux-lite-release
   build-linux-clang-no-test:
     jobs:
       - build-linux-clang-no-test
-  build-linux-clang10-no-test:
+  build-linux-clang10-asan:
     jobs:
-      - build-linux-clang10-no-test
+      - build-linux-clang10-asan
+  build-linux-clang10-mini-tsan:
+    jobs:
+      - build-linux-clang10-mini-tsan
+  build-linux-clang10-ubsan:
+    jobs:
+      - build-linux-clang10-ubsan
+  build-linux-clang10-clang-analyze:
+    jobs:
+      - build-linux-clang10-clang-analyze
   build-linux-cmake:
     jobs:
       - build-linux-cmake
+  build-linux-unity:
+    jobs:
+      - build-linux-unity
   build-windows:
     jobs:
       - build-windows
+  build-windows-vs2017:
+    jobs:
+      - build-windows:
+          vs_year: "2017"
+          cmake_generator: "Visual Studio 15 Win64"
+  build-windows-vs2015:
+    jobs:
+      - build-windows:
+          vs_year: "2015"
+          cmake_generator: "Visual Studio 14 Win64"
+  build-windows-cxx20:
+    jobs:
+      - build-windows:
+          extra_cmake_opt: -DCMAKE_CXX_STANDARD=20
+  build-java:
+    jobs:
+      - build-linux-java
+  build-examples:
+    jobs:
+      - build-examples
+  build-linux-non-shm:
+    jobs:
+      - build-linux-non-shm:
+          start_test: ""
+          end_test: "db_options_test" # make sure unique in src.mk
+      - build-linux-non-shm:
+          start_test: "db_options_test" # make sure unique in src.mk
+          end_test: "filename_test" # make sure unique in src.mk
+      - build-linux-non-shm:
+          start_test: "filename_test" # make sure unique in src.mk
+          end_test: "statistics_test" # make sure unique in src.mk
+      - build-linux-non-shm:
+          start_test: "statistics_test" # make sure unique in src.mk
+          end_test: ""
+  build-linux-gcc-4-8:
+    jobs:
+      - build-linux-gcc-4-8
+  build-macos:
+    jobs:
+      - build-macos

.circleci/ubsan_suppression_list.txt

@@ -0,0 +1,6 @@
+# Supress UBSAN warnings related to stl_tree.h, e.g.
+# UndefinedBehaviorSanitizer: undefined-behavior /usr/bin/../lib/gcc/x86_64-linux-gnu/5.4.0/../../../../include/c++/5.4.0/bits/stl_tree.h:1505:43 in 
+# /usr/bin/../lib/gcc/x86_64-linux-gnu/5.4.0/../../../../include/c++/5.4.0/bits/stl_tree.h:1505:43:
+# runtime error: upcast of address 0x000001fa8820 with insufficient space for an object of type
+# 'std::_Rb_tree_node<std::pair<const std::__cxx11::basic_string<char>, rocksdb::(anonymous namespace)::LockHoldingInfo> >'
+src:*bits/stl_tree.h

.circleci/vs2015_install.ps1

@@ -0,0 +1,23 @@
+$VS_DOWNLOAD_LINK = "https://go.microsoft.com/fwlink/?LinkId=691126"
+$COLLECT_DOWNLOAD_LINK = "https://aka.ms/vscollect.exe"
+curl.exe --retry 3 -kL $VS_DOWNLOAD_LINK --output vs_installer.exe
+if ($LASTEXITCODE -ne 0) {
+    echo "Download of the VS 2015 installer failed"
+    exit 1
+}
+$VS_INSTALL_ARGS = @("/Quiet", "/NoRestart")
+$process = Start-Process "${PWD}\vs_installer.exe" -ArgumentList $VS_INSTALL_ARGS -NoNewWindow -Wait -PassThru
+Remove-Item -Path vs_installer.exe -Force
+$exitCode = $process.ExitCode
+if (($exitCode -ne 0) -and ($exitCode -ne 3010)) {
+    echo "VS 2017 installer exited with code $exitCode, which should be one of [0, 3010]."
+    curl.exe --retry 3 -kL $COLLECT_DOWNLOAD_LINK --output Collect.exe
+    if ($LASTEXITCODE -ne 0) {
+        echo "Download of the VS Collect tool failed."
+        exit 1
+    }
+    Start-Process "${PWD}\Collect.exe" -NoNewWindow -Wait -PassThru
+    New-Item -Path "C:\w\build-results" -ItemType "directory" -Force
+    Copy-Item -Path "C:\Users\circleci\AppData\Local\Temp\vslogs.zip" -Destination "C:\w\build-results\"
+    exit 1
+}

.circleci/vs2017_install.ps1

@@ -0,0 +1,34 @@
+$VS_DOWNLOAD_LINK = "https://aka.ms/vs/15/release/vs_buildtools.exe"
+$COLLECT_DOWNLOAD_LINK = "https://aka.ms/vscollect.exe"
+$VS_INSTALL_ARGS = @("--nocache","--quiet","--wait", "--add Microsoft.VisualStudio.Workload.VCTools",
+                                                     "--add Microsoft.VisualStudio.Component.VC.Tools.14.13",
+                                                     "--add Microsoft.Component.MSBuild",
+                                                     "--add Microsoft.VisualStudio.Component.Roslyn.Compiler",
+                                                     "--add Microsoft.VisualStudio.Component.TextTemplating",
+                                                     "--add Microsoft.VisualStudio.Component.VC.CoreIde",
+                                                     "--add Microsoft.VisualStudio.Component.VC.Redist.14.Latest",
+                                                     "--add Microsoft.VisualStudio.ComponentGroup.NativeDesktop.Core",
+                                                     "--add Microsoft.VisualStudio.Component.VC.Tools.x86.x64",
+                                                     "--add Microsoft.VisualStudio.ComponentGroup.NativeDesktop.Win81")
+
+curl.exe --retry 3 -kL $VS_DOWNLOAD_LINK --output vs_installer.exe
+if ($LASTEXITCODE -ne 0) {
+    echo "Download of the VS 2017 installer failed"
+    exit 1
+}
+
+$process = Start-Process "${PWD}\vs_installer.exe" -ArgumentList $VS_INSTALL_ARGS -NoNewWindow -Wait -PassThru
+Remove-Item -Path vs_installer.exe -Force
+$exitCode = $process.ExitCode
+if (($exitCode -ne 0) -and ($exitCode -ne 3010)) {
+    echo "VS 2017 installer exited with code $exitCode, which should be one of [0, 3010]."
+    curl.exe --retry 3 -kL $COLLECT_DOWNLOAD_LINK --output Collect.exe
+    if ($LASTEXITCODE -ne 0) {
+        echo "Download of the VS Collect tool failed."
+        exit 1
+    }
+    Start-Process "${PWD}\Collect.exe" -NoNewWindow -Wait -PassThru
+    New-Item -Path "C:\w\build-results" -ItemType "directory" -Force
+    Copy-Item -Path "C:\Users\circleci\AppData\Local\Temp\vslogs.zip" -Destination "C:\w\build-results\"
+    exit 1
+}

.gitignore

@@ -1,4 +1,5 @@
 make_config.mk
+rocksdb.pc
 
 *.a
 *.arc
@@ -52,6 +53,7 @@ db_test2
 trace_analyzer
 trace_analyzer_test
 block_cache_trace_analyzer
+io_tracer_parser
 .DS_Store
 .vs
 .vscode

.travis.yml

@@ -87,134 +87,134 @@ matrix:
   - os: linux
     compiler: clang
   # Exclude all but most unique cmake variants for pull requests, but build all in branches
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: amd64
     env: JOB_NAME=cmake
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: amd64
     env: JOB_NAME=cmake-gcc8
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: amd64
     env: JOB_NAME=cmake-gcc9
   # Exclude most osx, arm64 and ppc64le tests for pull requests, but build in branches
-  # Temporarily disable ppc64le unit tests in PRs until Travis gets its act together (#6653)
-  - if: type = pull_request
+  # Temporarily disable ppc64le cmake test while snapd is broken
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
-    env: TEST_GROUP=platform_dependent
+    env: JOB_NAME=cmake
   # NB: the cmake build is a partial java test
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: osx
     env: TEST_GROUP=1
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: TEST_GROUP=1
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: TEST_GROUP=1
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: osx
     env: TEST_GROUP=2
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: TEST_GROUP=2
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: TEST_GROUP=2
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: osx
     env: TEST_GROUP=3
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: TEST_GROUP=3
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: TEST_GROUP=3
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: osx
     env: TEST_GROUP=4
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: TEST_GROUP=4
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: TEST_GROUP=4
-  - if: type = pull_request AND commit_message !~ /java/
+  - if: type = pull_request AND commit_message !~ /FULL_CI/ AND commit_message !~ /java/
     os : osx
     env: JOB_NAME=java_test
-  - if: type = pull_request AND commit_message !~ /java/
+  - if: type = pull_request AND commit_message !~ /FULL_CI/ AND commit_message !~ /java/
     os : linux
     arch: arm64
     env: JOB_NAME=java_test
-  - if: type = pull_request AND commit_message !~ /java/
+  - if: type = pull_request AND commit_message !~ /FULL_CI/ AND commit_message !~ /java/
     os: linux
     arch: ppc64le
     env: JOB_NAME=java_test
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : osx
     env: JOB_NAME=lite_build
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: JOB_NAME=lite_build
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: JOB_NAME=lite_build
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : osx
     env: JOB_NAME=examples
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: JOB_NAME=examples
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: JOB_NAME=examples
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: JOB_NAME=cmake-gcc8
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: JOB_NAME=cmake-gcc8
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: JOB_NAME=cmake-gcc9
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: JOB_NAME=cmake-gcc9
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: JOB_NAME=cmake-gcc9-c++20
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: JOB_NAME=cmake-gcc9-c++20
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : osx
     env: JOB_NAME=status_checked
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os : linux
     arch: arm64
     env: JOB_NAME=status_checked
-  - if: type = pull_request
+  - if: type = pull_request AND commit_message !~ /FULL_CI/
     os: linux
     arch: ppc64le
     env: JOB_NAME=status_checked
@@ -224,38 +224,33 @@ install:
       PATH=$PATH:/usr/local/opt/ccache/libexec;
     fi
   - if [ "${JOB_NAME}" == cmake-gcc8 ]; then
-      sudo apt-get install -y g++-8;
+      sudo apt-get install -y g++-8 || exit $?;
       CC=gcc-8 && CXX=g++-8;
     fi
   - if [ "${JOB_NAME}" == cmake-gcc9 ] || [ "${JOB_NAME}" == cmake-gcc9-c++20 ]; then
-      sudo apt-get install -y g++-9;
+      sudo apt-get install -y g++-9 || exit $?;
       CC=gcc-9 && CXX=g++-9;
     fi
   - if [ "${JOB_NAME}" == cmake-mingw ]; then
-      sudo apt-get install -y mingw-w64 ;
+      sudo apt-get install -y mingw-w64 || exit $?;
     fi
   - if [ "${JOB_NAME}" == make-gcc4.8 ]; then
-      sudo apt-get install -y g++-4.8 ;
+      sudo apt-get install -y g++-4.8 || exit $?;
       CC=gcc-4.8 && CXX=g++-4.8;
     fi
   - if [[ "${JOB_NAME}" == cmake* ]] && [ "${TRAVIS_OS_NAME}" == linux ]; then
-      CMAKE_DIST_URL="https://rocksdb-deps.s3-us-west-2.amazonaws.com/cmake/cmake-3.14.5-Linux-$(uname -m).tar.bz2";
-      TAR_OPT="--strip-components=1 -xj";
-      if [ "aarch64" == "$(uname -m)" ]; then
-        sudo apt-get install -y libuv1 librhash0;
-        sudo apt-get upgrade -y libstdc++6;
-      fi;
-      mkdir cmake-dist && curl --silent --fail --show-error --location "${CMAKE_DIST_URL}" | tar -C cmake-dist ${TAR_OPT} && export PATH=$PWD/cmake-dist/bin:$PATH;
+      sudo apt-get install snapd && sudo snap install cmake --beta --classic || exit $?;
+      export PATH=/snap/bin:$PATH;
     fi
   - |
     if [[ "${JOB_NAME}" == java_test || "${JOB_NAME}" == cmake* ]]; then
       # Ensure JDK 8
       if [ "${TRAVIS_OS_NAME}" == osx ]; then
-        brew tap AdoptOpenJDK/openjdk
-        brew cask install adoptopenjdk8
+        brew tap AdoptOpenJDK/openjdk || exit $?
+        brew cask install adoptopenjdk8 || exit $?
         export JAVA_HOME=$(/usr/libexec/java_home)
       else
-        sudo apt-get install -y openjdk-8-jdk
+        sudo apt-get install -y openjdk-8-jdk || exit $?
         export PATH=/usr/lib/jvm/java-8-openjdk-$(dpkg --print-architecture)/bin:$PATH
         export JAVA_HOME=/usr/lib/jvm/java-8-openjdk-$(dpkg --print-architecture)
       fi
@@ -274,19 +269,19 @@ script:
   - if [ `command -v ccache` ]; then ccache -C; fi
   - case $TEST_GROUP in
     platform_dependent)
-      OPT=-DTRAVIS V=1 ROCKSDBTESTS_END=db_block_cache_test make -j4 all_but_some_tests check_some
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_PLATFORM_DEPENDENT=only make -j4 all_but_some_tests check_some
       ;;
     1)
-      OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=db_block_cache_test ROCKSDBTESTS_END=db_iter_test make -j4 check_some
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_PLATFORM_DEPENDENT=exclude ROCKSDBTESTS_END=backupable_db_test make -j4 check_some
       ;;
     2)
-      OPT="-DTRAVIS -DROCKSDB_NAMESPACE=alternative_rocksdb_ns" V=1 make -j4 tools && OPT="-DTRAVIS -DROCKSDB_NAMESPACE=alternative_rocksdb_ns" V=1 ROCKSDBTESTS_START=db_iter_test ROCKSDBTESTS_END=options_file_test make -j4 check_some
+      OPT="-DTRAVIS -DROCKSDB_NAMESPACE=alternative_rocksdb_ns" V=1 make -j4 tools && OPT="-DTRAVIS -DROCKSDB_NAMESPACE=alternative_rocksdb_ns" V=1 ROCKSDBTESTS_PLATFORM_DEPENDENT=exclude ROCKSDBTESTS_START=backupable_db_test ROCKSDBTESTS_END=db_universal_compaction_test make -j4 check_some
       ;;
     3)
-      OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=options_file_test ROCKSDBTESTS_END=write_prepared_transaction_test make -j4 check_some
+      OPT=-DTRAVIS LIB_MODE=shared V=1 ROCKSDBTESTS_PLATFORM_DEPENDENT=exclude ROCKSDBTESTS_START=db_universal_compaction_test ROCKSDBTESTS_END=table_properties_collector_test make -j4 check_some
       ;;
     4)
-      OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=write_prepared_transaction_test make -j4 check_some
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_PLATFORM_DEPENDENT=exclude ROCKSDBTESTS_START=table_properties_collector_test make -j4 check_some
       ;;
     esac
   - case $JOB_NAME in

CMakeLists.txt

@@ -107,6 +107,7 @@ else()
     endif()
   endif()
 
+  set(GFLAGS_LIB)
   if(WITH_GFLAGS)
     # Config with namespace available since gflags 2.2.2
     option(GFLAGS_USE_TARGET_NAMESPACE "Use gflags import target with namespace." ON)
@@ -114,15 +115,17 @@ else()
     if(gflags_FOUND)
       if(TARGET ${GFLAGS_TARGET})
         # Config with GFLAGS_TARGET available since gflags 2.2.0
-        list(APPEND THIRDPARTY_LIBS ${GFLAGS_TARGET})
+        set(GFLAGS_LIB ${GFLAGS_TARGET})
       else()
         # Config with GFLAGS_LIBRARIES available since gflags 2.1.0
-        list(APPEND THIRDPARTY_LIBS ${GFLAGS_LIBRARIES})
+        set(GFLAGS_LIB ${GFLAGS_LIBRARIES})
       endif()
     else()
       find_package(gflags REQUIRED)
-      list(APPEND THIRDPARTY_LIBS gflags::gflags)
+        set(GFLAGS_LIB gflags::gflags)
     endif()
+    include_directories(${GFLAGS_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${GFLAGS_LIB})
     add_definitions(-DGFLAGS=1)
   endif()
 
@@ -251,12 +254,30 @@ endif(CMAKE_SYSTEM_PROCESSOR MATCHES "aarch64|AARCH64")
 
 option(PORTABLE "build a portable binary" OFF)
 option(FORCE_SSE42 "force building with SSE4.2, even when PORTABLE=ON" OFF)
+option(FORCE_AVX "force building with AVX, even when PORTABLE=ON" OFF)
+option(FORCE_AVX2 "force building with AVX2, even when PORTABLE=ON" OFF)
 if(PORTABLE)
   # MSVC does not need a separate compiler flag to enable SSE4.2; if nmmintrin.h
   # is available, it is available by default.
   if(FORCE_SSE42 AND NOT MSVC)
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse4.2 -mpclmul")
   endif()
+  if(MSVC)
+    if(FORCE_AVX)
+      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX")
+    endif()
+    # MSVC automatically enables BMI / lzcnt with AVX2.
+    if(FORCE_AVX2)
+      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX2")
+    endif()
+  else()
+    if(FORCE_AVX)
+      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mavx")
+    endif()
+    if(FORCE_AVX2)
+      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mavx2 -mbmi -mlzcnt")
+    endif()
+  endif()
 else()
   if(MSVC)
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX2")
@@ -370,6 +391,12 @@ if(NOT WITH_DYNAMIC_EXTENSION)
   add_definitions(-DROCKSDB_NO_DYNAMIC_EXTENSION)
 endif()
 
+option(ASSERT_STATUS_CHECKED "build with assert status checked" OFF)
+if (ASSERT_STATUS_CHECKED)
+  message(STATUS "Build with assert status checked")
+  add_definitions(-DROCKSDB_ASSERT_STATUS_CHECKED)
+endif()
+
 if(DEFINED USE_RTTI)
   if(USE_RTTI)
     message(STATUS "Enabling RTTI")
@@ -508,7 +535,11 @@ if(HAVE_PTHREAD_MUTEX_ADAPTIVE_NP)
 endif()
 
 include(CheckCXXSymbolExists)
-check_cxx_symbol_exists(malloc_usable_size malloc.h HAVE_MALLOC_USABLE_SIZE)
+if(CMAKE_SYSTEM_NAME MATCHES "^FreeBSD")
+  check_cxx_symbol_exists(malloc_usable_size malloc_np.h HAVE_MALLOC_USABLE_SIZE)
+else()
+  check_cxx_symbol_exists(malloc_usable_size malloc.h HAVE_MALLOC_USABLE_SIZE)
+endif()
 if(HAVE_MALLOC_USABLE_SIZE)
   add_definitions(-DROCKSDB_MALLOC_USABLE_SIZE)
 endif()
@@ -539,10 +570,12 @@ set(SOURCES
         cache/sharded_cache.cc
         db/arena_wrapped_db_iter.cc
         db/blob/blob_file_addition.cc
+        db/blob/blob_file_builder.cc
         db/blob/blob_file_garbage.cc
         db/blob/blob_file_meta.cc
+        db/blob/blob_file_reader.cc
         db/blob/blob_log_format.cc
-        db/blob/blob_log_reader.cc
+        db/blob/blob_log_sequential_reader.cc
         db/blob/blob_log_writer.cc
         db/builder.cc
         db/c.cc
@@ -555,6 +588,7 @@ set(SOURCES
         db/compaction/compaction_picker_fifo.cc
         db/compaction/compaction_picker_level.cc
         db/compaction/compaction_picker_universal.cc
+        db/compaction/sst_partitioner.cc
         db/convenience.cc
         db/db_filesnapshot.cc
         db/db_impl/db_impl.cc
@@ -587,6 +621,8 @@ set(SOURCES
         db/memtable_list.cc
         db/merge_helper.cc
         db/merge_operator.cc
+        db/output_validator.cc
+        db/periodic_work_scheduler.cc
         db/range_del_aggregator.cc
         db/range_tombstone_fragmenter.cc
         db/repair.cc
@@ -599,6 +635,7 @@ set(SOURCES
         db/version_edit.cc
         db/version_edit_handler.cc
         db/version_set.cc
+        db/wal_edit.cc
         db/wal_manager.cc
         db/write_batch.cc
         db/write_batch_base.cc
@@ -609,6 +646,7 @@ set(SOURCES
         env/env_encryption.cc
         env/env_hdfs.cc
         env/file_system.cc
+        env/file_system_tracer.cc
         env/mock_env.cc
         file/delete_scheduler.cc
         file/file_prefetch_buffer.cc
@@ -647,6 +685,7 @@ set(SOURCES
         monitoring/thread_status_util.cc
         monitoring/thread_status_util_debug.cc
         options/cf_options.cc
+        options/configurable.cc
         options/db_options.cc
         options/options.cc
         options/options_helper.cc
@@ -694,8 +733,10 @@ set(SOURCES
         table/plain/plain_table_index.cc
         table/plain/plain_table_key_coding.cc
         table/plain/plain_table_reader.cc
+        table/sst_file_dumper.cc
         table/sst_file_reader.cc
         table/sst_file_writer.cc
+        table/table_factory.cc
         table/table_properties.cc
         table/two_level_iterator.cc
         test_util/sync_point.cc
@@ -704,12 +745,14 @@ set(SOURCES
         test_util/transaction_test_util.cc
         tools/block_cache_analyzer/block_cache_trace_analyzer.cc
         tools/dump/db_dump_tool.cc
+        tools/io_tracer_parser_tool.cc
         tools/ldb_cmd.cc
         tools/ldb_tool.cc
         tools/sst_dump_tool.cc
         tools/trace_analyzer_tool.cc
         trace_replay/trace_replay.cc
         trace_replay/block_cache_tracer.cc
+        trace_replay/io_tracer.cc
         util/coding.cc
         util/compaction_job_stats_impl.cc
         util/comparator.cc
@@ -743,6 +786,8 @@ set(SOURCES
         utilities/debug.cc
         utilities/env_mirror.cc
         utilities/env_timed.cc
+        utilities/fault_injection_env.cc
+        utilities/fault_injection_fs.cc
         utilities/leveldb_options/leveldb_options.cc
         utilities/memory/memory_util.cc
         utilities/merge_operators/bytesxor.cc
@@ -764,6 +809,8 @@ set(SOURCES
         utilities/simulator_cache/sim_cache.cc
         utilities/table_properties_collectors/compact_on_deletion_collector.cc
         utilities/trace/file_trace_reader_writer.cc
+        utilities/transactions/lock/lock_tracker.cc
+        utilities/transactions/lock/point_lock_tracker.cc
         utilities/transactions/optimistic_transaction_db_impl.cc
         utilities/transactions/optimistic_transaction.cc
         utilities/transactions/pessimistic_transaction.cc
@@ -978,11 +1025,15 @@ endif()
 
 option(WITH_ALL_TESTS "Build all test, rather than a small subset" ON)
 
-if(WITH_TESTS)
+if(WITH_TESTS OR WITH_BENCHMARK_TOOLS)
   add_subdirectory(third-party/gtest-1.8.1/fused-src/gtest)
   add_library(testharness STATIC
+  test_util/mock_time_env.cc
   test_util/testharness.cc)
   target_link_libraries(testharness gtest)
+endif()
+
+if(WITH_TESTS)
   set(TESTS
         db/db_basic_test.cc
         env/env_basic_test.cc
@@ -992,7 +1043,9 @@ if(WITH_TESTS)
         cache/cache_test.cc
         cache/lru_cache_test.cc
         db/blob/blob_file_addition_test.cc
+        db/blob/blob_file_builder_test.cc
         db/blob/blob_file_garbage_test.cc
+        db/blob/blob_file_reader_test.cc
         db/blob/db_blob_index_test.cc
         db/column_family_test.cc
         db/compact_files_test.cc
@@ -1052,6 +1105,7 @@ if(WITH_TESTS)
         db/merge_test.cc
         db/options_file_test.cc
         db/perf_context_test.cc
+        db/periodic_work_scheduler_test.cc
         db/plain_table_db_test.cc
         db/prefix_test.cc
         db/range_del_aggregator_test.cc
@@ -1062,6 +1116,7 @@ if(WITH_TESTS)
         db/version_edit_test.cc
         db/version_set_test.cc
         db/wal_manager_test.cc
+        db/wal_edit_test.cc
         db/write_batch_test.cc
         db/write_callback_test.cc
         db/write_controller_test.cc
@@ -1069,6 +1124,7 @@ if(WITH_TESTS)
         env/io_posix_test.cc
         env/mock_env_test.cc
         file/delete_scheduler_test.cc
+        file/prefetch_test.cc
         file/random_access_file_reader_test.cc
         logging/auto_roll_logger_test.cc
         logging/env_logger_test.cc
@@ -1082,6 +1138,7 @@ if(WITH_TESTS)
         monitoring/iostats_context_test.cc
         monitoring/statistics_test.cc
         monitoring/stats_history_test.cc
+        options/configurable_test.cc
         options/options_settable_test.cc
         options/options_test.cc
         table/block_based/block_based_filter_block_test.cc
@@ -1098,7 +1155,9 @@ if(WITH_TESTS)
         table/table_test.cc
         table/block_fetcher_test.cc
         test_util/testutil_test.cc
+        trace_replay/io_tracer_test.cc
         tools/block_cache_analyzer/block_cache_trace_analyzer_test.cc
+        tools/io_tracer_parser_test.cc
         tools/ldb_cmd_test.cc
         tools/reduce_levels_test.cc
         tools/sst_dump_test.cc
@@ -1161,8 +1220,6 @@ if(WITH_TESTS)
       db/db_test_util.cc
       monitoring/thread_status_updater_debug.cc
       table/mock_table.cc
-      test_util/fault_injection_test_env.cc
-      test_util/fault_injection_test_fs.cc
       utilities/cassandra/test_utils.cc
   )
   enable_testing()
@@ -1188,7 +1245,7 @@ if(WITH_TESTS)
         EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
         OUTPUT_NAME ${exename}${ARTIFACT_SUFFIX}
       )
-      target_link_libraries(${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX} testharness gtest ${ROCKSDB_LIB})
+      target_link_libraries(${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX} testharness gtest ${THIRDPARTY_LIBS} ${ROCKSDB_LIB})
       if(NOT "${exename}" MATCHES "db_sanity_test")
         add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
         add_dependencies(check ${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX})
@@ -1212,9 +1269,8 @@ if(WITH_TESTS)
 
   if(ROCKSDB_LIB_FOR_C)
     set(C_TESTS db/c_test.c)
-    # C executables must link to a shared object
     add_executable(c_test db/c_test.c)
-    target_link_libraries(c_test ${ROCKSDB_SHARED_LIB} testharness)
+    target_link_libraries(c_test ${ROCKSDB_LIB_FOR_C} testharness)
     add_test(NAME c_test COMMAND c_test${ARTIFACT_SUFFIX})
     add_dependencies(check c_test)
   endif()
@@ -1225,37 +1281,37 @@ if(WITH_BENCHMARK_TOOLS)
     tools/db_bench.cc
     tools/db_bench_tool.cc)
   target_link_libraries(db_bench
-    ${ROCKSDB_LIB})
+    ${ROCKSDB_LIB} ${THIRDPARTY_LIBS})
 
   add_executable(cache_bench
     cache/cache_bench.cc)
   target_link_libraries(cache_bench
-    ${ROCKSDB_LIB})
+    ${ROCKSDB_LIB} ${GFLAGS_LIB})
 
   add_executable(memtablerep_bench
     memtable/memtablerep_bench.cc)
   target_link_libraries(memtablerep_bench
-    ${ROCKSDB_LIB})
+    ${ROCKSDB_LIB} ${GFLAGS_LIB})
 
   add_executable(range_del_aggregator_bench
     db/range_del_aggregator_bench.cc)
   target_link_libraries(range_del_aggregator_bench
-    ${ROCKSDB_LIB})
+    ${ROCKSDB_LIB} ${GFLAGS_LIB})
 
   add_executable(table_reader_bench
     table/table_reader_bench.cc)
   target_link_libraries(table_reader_bench
-    ${ROCKSDB_LIB} testharness)
+    ${ROCKSDB_LIB} testharness ${GFLAGS_LIB})
 
   add_executable(filter_bench
     util/filter_bench.cc)
   target_link_libraries(filter_bench
-    ${ROCKSDB_LIB})
+    ${ROCKSDB_LIB} ${GFLAGS_LIB})
 
   add_executable(hash_table_bench
     utilities/persistent_cache/hash_table_bench.cc)
   target_link_libraries(hash_table_bench
-    ${ROCKSDB_LIB})
+    ${ROCKSDB_LIB} ${GFLAGS_LIB})
 endif()
 
 if(WITH_CORE_TOOLS OR WITH_TOOLS)

HISTORY.md

@@ -1,23 +1,114 @@
 # Rocksdb Change Log
 ## Unreleased
+### Bug Fixes
+* Fixed a bug in the following combination of features: indexes with user keys (`format_version >= 3`), indexes are partitioned (`index_type == kTwoLevelIndexSearch`), and some index partitions are pinned in memory (`BlockBasedTableOptions::pin_l0_filter_and_index_blocks_in_cache`). The bug could cause keys to be truncated when read from the index leading to wrong read results or other unexpected behavior.
+* Fixed a bug when indexes are partitioned (`index_type == kTwoLevelIndexSearch`), some index partitions are pinned in memory (`BlockBasedTableOptions::pin_l0_filter_and_index_blocks_in_cache`), and partitions reads could be mixed between block cache and directly from the file (e.g., with `enable_index_compression == 1` and `mmap_read == 1`, partitions that were stored uncompressed due to poor compression ratio would be read directly from the file via mmap, while partitions that were stored compressed would be read from block cache). The bug could cause index partitions to be mistakenly considered empty during reads leading to wrong read results.
+* Since 6.12, memtable lookup should report unrecognized value_type as corruption (#7121).
+* Since 6.14, fix false positive flush/compaction `Status::Corruption` failure when `paranoid_file_checks == true` and range tombstones were written to the compaction output files.
+
+### Public API Change
+* Deprecate `BlockBasedTableOptions::pin_l0_filter_and_index_blocks_in_cache` and `BlockBasedTableOptions::pin_top_level_index_and_filter`. These options still take effect until users migrate to the replacement APIs in `BlockBasedTableOptions::metadata_cache_options`. Migration guidance can be found in the API comments on the deprecated options.
+
+## 6.14 (10/09/2020)
+### Bug fixes
+* Fixed a bug after a `CompactRange()` with `CompactRangeOptions::change_level` set fails due to a conflict in the level change step, which caused all subsequent calls to `CompactRange()` with `CompactRangeOptions::change_level` set to incorrectly fail with a `Status::NotSupported("another thread is refitting")` error.
+* Fixed a bug that the bottom most level compaction could still be a trivial move even if `BottommostLevelCompaction.kForce` or `kForceOptimized` is set.
+
+### Public API Change
+* The methods to create and manage EncrypedEnv have been changed.  The EncryptionProvider is now passed to NewEncryptedEnv as a shared pointer, rather than a raw pointer.  Comparably, the CTREncryptedProvider now takes a shared pointer, rather than a reference, to a BlockCipher.  CreateFromString methods have been added to BlockCipher and EncryptionProvider to provide a single API by which different ciphers and providers can be created, respectively.
+* The internal classes (CTREncryptionProvider, ROT13BlockCipher, CTRCipherStream) associated with the EncryptedEnv have been moved out of the public API.  To create a CTREncryptionProvider, one can either use EncryptionProvider::NewCTRProvider, or EncryptionProvider::CreateFromString("CTR").  To create a new ROT13BlockCipher, one can either use BlockCipher::NewROT13Cipher or BlockCipher::CreateFromString("ROT13").
+* The EncryptionProvider::AddCipher method has been added to allow keys to be added to an EncryptionProvider.  This API will allow future providers to support multiple cipher keys.
+* Add a new option "allow_data_in_errors". When this new option is set by users, it allows users to opt-in to get error messages containing corrupted keys/values. Corrupt keys, values will be logged in the messages, logs, status etc. that will help users with the useful information regarding affected data. By default value of this option is set false to prevent users data to be exposed in the messages so currently, data will be redacted from logs, messages, status by default.
+* AdvancedColumnFamilyOptions::force_consistency_checks is now true by default, for more proactive DB corruption detection at virtually no cost (estimated two extra CPU cycles per million on a major production workload). Corruptions reported by these checks now mention "force_consistency_checks" in case a false positive corruption report is suspected and the option needs to be disabled (unlikely). Since existing column families have a saved setting for force_consistency_checks, only new column families will pick up the new default.
+
+### General Improvements
+* The settings of the DBOptions and ColumnFamilyOptions are now managed by Configurable objects (see New Features).  The same convenience methods to configure these options still exist but the backend implementation has been unified under a common implementation.
+
+### New Features
+* Methods to configure serialize, and compare -- such as TableFactory -- are exposed directly through the Configurable base class (from which these objects inherit).  This change will allow for better and more thorough configuration management and retrieval in the future.  The options for a Configurable object can be set via the ConfigureFromMap, ConfigureFromString, or ConfigureOption method.  The serialized version of the options of an object can be retrieved via the GetOptionString, ToString, or GetOption methods.  The list of options supported by an object can be obtained via the GetOptionNames method.  The "raw" object (such as the BlockBasedTableOption) for an option may be retrieved via the GetOptions method.  Configurable options can be compared via the AreEquivalent method.  The settings within a Configurable object may be validated via the ValidateOptions method.  The object may be intialized (at which point only mutable options may be updated) via the PrepareOptions method.
+* Introduce options.check_flush_compaction_key_order with default value to be true. With this option, during flush and compaction, key order will be checked when writing to each SST file. If the order is violated, the flush or compaction will fail.
+* Added is_full_compaction to CompactionJobStats, so that the information is available through the EventListener interface.
+* Add more stats for MultiGet in Histogram to get number of data blocks, index blocks, filter blocks and sst files read from file system per level.
+
+## 6.13 (09/12/2020)
+### Bug fixes
+* Fix a performance regression introduced in 6.4 that makes a upper bound check for every Next() even if keys are within a data block that is within the upper bound.
+* Fix a possible corruption to the LSM state (overlapping files within a level) when a `CompactRange()` for refitting levels (`CompactRangeOptions::change_level == true`) and another manual compaction are executed in parallel.
+* Sanitize `recycle_log_file_num` to zero when the user attempts to enable it in combination with `WALRecoveryMode::kTolerateCorruptedTailRecords`. Previously the two features were allowed together, which compromised the user's configured crash-recovery guarantees.
+* Fix a bug where a level refitting in CompactRange() might race with an automatic compaction that puts the data to the target level of the refitting. The bug has been there for years.
+* Fixed a bug in version 6.12 in which BackupEngine::CreateNewBackup could fail intermittently with non-OK status when backing up a read-write DB configured with a DBOptions::file_checksum_gen_factory.
+* Fix useless no-op compactions scheduled upon snapshot release when options.disable-auto-compactions = true.
+* Fix a bug when max_write_buffer_size_to_maintain is set, immutable flushed memtable destruction is delayed until the next super version is installed. A memtable is not added to delete list because of its reference hold by super version and super version doesn't switch because of empt delete list. So memory usage keeps on increasing beyond write_buffer_size + max_write_buffer_size_to_maintain.
+* Avoid converting MERGES to PUTS when allow_ingest_behind is true.
+* Fix compression dictionary sampling together with `SstFileWriter`. Previously, the dictionary would be trained/finalized immediately with zero samples. Now, the whole `SstFileWriter` file is buffered in memory and then sampled.
+* Fix a bug with `avoid_unnecessary_blocking_io=1` and creating backups (BackupEngine::CreateNewBackup) or checkpoints (Checkpoint::Create). With this setting and WAL enabled, these operations could randomly fail with non-OK status.
+* Fix a bug in which bottommost compaction continues to advance the underlying InternalIterator to skip tombstones even after shutdown.
+
+### New Features
+* A new field `std::string requested_checksum_func_name` is added to `FileChecksumGenContext`, which enables the checksum factory to create generators for a suite of different functions.
+* Added a new subcommand, `ldb unsafe_remove_sst_file`, which removes a lost or corrupt SST file from a DB's metadata. This command involves data loss and must not be used on a live DB.
+
+### Performance Improvements
+* Reduce thread number for multiple DB instances by re-using one global thread for statistics dumping and persisting.
+* Reduce write-amp in heavy write bursts in `kCompactionStyleLevel` compaction style with `level_compaction_dynamic_level_bytes` set.
+* BackupEngine incremental backups no longer read DB table files that are already saved to a shared part of the backup directory, unless `share_files_with_checksum` is used with `kLegacyCrc32cAndFileSize` naming (discouraged).
+  * For `share_files_with_checksum`, we are confident there is no regression (vs. pre-6.12) in detecting DB or backup corruption at backup creation time, mostly because the old design did not leverage this extra checksum computation for detecting inconsistencies at backup creation time.
+  * For `share_table_files` without "checksum" (not recommended), there is a regression in detecting fundamentally unsafe use of the option, greatly mitigated by file size checking (under "Behavior Changes"). Almost no reason to use `share_files_with_checksum=false` should remain.
+  * `DB::VerifyChecksum` and `BackupEngine::VerifyBackup` with checksum checking are still able to catch corruptions that `CreateNewBackup` does not.
+
+### Public API Change
+* Expose kTypeDeleteWithTimestamp in EntryType and update GetEntryType() accordingly.
+* Added file_checksum and file_checksum_func_name to TableFileCreationInfo, which can pass the table file checksum information through the OnTableFileCreated callback during flush and compaction.
+* A warning is added to `DB::DeleteFile()` API describing its known problems and deprecation plan.
+* Add a new stats level, i.e. StatsLevel::kExceptTickers (PR7329) to exclude tickers even if application passes a non-null Statistics object.
+* Added a new status code IOStatus::IOFenced() for the Env/FileSystem to indicate that writes from this instance are fenced off. Like any other background error, this error is returned to the user in Put/Merge/Delete/Flush calls and can be checked using Status::IsIOFenced().
+
+### Behavior Changes
+* File abstraction `FSRandomAccessFile.Prefetch()` default return status is changed from `OK` to `NotSupported`. If the user inherited file doesn't implement prefetch, RocksDB will create internal prefetch buffer to improve read performance.
+* When retryabel IO error happens during Flush (manifest write error is excluded) and WAL is disabled, originally it is mapped to kHardError. Now,it is mapped to soft error. So DB will not stall the writes unless the memtable is full. At the same time, when auto resume is triggered to recover the retryable IO error during Flush, SwitchMemtable is not called to avoid generating to many small immutable memtables. If WAL is enabled, no behavior changes.
+* When considering whether a table file is already backed up in a shared part of backup directory, BackupEngine would already query the sizes of source (DB) and pre-existing destination (backup) files. BackupEngine now uses these file sizes to detect corruption, as at least one of (a) old backup, (b) backup in progress, or (c) current DB is corrupt if there's a size mismatch.
+
+### Others
+* Error in prefetching partitioned index blocks will not be swallowed. It will fail the query and return the IOError users.
+
+## 6.12 (2020-07-28)
 ### Public API Change
 * Encryption file classes now exposed for inheritance in env_encryption.h
+* File I/O listener is extended to cover more I/O operations. Now class `EventListener` in listener.h contains new callback functions: `OnFileFlushFinish()`, `OnFileSyncFinish()`, `OnFileRangeSyncFinish()`, `OnFileTruncateFinish()`, and ``OnFileCloseFinish()``.
+* `FileOperationInfo` now reports `duration` measured by `std::chrono::steady_clock` and `start_ts` measured by `std::chrono::system_clock` instead of start and finish timestamps measured by `system_clock`. Note that `system_clock` is called before `steady_clock` in program order at operation starts.
+* `DB::GetDbSessionId(std::string& session_id)` is added. `session_id` stores a unique identifier that gets reset every time the DB is opened. This DB session ID should be unique among all open DB instances on all hosts, and should be unique among re-openings of the same or other DBs. This identifier is recorded in the LOG file on the line starting with "DB Session ID:".
+* `DB::OpenForReadOnly()` now returns `Status::NotFound` when the specified DB directory does not exist. Previously the error returned depended on the underlying `Env`. This change is available in all 6.11 releases as well.
+* A parameter `verify_with_checksum` is added to `BackupEngine::VerifyBackup`, which is false by default. If it is ture, `BackupEngine::VerifyBackup` verifies checksums and file sizes of backup files. Pass `false` for `verify_with_checksum` to maintain the previous behavior and performance of `BackupEngine::VerifyBackup`, by only verifying sizes of backup files.
 
 ### Behavior Changes
 * Best-efforts recovery ignores CURRENT file completely. If CURRENT file is missing during recovery, best-efforts recovery still proceeds with MANIFEST file(s).
 * In best-efforts recovery, an error that is not Corruption or IOError::kNotFound or IOError::kPathNotFound will be overwritten silently. Fix this by checking all non-ok cases and return early.
+* When `file_checksum_gen_factory` is set to `GetFileChecksumGenCrc32cFactory()`, BackupEngine will compare the crc32c checksums of table files computed when creating a backup to the expected checksums stored in the DB manifest, and will fail `CreateNewBackup()` on mismatch (corruption). If the `file_checksum_gen_factory` is not set or set to any other customized factory, there is no checksum verification to detect if SST files in a DB are corrupt when read, copied, and independently checksummed by BackupEngine.
+* When a DB sets `stats_dump_period_sec > 0`, either as the initial value for DB open or as a dynamic option change, the first stats dump is staggered in the following X seconds, where X is an integer in `[0, stats_dump_period_sec)`. Subsequent stats dumps are still spaced `stats_dump_period_sec` seconds apart.
+* When the paranoid_file_checks option is true, a hash is generated of all keys and values are generated when the SST file is written, and then the values are read back in to validate the file.  A corruption is signaled if the two hashes do not match.
+
+### Bug fixes
 * Compressed block cache was automatically disabled with read-only DBs by mistake. Now it is fixed: compressed block cache will be in effective with read-only DB too.
 * Fix a bug of wrong iterator result if another thread finishes an update and a DB flush between two statement.
-
-### Public API Change
-* `DB::GetDbSessionId(std::string& session_id)` is added. `session_id` stores a unique identifier that gets reset every time the DB is opened. This DB session ID should be unique among all open DB instances on all hosts, and should be unique among re-openings of the same or other DBs. This identifier is recorded in the LOG file on the line starting with "DB Session ID:".
+* Disable file deletion after MANIFEST write/sync failure until db re-open or Resume() so that subsequent re-open will not see MANIFEST referencing deleted SSTs.
+* Fix a bug when index_type == kTwoLevelIndexSearch in PartitionedIndexBuilder to update FlushPolicy to point to internal key partitioner when it changes from user-key mode to internal-key mode in index partition.
+* Make compaction report InternalKey corruption while iterating over the input.
+* Fix a bug which may cause MultiGet to be slow because it may read more data than requested, but this won't affect correctness. The bug was introduced in 6.10 release.
+* Fail recovery and report once hitting a physical log record checksum mismatch, while reading MANIFEST. RocksDB should not continue processing the MANIFEST any further.
+* Fixed a bug in size-amp-triggered and periodic-triggered universal compaction, where the compression settings for the first input level were used rather than the compression settings for the output (bottom) level.
 
 ### New Features
 * DB identity (`db_id`) and DB session identity (`db_session_id`) are added to table properties and stored in SST files. SST files generated from SstFileWriter and Repairer have DB identity “SST Writer” and “DB Repairer”, respectively. Their DB session IDs are generated in the same way as `DB::GetDbSessionId`. The session ID for SstFileWriter (resp., Repairer) resets every time `SstFileWriter::Open` (resp., `Repairer::Run`) is called.
 * Added experimental option BlockBasedTableOptions::optimize_filters_for_memory for reducing allocated memory size of Bloom filters (~10% savings with Jemalloc) while preserving the same general accuracy. To have an effect, the option requires format_version=5 and malloc_usable_size. Enabling this option is forward and backward compatible with existing format_version=5.
+* `BackupableDBOptions::share_files_with_checksum_naming` is added with new default behavior for naming backup files with `share_files_with_checksum`, to address performance and backup integrity issues. See API comments for details.
+* Added auto resume function to automatically recover the DB from background Retryable IO Error. When retryable IOError happens during flush and WAL write, the error is mapped to Hard Error and DB will be in read mode. When retryable IO Error happens during compaction, the error will be mapped to Soft Error. DB is still in write/read mode. Autoresume function will create a thread for a DB to call DB->ResumeImpl() to try the recover for Retryable IO Error during flush and WAL write. Compaction will be rescheduled by itself if retryable IO Error happens. Auto resume may also cause other Retryable IO Error during the recovery, so the recovery will fail. Retry the auto resume may solve the issue, so we use max_bgerror_resume_count to decide how many resume cycles will be tried in total. If it is <=0, auto resume retryable IO Error is disabled. Default is INT_MAX, which will lead to a infinit auto resume. bgerror_resume_retry_interval decides the time interval between two auto resumes.
+* Option `max_subcompactions` can be set dynamically using DB::SetDBOptions().
+* Added experimental ColumnFamilyOptions::sst_partitioner_factory to define determine the partitioning of sst files. This helps compaction to split the files on interesting boundaries (key prefixes) to make propagation of sst files less write amplifying (covering the whole key space).
 
-### Bug Fixes
-* Fail recovery and report once hitting a physical log record checksum mismatch, while reading MANIFEST. RocksDB should not continue processing the MANIFEST any further.
+### Performance Improvements
+* Eliminate key copies for internal comparisons while accessing ingested block-based tables.
+* Reduce key comparisons during random access in all block-based tables.
+* BackupEngine avoids unnecessary repeated checksum computation for backing up a table file to the `shared_checksum` directory when using `share_files_with_checksum_naming = kUseDbSessionId` (new default), except on SST files generated before this version of RocksDB, which fall back on using `kLegacyCrc32cAndFileSize`.
 
 ## 6.11 (6/12/2020)
 ### Bug Fixes

LANGUAGE-BINDINGS.md

@@ -10,7 +10,9 @@ This is the list of all known third-party language bindings for RocksDB. If some
 * Ruby - http://rubygems.org/gems/rocksdb-ruby
 * Haskell - https://hackage.haskell.org/package/rocksdb-haskell
 * PHP - https://github.com/Photonios/rocksdb-php
-* C# - https://github.com/warrenfalk/rocksdb-sharp
+* C#
+    * https://github.com/warrenfalk/rocksdb-sharp
+    * https://github.com/curiosity-ai/rocksdb-sharp
 * Rust
     * https://github.com/pingcap/rust-rocksdb (used in production fork of https://github.com/spacejam/rust-rocksdb)
     * https://github.com/spacejam/rust-rocksdb

README.md

@@ -1,7 +1,8 @@
 ## RocksDB: A Persistent Key-Value Store for Flash and RAM Storage
 
-[![Linux/Mac Build Status](https://travis-ci.org/facebook/rocksdb.svg?branch=master)](https://travis-ci.org/facebook/rocksdb)
-[![Windows Build status](https://ci.appveyor.com/api/projects/status/fbgfu0so3afcno78/branch/master?svg=true)](https://ci.appveyor.com/project/Facebook/rocksdb/branch/master)
+[![CircleCI Status](https://circleci.com/gh/facebook/rocksdb.svg?style=svg)](https://circleci.com/gh/facebook/rocksdb)
+[![TravisCI Status](https://travis-ci.org/facebook/rocksdb.svg?branch=master)](https://travis-ci.org/facebook/rocksdb)
+[![Appveyor Build status](https://ci.appveyor.com/api/projects/status/fbgfu0so3afcno78/branch/master?svg=true)](https://ci.appveyor.com/project/Facebook/rocksdb/branch/master)
 [![PPC64le Build Status](http://140.211.168.68:8080/buildStatus/icon?job=Rocksdb)](http://140.211.168.68:8080/job/Rocksdb)
 
 RocksDB is developed and maintained by Facebook Database Engineering Team.

TARGETS

@@ -25,13 +25,16 @@ ROCKSDB_EXTERNAL_DEPS = [
     ("gflags", None, "gflags"),
     ("lz4", None, "lz4"),
     ("zstd", None),
-    ("tbb", None),
 ]
 
 ROCKSDB_OS_DEPS = [
     (
         "linux",
-        ["third-party//numa:numa", "third-party//liburing:uring"],
+        ["third-party//numa:numa", "third-party//liburing:uring", "third-party//tbb:tbb"],
+    ),
+    (
+        "macos",
+        ["third-party//tbb:tbb"],
     ),
 ]
 
@@ -49,17 +52,27 @@ ROCKSDB_OS_PREPROCESSOR_FLAGS = [
             "-DHAVE_SSE42",
             "-DLIBURING",
             "-DNUMA",
+            "-DROCKSDB_PLATFORM_POSIX",
+            "-DROCKSDB_LIB_IO_POSIX",
+            "-DTBB",
         ],
     ),
     (
         "macos",
-        ["-DOS_MACOSX"],
+        [
+            "-DOS_MACOSX",
+            "-DROCKSDB_PLATFORM_POSIX",
+            "-DROCKSDB_LIB_IO_POSIX",
+            "-DTBB",
+        ],
+    ),
+    (
+        "windows",
+        [ "-DOS_WIN", "-DWIN32", "-D_MBCS", "-DWIN64", "-DNOMINMAX" ]
     ),
 ]
 
 ROCKSDB_PREPROCESSOR_FLAGS = [
-    "-DROCKSDB_PLATFORM_POSIX",
-    "-DROCKSDB_LIB_IO_POSIX",
     "-DROCKSDB_SUPPORT_THREAD_LOCAL",
 
     # Flags to enable libs we include
@@ -70,7 +83,6 @@ ROCKSDB_PREPROCESSOR_FLAGS = [
     "-DZSTD",
     "-DZSTD_STATIC_LINKING_ONLY",
     "-DGFLAGS=gflags",
-    "-DTBB",
 
     # Added missing flags from output of build_detect_platform
     "-DROCKSDB_BACKTRACE",
@@ -122,10 +134,12 @@ cpp_library(
         "cache/sharded_cache.cc",
         "db/arena_wrapped_db_iter.cc",
         "db/blob/blob_file_addition.cc",
+        "db/blob/blob_file_builder.cc",
         "db/blob/blob_file_garbage.cc",
         "db/blob/blob_file_meta.cc",
+        "db/blob/blob_file_reader.cc",
         "db/blob/blob_log_format.cc",
-        "db/blob/blob_log_reader.cc",
+        "db/blob/blob_log_sequential_reader.cc",
         "db/blob/blob_log_writer.cc",
         "db/builder.cc",
         "db/c.cc",
@@ -138,6 +152,7 @@ cpp_library(
         "db/compaction/compaction_picker_fifo.cc",
         "db/compaction/compaction_picker_level.cc",
         "db/compaction/compaction_picker_universal.cc",
+        "db/compaction/sst_partitioner.cc",
         "db/convenience.cc",
         "db/db_filesnapshot.cc",
         "db/db_impl/db_impl.cc",
@@ -170,6 +185,8 @@ cpp_library(
         "db/memtable_list.cc",
         "db/merge_helper.cc",
         "db/merge_operator.cc",
+        "db/output_validator.cc",
+        "db/periodic_work_scheduler.cc",
         "db/range_del_aggregator.cc",
         "db/range_tombstone_fragmenter.cc",
         "db/repair.cc",
@@ -182,6 +199,7 @@ cpp_library(
         "db/version_edit.cc",
         "db/version_edit_handler.cc",
         "db/version_set.cc",
+        "db/wal_edit.cc",
         "db/wal_manager.cc",
         "db/write_batch.cc",
         "db/write_batch_base.cc",
@@ -193,6 +211,7 @@ cpp_library(
         "env/env_hdfs.cc",
         "env/env_posix.cc",
         "env/file_system.cc",
+        "env/file_system_tracer.cc",
         "env/fs_posix.cc",
         "env/io_posix.cc",
         "env/mock_env.cc",
@@ -234,12 +253,19 @@ cpp_library(
         "monitoring/thread_status_util.cc",
         "monitoring/thread_status_util_debug.cc",
         "options/cf_options.cc",
+        "options/configurable.cc",
         "options/db_options.cc",
         "options/options.cc",
         "options/options_helper.cc",
         "options/options_parser.cc",
         "port/port_posix.cc",
         "port/stack_trace.cc",
+        "port/win/env_default.cc",
+        "port/win/env_win.cc",
+        "port/win/io_win.cc",
+        "port/win/port_win.cc",
+        "port/win/win_logger.cc",
+        "port/win/win_thread.cc",
         "table/adaptive/adaptive_table_factory.cc",
         "table/block_based/binary_search_index_reader.cc",
         "table/block_based/block.cc",
@@ -282,14 +308,17 @@ cpp_library(
         "table/plain/plain_table_index.cc",
         "table/plain/plain_table_key_coding.cc",
         "table/plain/plain_table_reader.cc",
+        "table/sst_file_dumper.cc",
         "table/sst_file_reader.cc",
         "table/sst_file_writer.cc",
+        "table/table_factory.cc",
         "table/table_properties.cc",
         "table/two_level_iterator.cc",
         "test_util/sync_point.cc",
         "test_util/sync_point_impl.cc",
         "test_util/transaction_test_util.cc",
         "tools/dump/db_dump_tool.cc",
+        "tools/io_tracer_parser_tool.cc",
         "tools/ldb_cmd.cc",
         "tools/ldb_tool.cc",
         "tools/sst_dump_tool.cc",
@@ -331,6 +360,8 @@ cpp_library(
         "utilities/debug.cc",
         "utilities/env_mirror.cc",
         "utilities/env_timed.cc",
+        "utilities/fault_injection_env.cc",
+        "utilities/fault_injection_fs.cc",
         "utilities/leveldb_options/leveldb_options.cc",
         "utilities/memory/memory_util.cc",
         "utilities/merge_operators/bytesxor.cc",
@@ -352,6 +383,8 @@ cpp_library(
         "utilities/simulator_cache/sim_cache.cc",
         "utilities/table_properties_collectors/compact_on_deletion_collector.cc",
         "utilities/trace/file_trace_reader_writer.cc",
+        "utilities/transactions/lock/lock_tracker.cc",
+        "utilities/transactions/lock/point_lock_tracker.cc",
         "utilities/transactions/optimistic_transaction.cc",
         "utilities/transactions/optimistic_transaction_db_impl.cc",
         "utilities/transactions/pessimistic_transaction.cc",
@@ -377,6 +410,295 @@ cpp_library(
     preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
     deps = [],
     external_deps = ROCKSDB_EXTERNAL_DEPS,
+    link_whole = False,
+)
+
+cpp_library(
+    name = "rocksdb_whole_archive_lib",
+    srcs = [
+        "cache/cache.cc",
+        "cache/clock_cache.cc",
+        "cache/lru_cache.cc",
+        "cache/sharded_cache.cc",
+        "db/arena_wrapped_db_iter.cc",
+        "db/blob/blob_file_addition.cc",
+        "db/blob/blob_file_builder.cc",
+        "db/blob/blob_file_garbage.cc",
+        "db/blob/blob_file_meta.cc",
+        "db/blob/blob_file_reader.cc",
+        "db/blob/blob_log_format.cc",
+        "db/blob/blob_log_sequential_reader.cc",
+        "db/blob/blob_log_writer.cc",
+        "db/builder.cc",
+        "db/c.cc",
+        "db/column_family.cc",
+        "db/compacted_db_impl.cc",
+        "db/compaction/compaction.cc",
+        "db/compaction/compaction_iterator.cc",
+        "db/compaction/compaction_job.cc",
+        "db/compaction/compaction_picker.cc",
+        "db/compaction/compaction_picker_fifo.cc",
+        "db/compaction/compaction_picker_level.cc",
+        "db/compaction/compaction_picker_universal.cc",
+        "db/compaction/sst_partitioner.cc",
+        "db/convenience.cc",
+        "db/db_filesnapshot.cc",
+        "db/db_impl/db_impl.cc",
+        "db/db_impl/db_impl_compaction_flush.cc",
+        "db/db_impl/db_impl_debug.cc",
+        "db/db_impl/db_impl_experimental.cc",
+        "db/db_impl/db_impl_files.cc",
+        "db/db_impl/db_impl_open.cc",
+        "db/db_impl/db_impl_readonly.cc",
+        "db/db_impl/db_impl_secondary.cc",
+        "db/db_impl/db_impl_write.cc",
+        "db/db_info_dumper.cc",
+        "db/db_iter.cc",
+        "db/dbformat.cc",
+        "db/error_handler.cc",
+        "db/event_helpers.cc",
+        "db/experimental.cc",
+        "db/external_sst_file_ingestion_job.cc",
+        "db/file_indexer.cc",
+        "db/flush_job.cc",
+        "db/flush_scheduler.cc",
+        "db/forward_iterator.cc",
+        "db/import_column_family_job.cc",
+        "db/internal_stats.cc",
+        "db/log_reader.cc",
+        "db/log_writer.cc",
+        "db/logs_with_prep_tracker.cc",
+        "db/malloc_stats.cc",
+        "db/memtable.cc",
+        "db/memtable_list.cc",
+        "db/merge_helper.cc",
+        "db/merge_operator.cc",
+        "db/output_validator.cc",
+        "db/periodic_work_scheduler.cc",
+        "db/range_del_aggregator.cc",
+        "db/range_tombstone_fragmenter.cc",
+        "db/repair.cc",
+        "db/snapshot_impl.cc",
+        "db/table_cache.cc",
+        "db/table_properties_collector.cc",
+        "db/transaction_log_impl.cc",
+        "db/trim_history_scheduler.cc",
+        "db/version_builder.cc",
+        "db/version_edit.cc",
+        "db/version_edit_handler.cc",
+        "db/version_set.cc",
+        "db/wal_edit.cc",
+        "db/wal_manager.cc",
+        "db/write_batch.cc",
+        "db/write_batch_base.cc",
+        "db/write_controller.cc",
+        "db/write_thread.cc",
+        "env/env.cc",
+        "env/env_chroot.cc",
+        "env/env_encryption.cc",
+        "env/env_hdfs.cc",
+        "env/env_posix.cc",
+        "env/file_system.cc",
+        "env/file_system_tracer.cc",
+        "env/fs_posix.cc",
+        "env/io_posix.cc",
+        "env/mock_env.cc",
+        "file/delete_scheduler.cc",
+        "file/file_prefetch_buffer.cc",
+        "file/file_util.cc",
+        "file/filename.cc",
+        "file/random_access_file_reader.cc",
+        "file/read_write_util.cc",
+        "file/readahead_raf.cc",
+        "file/sequence_file_reader.cc",
+        "file/sst_file_manager_impl.cc",
+        "file/writable_file_writer.cc",
+        "logging/auto_roll_logger.cc",
+        "logging/event_logger.cc",
+        "logging/log_buffer.cc",
+        "memory/arena.cc",
+        "memory/concurrent_arena.cc",
+        "memory/jemalloc_nodump_allocator.cc",
+        "memory/memkind_kmem_allocator.cc",
+        "memtable/alloc_tracker.cc",
+        "memtable/hash_linklist_rep.cc",
+        "memtable/hash_skiplist_rep.cc",
+        "memtable/skiplistrep.cc",
+        "memtable/vectorrep.cc",
+        "memtable/write_buffer_manager.cc",
+        "monitoring/histogram.cc",
+        "monitoring/histogram_windowing.cc",
+        "monitoring/in_memory_stats_history.cc",
+        "monitoring/instrumented_mutex.cc",
+        "monitoring/iostats_context.cc",
+        "monitoring/perf_context.cc",
+        "monitoring/perf_level.cc",
+        "monitoring/persistent_stats_history.cc",
+        "monitoring/statistics.cc",
+        "monitoring/thread_status_impl.cc",
+        "monitoring/thread_status_updater.cc",
+        "monitoring/thread_status_updater_debug.cc",
+        "monitoring/thread_status_util.cc",
+        "monitoring/thread_status_util_debug.cc",
+        "options/cf_options.cc",
+        "options/configurable.cc",
+        "options/db_options.cc",
+        "options/options.cc",
+        "options/options_helper.cc",
+        "options/options_parser.cc",
+        "port/port_posix.cc",
+        "port/stack_trace.cc",
+        "port/win/env_default.cc",
+        "port/win/env_win.cc",
+        "port/win/io_win.cc",
+        "port/win/port_win.cc",
+        "port/win/win_logger.cc",
+        "port/win/win_thread.cc",
+        "table/adaptive/adaptive_table_factory.cc",
+        "table/block_based/binary_search_index_reader.cc",
+        "table/block_based/block.cc",
+        "table/block_based/block_based_filter_block.cc",
+        "table/block_based/block_based_table_builder.cc",
+        "table/block_based/block_based_table_factory.cc",
+        "table/block_based/block_based_table_iterator.cc",
+        "table/block_based/block_based_table_reader.cc",
+        "table/block_based/block_builder.cc",
+        "table/block_based/block_prefetcher.cc",
+        "table/block_based/block_prefix_index.cc",
+        "table/block_based/data_block_footer.cc",
+        "table/block_based/data_block_hash_index.cc",
+        "table/block_based/filter_block_reader_common.cc",
+        "table/block_based/filter_policy.cc",
+        "table/block_based/flush_block_policy.cc",
+        "table/block_based/full_filter_block.cc",
+        "table/block_based/hash_index_reader.cc",
+        "table/block_based/index_builder.cc",
+        "table/block_based/index_reader_common.cc",
+        "table/block_based/parsed_full_filter_block.cc",
+        "table/block_based/partitioned_filter_block.cc",
+        "table/block_based/partitioned_index_iterator.cc",
+        "table/block_based/partitioned_index_reader.cc",
+        "table/block_based/reader_common.cc",
+        "table/block_based/uncompression_dict_reader.cc",
+        "table/block_fetcher.cc",
+        "table/cuckoo/cuckoo_table_builder.cc",
+        "table/cuckoo/cuckoo_table_factory.cc",
+        "table/cuckoo/cuckoo_table_reader.cc",
+        "table/format.cc",
+        "table/get_context.cc",
+        "table/iterator.cc",
+        "table/merging_iterator.cc",
+        "table/meta_blocks.cc",
+        "table/persistent_cache_helper.cc",
+        "table/plain/plain_table_bloom.cc",
+        "table/plain/plain_table_builder.cc",
+        "table/plain/plain_table_factory.cc",
+        "table/plain/plain_table_index.cc",
+        "table/plain/plain_table_key_coding.cc",
+        "table/plain/plain_table_reader.cc",
+        "table/sst_file_dumper.cc",
+        "table/sst_file_reader.cc",
+        "table/sst_file_writer.cc",
+        "table/table_factory.cc",
+        "table/table_properties.cc",
+        "table/two_level_iterator.cc",
+        "test_util/sync_point.cc",
+        "test_util/sync_point_impl.cc",
+        "test_util/transaction_test_util.cc",
+        "tools/dump/db_dump_tool.cc",
+        "tools/io_tracer_parser_tool.cc",
+        "tools/ldb_cmd.cc",
+        "tools/ldb_tool.cc",
+        "tools/sst_dump_tool.cc",
+        "trace_replay/block_cache_tracer.cc",
+        "trace_replay/io_tracer.cc",
+        "trace_replay/trace_replay.cc",
+        "util/build_version.cc",
+        "util/coding.cc",
+        "util/compaction_job_stats_impl.cc",
+        "util/comparator.cc",
+        "util/compression_context_cache.cc",
+        "util/concurrent_task_limiter_impl.cc",
+        "util/crc32c.cc",
+        "util/dynamic_bloom.cc",
+        "util/file_checksum_helper.cc",
+        "util/hash.cc",
+        "util/murmurhash.cc",
+        "util/random.cc",
+        "util/rate_limiter.cc",
+        "util/slice.cc",
+        "util/status.cc",
+        "util/string_util.cc",
+        "util/thread_local.cc",
+        "util/threadpool_imp.cc",
+        "util/xxhash.cc",
+        "utilities/backupable/backupable_db.cc",
+        "utilities/blob_db/blob_compaction_filter.cc",
+        "utilities/blob_db/blob_db.cc",
+        "utilities/blob_db/blob_db_impl.cc",
+        "utilities/blob_db/blob_db_impl_filesnapshot.cc",
+        "utilities/blob_db/blob_dump_tool.cc",
+        "utilities/blob_db/blob_file.cc",
+        "utilities/cassandra/cassandra_compaction_filter.cc",
+        "utilities/cassandra/format.cc",
+        "utilities/cassandra/merge_operator.cc",
+        "utilities/checkpoint/checkpoint_impl.cc",
+        "utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc",
+        "utilities/convenience/info_log_finder.cc",
+        "utilities/debug.cc",
+        "utilities/env_mirror.cc",
+        "utilities/env_timed.cc",
+        "utilities/fault_injection_env.cc",
+        "utilities/fault_injection_fs.cc",
+        "utilities/leveldb_options/leveldb_options.cc",
+        "utilities/memory/memory_util.cc",
+        "utilities/merge_operators/bytesxor.cc",
+        "utilities/merge_operators/max.cc",
+        "utilities/merge_operators/put.cc",
+        "utilities/merge_operators/sortlist.cc",
+        "utilities/merge_operators/string_append/stringappend.cc",
+        "utilities/merge_operators/string_append/stringappend2.cc",
+        "utilities/merge_operators/uint64add.cc",
+        "utilities/object_registry.cc",
+        "utilities/option_change_migration/option_change_migration.cc",
+        "utilities/options/options_util.cc",
+        "utilities/persistent_cache/block_cache_tier.cc",
+        "utilities/persistent_cache/block_cache_tier_file.cc",
+        "utilities/persistent_cache/block_cache_tier_metadata.cc",
+        "utilities/persistent_cache/persistent_cache_tier.cc",
+        "utilities/persistent_cache/volatile_tier_impl.cc",
+        "utilities/simulator_cache/cache_simulator.cc",
+        "utilities/simulator_cache/sim_cache.cc",
+        "utilities/table_properties_collectors/compact_on_deletion_collector.cc",
+        "utilities/trace/file_trace_reader_writer.cc",
+        "utilities/transactions/lock/lock_tracker.cc",
+        "utilities/transactions/lock/point_lock_tracker.cc",
+        "utilities/transactions/optimistic_transaction.cc",
+        "utilities/transactions/optimistic_transaction_db_impl.cc",
+        "utilities/transactions/pessimistic_transaction.cc",
+        "utilities/transactions/pessimistic_transaction_db.cc",
+        "utilities/transactions/snapshot_checker.cc",
+        "utilities/transactions/transaction_base.cc",
+        "utilities/transactions/transaction_db_mutex_impl.cc",
+        "utilities/transactions/transaction_lock_mgr.cc",
+        "utilities/transactions/transaction_util.cc",
+        "utilities/transactions/write_prepared_txn.cc",
+        "utilities/transactions/write_prepared_txn_db.cc",
+        "utilities/transactions/write_unprepared_txn.cc",
+        "utilities/transactions/write_unprepared_txn_db.cc",
+        "utilities/ttl/db_ttl_impl.cc",
+        "utilities/write_batch_with_index/write_batch_with_index.cc",
+        "utilities/write_batch_with_index/write_batch_with_index_internal.cc",
+    ],
+    auto_headers = AutoHeaders.RECURSIVE_GLOB,
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    os_deps = ROCKSDB_OS_DEPS,
+    os_preprocessor_flags = ROCKSDB_OS_PREPROCESSOR_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+    deps = [],
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
+    link_whole = True,
 )
 
 cpp_library(
@@ -384,8 +706,7 @@ cpp_library(
     srcs = [
         "db/db_test_util.cc",
         "table/mock_table.cc",
-        "test_util/fault_injection_test_env.cc",
-        "test_util/fault_injection_test_fs.cc",
+        "test_util/mock_time_env.cc",
         "test_util/testharness.cc",
         "test_util/testutil.cc",
         "tools/block_cache_analyzer/block_cache_trace_analyzer.cc",
@@ -402,6 +723,7 @@ cpp_library(
     external_deps = ROCKSDB_EXTERNAL_DEPS + [
         ("googletest", None, "gtest"),
     ],
+    link_whole = False,
 )
 
 cpp_library(
@@ -420,6 +742,7 @@ cpp_library(
     preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
     deps = [":rocksdb_lib"],
     external_deps = ROCKSDB_EXTERNAL_DEPS,
+    link_whole = False,
 )
 
 cpp_library(
@@ -448,6 +771,27 @@ cpp_library(
     external_deps = ROCKSDB_EXTERNAL_DEPS,
 )
 
+if not is_opt_mode:
+    cpp_binary(
+        name = "c_test_bin",
+        srcs = ["db/c_test.c"],
+        arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+        os_preprocessor_flags = ROCKSDB_OS_PREPROCESSOR_FLAGS,
+        compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+        deps = [":rocksdb_test_lib"],
+    )
+
+if not is_opt_mode:
+    custom_unittest(
+        "c_test",
+        command = [
+            native.package_name() + "/buckifier/rocks_test_runner.sh",
+            "$(location :{})".format("c_test_bin"),
+        ],
+        type = "simple",
+    )
+
 cpp_library(
     name = "env_basic_test_lib",
     srcs = ["env/env_basic_test.cc"],
@@ -459,6 +803,7 @@ cpp_library(
     preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
     deps = [":rocksdb_test_lib"],
     external_deps = ROCKSDB_EXTERNAL_DEPS,
+    link_whole = False,
 )
 
 # [test_name, test_src, test_type, extra_deps, extra_compiler_flags]
@@ -505,6 +850,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "blob_file_builder_test",
+        "db/blob/blob_file_builder_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "blob_file_garbage_test",
         "db/blob/blob_file_garbage_test.cc",
@@ -512,6 +864,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "blob_file_reader_test",
+        "db/blob/blob_file_reader_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "block_based_filter_block_test",
         "table/block_based/block_based_filter_block_test.cc",
@@ -561,13 +920,6 @@ ROCKS_TESTS = [
         [],
         [],
     ],
-    [
-        "c_test",
-        "db/c_test.c",
-        "serial",
-        [],
-        [],
-    ],
     [
         "cache_simulator_test",
         "utilities/simulator_cache/cache_simulator_test.cc",
@@ -634,7 +986,7 @@ ROCKS_TESTS = [
     [
         "column_family_test",
         "db/column_family_test.cc",
-        "serial",
+        "parallel",
         [],
         [],
     ],
@@ -687,6 +1039,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "configurable_test",
+        "options/configurable_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "corruption_test",
         "db/corruption_test.cc",
@@ -928,7 +1287,7 @@ ROCKS_TESTS = [
     [
         "db_test2",
         "db/db_test2.cc",
-        "serial",
+        "parallel",
         [],
         [],
     ],
@@ -1156,6 +1515,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "io_tracer_parser_test",
+        "tools/io_tracer_parser_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "io_tracer_test",
         "trace_replay/io_tracer_test.cc",
@@ -1324,6 +1690,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "periodic_work_scheduler_test",
+        "db/periodic_work_scheduler_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "persistent_cache_test",
         "utilities/persistent_cache/persistent_cache_test.cc",
@@ -1338,6 +1711,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "prefetch_test",
+        "file/prefetch_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "prefix_test",
         "db/prefix_test.cc",
@@ -1643,7 +2023,7 @@ ROCKS_TESTS = [
         srcs = [test_cc],
         arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
         os_preprocessor_flags = ROCKSDB_OS_PREPROCESSOR_FLAGS,
-        compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        compiler_flags = ROCKSDB_COMPILER_FLAGS + extra_compiler_flags,
         preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
         deps = [":rocksdb_test_lib"] + extra_deps,
         external_deps = ROCKSDB_EXTERNAL_DEPS + [

appveyor.yml

@@ -24,15 +24,6 @@ environment:
     - APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2017
       CMAKE_GENERATOR: Visual Studio 15 Win64
       DEV_ENV: C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\Common7\IDE\devenv.com
-    - APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2019
-      CMAKE_GENERATOR: Visual Studio 16
-      CMAKE_PLATEFORM_NAME: x64
-      DEV_ENV: C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\Common7\IDE\devenv.com
-    - APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2019
-      CMAKE_GENERATOR: Visual Studio 16
-      CMAKE_PLATEFORM_NAME: x64
-      CMAKE_OPT: -DCMAKE_CXX_STANDARD=20
-      DEV_ENV: C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\Common7\IDE\devenv.com
 
 install:
   - md %THIRDPARTY_HOME%

buckifier/buckify_rocksdb.py

@@ -48,8 +48,8 @@ def parse_src_mk(repo_path):
         if '=' in line:
             current_src = line.split('=')[0].strip()
             src_files[current_src] = []
-        elif '.cc' in line:
-            src_path = line.split('.cc')[0].strip() + '.cc'
+        elif '.c' in line:
+            src_path = line.split('\\')[0].strip()
             src_files[current_src].append(src_path)
     return src_files
 
@@ -69,27 +69,11 @@ def get_cc_files(repo_path):
     return cc_files
 
 
-# Get tests from Makefile
-def get_tests(repo_path):
+# Get parallel tests from Makefile
+def get_parallel_tests(repo_path):
     Makefile = repo_path + "/Makefile"
 
-    # Dictionary TEST_NAME => IS_PARALLEL
-    tests = {}
-
-    found_tests = False
-    for line in open(Makefile):
-        line = line.strip()
-        if line.startswith("TESTS ="):
-            found_tests = True
-        elif found_tests:
-            if line.endswith("\\"):
-                # remove the trailing \
-                line = line[:-1]
-                line = line.strip()
-                tests[line] = False
-            else:
-                # we consumed all the tests
-                break
+    s = set({})
 
     found_parallel_tests = False
     for line in open(Makefile):
@@ -101,12 +85,12 @@ def get_tests(repo_path):
                 # remove the trailing \
                 line = line[:-1]
                 line = line.strip()
-                tests[line] = True
+                s.add(line)
             else:
                 # we consumed all the parallel tests
                 break
 
-    return tests
+    return s
 
 # Parse extra dependencies passed by user from command line
 def get_dependencies():
@@ -139,10 +123,10 @@ def generate_targets(repo_path, deps_map):
     src_mk = parse_src_mk(repo_path)
     # get all .cc files
     cc_files = get_cc_files(repo_path)
-    # get tests from Makefile
-    tests = get_tests(repo_path)
+    # get parallel tests from Makefile
+    parallel_tests = get_parallel_tests(repo_path)
 
-    if src_mk is None or cc_files is None or tests is None:
+    if src_mk is None or cc_files is None or parallel_tests is None:
         return False
 
     TARGETS = TARGETSBuilder("%s/TARGETS" % repo_path)
@@ -152,6 +136,15 @@ def generate_targets(repo_path, deps_map):
         "rocksdb_lib",
         src_mk["LIB_SOURCES"] +
         src_mk["TOOL_LIB_SOURCES"])
+    # rocksdb_whole_archive_lib
+    TARGETS.add_library(
+        "rocksdb_whole_archive_lib",
+        src_mk["LIB_SOURCES"] +
+        src_mk["TOOL_LIB_SOURCES"],
+        deps=None,
+        headers=None,
+        extra_external_deps="",
+        link_whole=True)
     # rocksdb_test_lib
     TARGETS.add_library(
         "rocksdb_test_lib",
@@ -178,32 +171,43 @@ def generate_targets(repo_path, deps_map):
         + ["test_util/testutil.cc"])
 
     print("Extra dependencies:\n{0}".format(json.dumps(deps_map)))
-    # test for every test we found in the Makefile
+
+    # Dictionary test executable name -> relative source file path
+    test_source_map = {}
+    print(src_mk)
+
+    # c_test.c is added through TARGETS.add_c_test(). If there
+    # are more than one .c test file, we need to extend
+    # TARGETS.add_c_test() to include other C tests too.
+    for test_src in src_mk.get("TEST_MAIN_SOURCES_C", []):
+        if test_src != 'db/c_test.c':
+            print("Don't know how to deal with " + test_src)
+            return False
+    TARGETS.add_c_test()
+
+    for test_src in src_mk.get("TEST_MAIN_SOURCES", []):
+        test = test_src.split('.c')[0].strip().split('/')[-1].strip()
+        test_source_map[test] = test_src
+        print("" + test + " " + test_src)
+
     for target_alias, deps in deps_map.items():
-        for test in sorted(tests):
-            match_src = [src for src in cc_files if ("/%s.c" % test) in src]
-            if len(match_src) == 0:
-                print(ColorString.warning("Cannot find .cc file for %s" % test))
-                continue
-            elif len(match_src) > 1:
-                print(ColorString.warning("Found more than one .cc for %s" % test))
-                print(match_src)
+        for test, test_src in sorted(test_source_map.items()):
+            if len(test) == 0:
+                print(ColorString.warning("Failed to get test name for %s" % test_src))
                 continue
 
-            assert(len(match_src) == 1)
-            is_parallel = tests[test]
             test_target_name = \
                 test if not target_alias else test + "_" + target_alias
             TARGETS.register_test(
                 test_target_name,
-                match_src[0],
-                is_parallel,
+                test_src,
+                test in parallel_tests,
                 json.dumps(deps['extra_deps']),
                 json.dumps(deps['extra_compiler_flags']))
 
             if test in _EXPORTED_TEST_LIBS:
                 test_library = "%s_lib" % test_target_name
-                TARGETS.add_library(test_library, match_src, [":rocksdb_test_lib"])
+                TARGETS.add_library(test_library, [test_src], [":rocksdb_test_lib"])
     TARGETS.flush_tests()
 
     print(ColorString.info("Generated TARGETS Summary:"))

buckifier/targets_builder.py

@@ -27,9 +27,9 @@ def pretty_list(lst, indent=8):
 class TARGETSBuilder(object):
     def __init__(self, path):
         self.path = path
-        self.targets_file = open(path, 'w')
+        self.targets_file = open(path, 'wb')
         header = targets_cfg.rocksdb_target_header_template
-        self.targets_file.write(header)
+        self.targets_file.write(header.encode("utf-8"))
         self.total_lib = 0
         self.total_bin = 0
         self.total_test = 0
@@ -39,7 +39,7 @@ class TARGETSBuilder(object):
         self.targets_file.close()
 
     def add_library(self, name, srcs, deps=None, headers=None,
-                    extra_external_deps=""):
+                    extra_external_deps="", link_whole=False):
         headers_attr_prefix = ""
         if headers is None:
             headers_attr_prefix = "auto_"
@@ -52,7 +52,8 @@ class TARGETSBuilder(object):
             headers_attr_prefix=headers_attr_prefix,
             headers=headers,
             deps=pretty_list(deps),
-            extra_external_deps=extra_external_deps))
+            extra_external_deps=extra_external_deps,
+            link_whole=link_whole).encode("utf-8"))
         self.total_lib = self.total_lib + 1
 
     def add_rocksdb_library(self, name, srcs, headers=None):
@@ -66,16 +67,40 @@ class TARGETSBuilder(object):
             name=name,
             srcs=pretty_list(srcs),
             headers_attr_prefix=headers_attr_prefix,
-            headers=headers))
+            headers=headers).encode("utf-8"))
         self.total_lib = self.total_lib + 1
 
     def add_binary(self, name, srcs, deps=None):
-        self.targets_file.write(targets_cfg.binary_template % (
-            name,
-            pretty_list(srcs),
-            pretty_list(deps)))
+        self.targets_file.write(targets_cfg.binary_template.format(
+            name=name,
+            srcs=pretty_list(srcs),
+            deps=pretty_list(deps)).encode("utf-8"))
         self.total_bin = self.total_bin + 1
 
+    def add_c_test(self):
+        self.targets_file.write(b"""
+if not is_opt_mode:
+    cpp_binary(
+        name = "c_test_bin",
+        srcs = ["db/c_test.c"],
+        arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+        os_preprocessor_flags = ROCKSDB_OS_PREPROCESSOR_FLAGS,
+        compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+        deps = [":rocksdb_test_lib"],
+    )
+
+if not is_opt_mode:
+    custom_unittest(
+        "c_test",
+        command = [
+            native.package_name() + "/buckifier/rocks_test_runner.sh",
+            "$(location :{})".format("c_test_bin"),
+        ],
+        type = "simple",
+    )
+""")
+
     def register_test(self,
                       test_name,
                       src,
@@ -95,5 +120,5 @@ class TARGETSBuilder(object):
         self.total_test = self.total_test + 1
 
     def flush_tests(self):
-        self.targets_file.write(targets_cfg.unittests_template % self.tests_cfg)
+        self.targets_file.write(targets_cfg.unittests_template.format(tests=self.tests_cfg).encode("utf-8"))
         self.tests_cfg = ""

buckifier/targets_cfg.py

@@ -32,13 +32,16 @@ ROCKSDB_EXTERNAL_DEPS = [
     ("gflags", None, "gflags"),
     ("lz4", None, "lz4"),
     ("zstd", None),
-    ("tbb", None),
 ]
 
 ROCKSDB_OS_DEPS = [
     (
         "linux",
-        ["third-party//numa:numa", "third-party//liburing:uring"],
+        ["third-party//numa:numa", "third-party//liburing:uring", "third-party//tbb:tbb"],
+    ),
+    (
+        "macos",
+        ["third-party//tbb:tbb"],
     ),
 ]
 
@@ -56,17 +59,27 @@ ROCKSDB_OS_PREPROCESSOR_FLAGS = [
             "-DHAVE_SSE42",
             "-DLIBURING",
             "-DNUMA",
+            "-DROCKSDB_PLATFORM_POSIX",
+            "-DROCKSDB_LIB_IO_POSIX",
+            "-DTBB",
         ],
     ),
     (
         "macos",
-        ["-DOS_MACOSX"],
+        [
+            "-DOS_MACOSX",
+            "-DROCKSDB_PLATFORM_POSIX",
+            "-DROCKSDB_LIB_IO_POSIX",
+            "-DTBB",
+        ],
+    ),
+    (
+        "windows",
+        [ "-DOS_WIN", "-DWIN32", "-D_MBCS", "-DWIN64", "-DNOMINMAX" ]
     ),
 ]
 
 ROCKSDB_PREPROCESSOR_FLAGS = [
-    "-DROCKSDB_PLATFORM_POSIX",
-    "-DROCKSDB_LIB_IO_POSIX",
     "-DROCKSDB_SUPPORT_THREAD_LOCAL",
 
     # Flags to enable libs we include
@@ -77,7 +90,6 @@ ROCKSDB_PREPROCESSOR_FLAGS = [
     "-DZSTD",
     "-DZSTD_STATIC_LINKING_ONLY",
     "-DGFLAGS=gflags",
-    "-DTBB",
 
     # Added missing flags from output of build_detect_platform
     "-DROCKSDB_BACKTRACE",
@@ -134,6 +146,7 @@ cpp_library(
     preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
     deps = [{deps}],
     external_deps = ROCKSDB_EXTERNAL_DEPS{extra_external_deps},
+    link_whole = {link_whole},
 )
 """
 
@@ -154,12 +167,12 @@ cpp_library(
 
 binary_template = """
 cpp_binary(
-    name = "%s",
-    srcs = [%s],
+    name = "{name}",
+    srcs = [{srcs}],
     arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
     compiler_flags = ROCKSDB_COMPILER_FLAGS,
     preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
-    deps = [%s],
+    deps = [{deps}],
     external_deps = ROCKSDB_EXTERNAL_DEPS,
 )
 """
@@ -176,7 +189,7 @@ test_cfg_template = """    [
 unittests_template = """
 # [test_name, test_src, test_type, extra_deps, extra_compiler_flags]
 ROCKS_TESTS = [
-%s]
+{tests}]
 
 # Generate a test rule for each entry in ROCKS_TESTS
 # Do not build the tests in opt mode, since SyncPoint and other test code
@@ -187,7 +200,7 @@ ROCKS_TESTS = [
         srcs = [test_cc],
         arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
         os_preprocessor_flags = ROCKSDB_OS_PREPROCESSOR_FLAGS,
-        compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        compiler_flags = ROCKSDB_COMPILER_FLAGS + extra_compiler_flags,
         preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
         deps = [":rocksdb_test_lib"] + extra_deps,
         external_deps = ROCKSDB_EXTERNAL_DEPS + [

build_tools/build_detect_platform

@@ -60,8 +60,16 @@ if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
       source "$PWD/build_tools/fbcode_config4.8.1.sh"
     elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_5xx" ]; then
       source "$PWD/build_tools/fbcode_config.sh"
-    else
+    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_PLATFORM007" ]; then
       source "$PWD/build_tools/fbcode_config_platform007.sh"
+    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_PLATFORM009" ]; then
+      source "$PWD/build_tools/fbcode_config_platform009.sh"
+    elif [ -z "$USE_CLANG" ]; then
+        # Still use platform007 for gcc by default for build break on
+        # some hosts.
+      source "$PWD/build_tools/fbcode_config_platform007.sh"
+    else
+      source "$PWD/build_tools/fbcode_config_platform009.sh"
     fi
 fi
 
@@ -89,6 +97,16 @@ if test -z "$CXX"; then
     fi
 fi
 
+if test -z "$AR"; then
+    if [ -x "$(command -v gcc-ar)" ]; then
+        AR=gcc-ar
+    elif [ -x "$(command -v llvm-ar)" ]; then
+        AR=llvm-ar
+    else
+        AR=ar
+    fi
+fi
+
 # Detect OS
 if test -z "$TARGET_OS"; then
     TARGET_OS=`uname -s`
@@ -151,7 +169,7 @@ case "$TARGET_OS" in
         else
             PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -latomic"
         fi
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread -lrt"
+        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread -lrt -ldl"
         if test $ROCKSDB_USE_IO_URING; then
             # check for liburing
             $CXX $CFLAGS -x c++ - -luring -o /dev/null 2>/dev/null  <<EOF
@@ -296,24 +314,32 @@ EOF
         # Test whether gflags library is installed
         # http://gflags.github.io/gflags/
         # check if the namespace is gflags
-        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
+        if $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
           #include <gflags/gflags.h>
+          using namespace GFLAGS_NAMESPACE;
           int main() {}
 EOF
-        if [ "$?" = 0 ]; then
-            COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=1"
-            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
-        else
-          # check if namespace is google
-          $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
+        then
+          COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=1"
+          PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
+        # check if namespace is gflags
+        elif $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
+            #include <gflags/gflags.h>
+            using namespace gflags;
+            int main() {}
+EOF
+        then
+          COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=1 -DGFLAGS_NAMESPACE=gflags"
+          PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
+        # check if namespace is google
+        elif $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
             #include <gflags/gflags.h>
             using namespace google;
             int main() {}
 EOF
-          if [ "$?" = 0 ]; then
-              COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=google"
-              PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
-          fi
+        then
+          COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=1 -DGFLAGS_NAMESPACE=google"
+          PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
         fi
     fi
 
@@ -441,7 +467,7 @@ EOF
             COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_MALLOC_USABLE_SIZE"
         fi
     fi
-    
+
     if ! test $ROCKSDB_DISABLE_MEMKIND; then
         # Test whether memkind library is installed
         $CXX $CFLAGS $COMMON_FLAGS -lmemkind -x c++ - -o /dev/null 2>/dev/null  <<EOF
@@ -457,7 +483,7 @@ EOF
             JAVA_LDFLAGS="$JAVA_LDFLAGS -lmemkind"
         fi
     fi
-    
+
     if ! test $ROCKSDB_DISABLE_PTHREAD_MUTEX_ADAPTIVE_NP; then
         # Test whether PTHREAD_MUTEX_ADAPTIVE_NP mutex type is available
         $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
@@ -623,6 +649,13 @@ else
   if test "$USE_SSE"; then
     TRY_SSE_ETC="1"
   fi
+
+  if [[ "${PLATFORM}" == "OS_MACOSX" ]]; then
+    # For portability compile for macOS 10.12 (2016) or newer
+    COMMON_FLAGS="$COMMON_FLAGS -mmacosx-version-min=10.12"
+    PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -mmacosx-version-min=10.12"
+    PLATFORM_SHARED_LDFLAGS="$PLATFORM_SHARED_LDFLAGS -mmacosx-version-min=10.12"
+  fi
 fi
 
 if test "$TRY_SSE_ETC"; then
@@ -632,11 +665,18 @@ if test "$TRY_SSE_ETC"; then
   # It doesn't even really check that your current CPU is compatible.
   #
   # SSE4.2 available since nehalem, ca. 2008-2010
+  # Includes POPCNT for BitsSetToOne, BitParity
   TRY_SSE42="-msse4.2"
   # PCLMUL available since westmere, ca. 2010-2011
   TRY_PCLMUL="-mpclmul"
   # AVX2 available since haswell, ca. 2013-2015
   TRY_AVX2="-mavx2"
+  # BMI available since haswell, ca. 2013-2015
+  # Primarily for TZCNT for CountTrailingZeroBits
+  TRY_BMI="-mbmi"
+  # LZCNT available since haswell, ca. 2013-2015
+  # For FloorLog2
+  TRY_LZCNT="-mlzcnt"
 fi
 
 $CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_SSE42 -x c++ - -o /dev/null 2>/dev/null <<EOF
@@ -685,6 +725,34 @@ elif test "$USE_SSE"; then
   echo "warning: USE_SSE specified but compiler could not use AVX2 intrinsics, disabling" >&2
 fi
 
+$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_BMI -x c++ - -o /dev/null 2>/dev/null <<EOF
+  #include <cstdint>
+  #include <immintrin.h>
+  int main(int argc, char *argv[]) {
+    (void)argv;
+    return (int)_tzcnt_u64((uint64_t)argc);
+  }
+EOF
+if [ "$?" = 0 ]; then
+  COMMON_FLAGS="$COMMON_FLAGS $TRY_BMI -DHAVE_BMI"
+elif test "$USE_SSE"; then
+  echo "warning: USE_SSE specified but compiler could not use BMI intrinsics, disabling" >&2
+fi
+
+$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_LZCNT -x c++ - -o /dev/null 2>/dev/null <<EOF
+  #include <cstdint>
+  #include <immintrin.h>
+  int main(int argc, char *argv[]) {
+    (void)argv;
+    return (int)_lzcnt_u64((uint64_t)argc);
+  }
+EOF
+if [ "$?" = 0 ]; then
+  COMMON_FLAGS="$COMMON_FLAGS $TRY_LZCNT -DHAVE_LZCNT"
+elif test "$USE_SSE"; then
+  echo "warning: USE_SSE specified but compiler could not use LZCNT intrinsics, disabling" >&2
+fi
+
 $CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
   #include <cstdint>
   int main() {
@@ -741,6 +809,7 @@ ROCKSDB_PATCH=`build_tools/version.sh patch`
 
 echo "CC=$CC" >> "$OUTPUT"
 echo "CXX=$CXX" >> "$OUTPUT"
+echo "AR=$AR" >> "$OUTPUT"
 echo "PLATFORM=$PLATFORM" >> "$OUTPUT"
 echo "PLATFORM_LDFLAGS=$PLATFORM_LDFLAGS" >> "$OUTPUT"
 echo "JAVA_LDFLAGS=$JAVA_LDFLAGS" >> "$OUTPUT"

build_tools/dependencies_platform009.sh

@@ -0,0 +1,20 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+GCC_BASE=/mnt/gvfs/third-party2/gcc/1795efe5f06778c15a92c8f9a2aba5dc496d9d4d/9.x/centos7-native/3bed279
+CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/7318eaac22659b6ff2fe43918e4b69fd0772a8a7/9.0.0/platform009/651ee30
+LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/4959b39cfbe5965a37c861c4c327fa7c5c759b87/9.x/platform009/9202ce7
+GLIBC_BASE=/mnt/gvfs/third-party2/glibc/45ce3375cdc77ecb2520bbf8f0ecddd3f98efd7a/2.30/platform009/f259413
+SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/be4de3205e029101b18aa8103daa696c2bef3b19/1.1.3/platform009/7f3b187
+ZLIB_BASE=/mnt/gvfs/third-party2/zlib/3c160ac5c67e257501e24c6c1d00ad5e01d73db6/1.2.8/platform009/7f3b187
+BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/73a237ac5bc0a5f5d67b39b8d253cfebaab88684/1.0.6/platform009/7f3b187
+LZ4_BASE=/mnt/gvfs/third-party2/lz4/ec6573523b0ce55ef6373a4801189027cf07bb2c/1.9.1/platform009/7f3b187
+ZSTD_BASE=/mnt/gvfs/third-party2/zstd/64c58a207d2495e83abc57a500a956df09b79a7c/1.4.x/platform009/ba86d1f
+GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/824d0a8a5abb5b121afd1b35fc3896407ea50092/2.2.0/platform009/7f3b187
+JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/d9aef9feb850b168a68736420f217b01cce11a89/master/platform009/c305944
+NUMA_BASE=/mnt/gvfs/third-party2/numa/0af65f71e23a67bf65dc91b11f95caa39325c432/2.0.11/platform009/7f3b187
+LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/02486dac347645d31dce116f44e1de3177315be2/1.4/platform009/5191652
+TBB_BASE=/mnt/gvfs/third-party2/tbb/2e0ec671e550bfca347300bf3f789d9c0fff24ad/2018_U5/platform009/7f3b187
+LIBURING_BASE=/mnt/gvfs/third-party2/liburing/70dbd9cfee63a25611417d09433a86d7711b3990/20200729/platform009/7f3b187
+KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/32b8a2407b634df3f8f948ba373fc4acc6a18296/fb/platform009/da39a3e
+BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/08634589372fa5f237bfd374e8c644a8364e78c1/2.32/platform009/ba86d1f/
+VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/6ae525939ad02e5e676855082fbbc7828dbafeac/3.15.0/platform009/7f3b187
+LUA_BASE=/mnt/gvfs/third-party2/lua/162efd9561a3d21f6869f4814011e9cf1b3ff4dc/5.3.4/platform009/a6271c4

build_tools/fbcode_config.sh

@@ -109,6 +109,7 @@ if [ -z "$USE_CLANG" ]; then
   # gcc
   CC="$GCC_BASE/bin/gcc"
   CXX="$GCC_BASE/bin/g++"
+  AR="$GCC_BASE/bin/gcc-ar"
 
   CFLAGS+=" -B$BINUTILS/gold"
   CFLAGS+=" -isystem $GLIBC_INCLUDE"
@@ -119,6 +120,7 @@ else
   CLANG_INCLUDE="$CLANG_LIB/clang/stable/include"
   CC="$CLANG_BIN/clang"
   CXX="$CLANG_BIN/clang++"
+  AR="$CLANG_BIN/llvm-ar"
 
   KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
 

build_tools/fbcode_config4.8.1.sh

@@ -69,6 +69,7 @@ if [ -z "$USE_CLANG" ]; then
   # gcc
   CC="$GCC_BASE/bin/gcc"
   CXX="$GCC_BASE/bin/g++"
+  CXX="$GCC_BASE/bin/gcc-ar"
 
   CFLAGS="-B$BINUTILS/gold -m64 -mtune=generic"
   CFLAGS+=" -isystem $GLIBC_INCLUDE"
@@ -81,6 +82,7 @@ else
   CLANG_INCLUDE="$CLANG_LIB/clang/*/include"
   CC="$CLANG_BIN/clang"
   CXX="$CLANG_BIN/clang++"
+  AR="$CLANG_BIN/llvm-ar"
 
   KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include/"
 

build_tools/fbcode_config_platform007.sh

@@ -118,6 +118,7 @@ if [ -z "$USE_CLANG" ]; then
   # gcc
   CC="$GCC_BASE/bin/gcc"
   CXX="$GCC_BASE/bin/g++"
+  AR="$GCC_BASE/bin/gcc-ar"
 
   CFLAGS+=" -B$BINUTILS/gold"
   CFLAGS+=" -isystem $LIBGCC_INCLUDE"
@@ -128,6 +129,7 @@ else
   CLANG_INCLUDE="$CLANG_LIB/clang/stable/include"
   CC="$CLANG_BIN/clang"
   CXX="$CLANG_BIN/clang++"
+  AR="$CLANG_BIN/llvm-ar"
 
   KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
 

build_tools/fbcode_config_platform009.sh

@@ -0,0 +1,172 @@
+#!/bin/sh
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+#
+# Set environment variables so that we can compile rocksdb using
+# fbcode settings.  It uses the latest g++ and clang compilers and also
+# uses jemalloc
+# Environment variables that change the behavior of this script:
+# PIC_BUILD -- if true, it will only take pic versions of libraries from fbcode. libraries that don't have pic variant will not be included
+
+
+BASEDIR=`dirname $BASH_SOURCE`
+source "$BASEDIR/dependencies_platform009.sh"
+
+CFLAGS=""
+
+# libgcc
+LIBGCC_INCLUDE="$LIBGCC_BASE/include/c++/9.3.0"
+LIBGCC_LIBS=" -L $LIBGCC_BASE/lib"
+
+# glibc
+GLIBC_INCLUDE="$GLIBC_BASE/include"
+GLIBC_LIBS=" -L $GLIBC_BASE/lib"
+
+# snappy
+SNAPPY_INCLUDE=" -I $SNAPPY_BASE/include/"
+if test -z $PIC_BUILD; then
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy.a"
+else
+  SNAPPY_LIBS=" $SNAPPY_BASE/lib/libsnappy_pic.a"
+fi
+CFLAGS+=" -DSNAPPY"
+
+if test -z $PIC_BUILD; then
+  # location of zlib headers and libraries
+  ZLIB_INCLUDE=" -I $ZLIB_BASE/include/"
+  ZLIB_LIBS=" $ZLIB_BASE/lib/libz.a"
+  CFLAGS+=" -DZLIB"
+
+  # location of bzip headers and libraries
+  BZIP_INCLUDE=" -I $BZIP2_BASE/include/"
+  BZIP_LIBS=" $BZIP2_BASE/lib/libbz2.a"
+  CFLAGS+=" -DBZIP2"
+
+  LZ4_INCLUDE=" -I $LZ4_BASE/include/"
+  LZ4_LIBS=" $LZ4_BASE/lib/liblz4.a"
+  CFLAGS+=" -DLZ4"
+fi
+
+ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
+if test -z $PIC_BUILD; then
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
+else
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
+fi
+CFLAGS+=" -DZSTD"
+
+# location of gflags headers and libraries
+GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
+if test -z $PIC_BUILD; then
+  GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags.a"
+else
+  GFLAGS_LIBS=" $GFLAGS_BASE/lib/libgflags_pic.a"
+fi
+CFLAGS+=" -DGFLAGS=gflags"
+
+# location of jemalloc
+JEMALLOC_INCLUDE=" -I $JEMALLOC_BASE/include/"
+JEMALLOC_LIB=" $JEMALLOC_BASE/lib/libjemalloc.a"
+
+if test -z $PIC_BUILD; then
+  # location of numa
+  NUMA_INCLUDE=" -I $NUMA_BASE/include/"
+  NUMA_LIB=" $NUMA_BASE/lib/libnuma.a"
+  CFLAGS+=" -DNUMA"
+
+  # location of libunwind
+  LIBUNWIND="$LIBUNWIND_BASE/lib/libunwind.a"
+fi
+
+# location of TBB
+TBB_INCLUDE=" -isystem $TBB_BASE/include/"
+if test -z $PIC_BUILD; then
+  TBB_LIBS="$TBB_BASE/lib/libtbb.a"
+else
+  TBB_LIBS="$TBB_BASE/lib/libtbb_pic.a"
+fi
+CFLAGS+=" -DTBB"
+
+# location of LIBURING
+LIBURING_INCLUDE=" -isystem $LIBURING_BASE/include/"
+if test -z $PIC_BUILD; then
+  LIBURING_LIBS="$LIBURING_BASE/lib/liburing.a"
+else
+  LIBURING_LIBS="$LIBURING_BASE/lib/liburing_pic.a"
+fi
+CFLAGS+=" -DLIBURING"
+
+test "$USE_SSE" || USE_SSE=1
+export USE_SSE
+test "$PORTABLE" || PORTABLE=1
+export PORTABLE
+
+BINUTILS="$BINUTILS_BASE/bin"
+AR="$BINUTILS/ar"
+
+DEPS_INCLUDE="$SNAPPY_INCLUDE $ZLIB_INCLUDE $BZIP_INCLUDE $LZ4_INCLUDE $ZSTD_INCLUDE $GFLAGS_INCLUDE $NUMA_INCLUDE $TBB_INCLUDE $LIBURING_INCLUDE"
+
+STDLIBS="-L $GCC_BASE/lib64"
+
+CLANG_BIN="$CLANG_BASE/bin"
+CLANG_LIB="$CLANG_BASE/lib"
+CLANG_SRC="$CLANG_BASE/../../src"
+
+CLANG_ANALYZER="$CLANG_BIN/clang++"
+CLANG_SCAN_BUILD="$CLANG_SRC/llvm/clang/tools/scan-build/bin/scan-build"
+
+if [ -z "$USE_CLANG" ]; then
+  # gcc
+  CC="$GCC_BASE/bin/gcc"
+  CXX="$GCC_BASE/bin/g++"
+  AR="$GCC_BASE/bin/gcc-ar"
+
+  CFLAGS+=" -B$BINUTILS/gold"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
+  CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  JEMALLOC=1
+else
+  # clang
+  CLANG_INCLUDE="$CLANG_LIB/clang/stable/include"
+  CC="$CLANG_BIN/clang"
+  CXX="$CLANG_BIN/clang++"
+  AR="$CLANG_BIN/llvm-ar"
+
+  KERNEL_HEADERS_INCLUDE="$KERNEL_HEADERS_BASE/include"
+
+  CFLAGS+=" -B$BINUTILS/gold -nostdinc -nostdlib"
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/9.x "
+  CFLAGS+=" -isystem $LIBGCC_BASE/include/c++/9.x/x86_64-facebook-linux "
+  CFLAGS+=" -isystem $GLIBC_INCLUDE"
+  CFLAGS+=" -isystem $LIBGCC_INCLUDE"
+  CFLAGS+=" -isystem $CLANG_INCLUDE"
+  CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE/linux "
+  CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE "
+  CFLAGS+=" -Wno-expansion-to-defined "
+  CXXFLAGS="-nostdinc++"
+fi
+
+CFLAGS+=" $DEPS_INCLUDE"
+CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DROCKSDB_SUPPORT_THREAD_LOCAL -DHAVE_SSE42 -DROCKSDB_IOURING_PRESENT"
+CXXFLAGS+=" $CFLAGS"
+
+EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS $LIBURING_LIBS"
+EXEC_LDFLAGS+=" -B$BINUTILS/gold"
+EXEC_LDFLAGS+=" -Wl,--dynamic-linker,/usr/local/fbcode/platform009/lib/ld.so"
+EXEC_LDFLAGS+=" $LIBUNWIND"
+EXEC_LDFLAGS+=" -Wl,-rpath=/usr/local/fbcode/platform009/lib"
+# required by libtbb
+EXEC_LDFLAGS+=" -ldl"
+
+PLATFORM_LDFLAGS="$LIBGCC_LIBS $GLIBC_LIBS $STDLIBS -lgcc -lstdc++"
+
+EXEC_LDFLAGS_SHARED="$SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $TBB_LIBS $LIBURING_LIBS"
+
+VALGRIND_VER="$VALGRIND_BASE/bin/"
+
+# lua not supported because it's on track for deprecation, I think
+LUA_PATH=
+LUA_LIB=
+
+USE_FOLLY_DISTRIBUTED_MUTEX=1
+
+export CC CXX AR CFLAGS CXXFLAGS EXEC_LDFLAGS EXEC_LDFLAGS_SHARED VALGRIND_VER JEMALLOC_LIB JEMALLOC_INCLUDE CLANG_ANALYZER CLANG_SCAN_BUILD LUA_PATH LUA_LIB

build_tools/make_package.sh

@@ -103,31 +103,26 @@ function main() {
   gem_install fpm
 
   make static_lib
-  make install INSTALL_PATH=package
-
-  cd package
-
-  LIB_DIR=lib
-  if [[ -z "$ARCH" ]]; then
-      ARCH=$(getconf LONG_BIT)
-  fi
-  if [[ ("$FPM_OUTPUT" = "rpm") && ($ARCH -eq 64) ]]; then
-      mv lib lib64
-      LIB_DIR=lib64
+  LIBDIR=/usr/lib
+  if [[ $FPM_OUTPUT = "rpm" ]]; then
+      LIBDIR=$(rpm --eval '%_libdir')
   fi
 
+  rm -rf package
+  make install DESTDIR=package PREFIX=/usr LIBDIR=$LIBDIR
+
   fpm \
     -s dir \
     -t $FPM_OUTPUT \
+    -C package \
     -n rocksdb \
     -v $1 \
-    --prefix /usr \
     --url http://rocksdb.org/ \
     -m rocksdb@fb.com \
     --license BSD \
     --vendor Facebook \
     --description "RocksDB is an embeddable persistent key-value store for fast storage." \
-    include $LIB_DIR
+    usr
 }
 
 # shellcheck disable=SC2068

build_tools/rocksdb-lego-determinator

@@ -421,6 +421,66 @@ STRESS_CRASH_TEST_COMMANDS="[
     }
 ]"
 
+#
+# RocksDB blackbox stress/crash test
+#
+BLACKBOX_STRESS_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb Blackbox Stress and Crash Test',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug stress tests',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 db_stress || $CONTRUN_NAME=db_stress $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            {
+                'name':'Build and run RocksDB debug blackbox crash tests',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 blackbox_crash_test || $CONTRUN_NAME=blackbox_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
+#
+# RocksDB whitebox stress/crash test
+#
+WHITEBOX_STRESS_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb Whitebox Stress and Crash Test',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug stress tests',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 db_stress || $CONTRUN_NAME=db_stress $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            {
+                'name':'Build and run RocksDB debug whitebox crash tests',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 whitebox_crash_test || $CONTRUN_NAME=whitebox_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB stress/crash test with atomic flush
 #
@@ -549,6 +609,54 @@ ASAN_CRASH_TEST_COMMANDS="[
     }
 ]"
 
+#
+# RocksDB blackbox crash testing under address sanitizer
+#
+ASAN_BLACKBOX_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb blackbox crash test under ASAN',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug blackbox asan_crash_test',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 blackbox_asan_crash_test || $CONTRUN_NAME=blackbox_asan_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
+#
+# RocksDB whitebox crash testing under address sanitizer
+#
+ASAN_WHITEBOX_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb whitebox crash test under ASAN',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug whitebox asan_crash_test',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 whitebox_asan_crash_test || $CONTRUN_NAME=whitebox_asan_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB crash testing with atomic flush under address sanitizer
 #
@@ -619,7 +727,7 @@ UBSAN_TEST_COMMANDS="[
 ]"
 
 #
-# RocksDB crash testing under udnefined behavior sanitizer
+# RocksDB crash testing under undefined behavior sanitizer
 #
 UBSAN_CRASH_TEST_COMMANDS="[
     {
@@ -642,6 +750,54 @@ UBSAN_CRASH_TEST_COMMANDS="[
     }
 ]"
 
+#
+# RocksDB crash testing under undefined behavior sanitizer
+#
+UBSAN_BLACKBOX_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb blackbox crash test under UBSAN',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug blackbox ubsan_crash_test',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH $CLANG make J=1 blackbox_ubsan_crash_test || $CONTRUN_NAME=blackbox_ubsan_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
+#
+# RocksDB crash testing under undefined behavior sanitizer
+#
+UBSAN_WHITEBOX_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb whitebox crash test under UBSAN',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Build and run RocksDB debug whitebox ubsan_crash_test',
+                'timeout': 86400,
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH $CLANG make J=1 whitebox_ubsan_crash_test || $CONTRUN_NAME=whitebox_ubsan_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB crash testing with atomic flush under undefined behavior sanitizer
 #
@@ -760,6 +916,54 @@ TSAN_CRASH_TEST_COMMANDS="[
     }
 ]"
 
+#
+# RocksDB blackbox crash test under TSAN
+#
+TSAN_BLACKBOX_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb Blackbox Crash Test under TSAN',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Compile and run',
+                'timeout': 86400,
+                'shell':'set -o pipefail && $SHM $DEBUG $TSAN $TSAN_CRASH CRASH_TEST_KILL_ODD=1887 make J=1 blackbox_crash_test || $CONTRUN_NAME=tsan_blackbox_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
+#
+# RocksDB whitebox crash test under TSAN
+#
+TSAN_WHITEBOX_CRASH_TEST_COMMANDS="[
+    {
+        'name':'Rocksdb Whitebox Crash Test under TSAN',
+        'oncall':'$ONCALL',
+        'executeLocal': 'true',
+        'timeout': 86400,
+        'steps': [
+            $CLEANUP_ENV,
+            {
+                'name':'Compile and run',
+                'timeout': 86400,
+                'shell':'set -o pipefail && $SHM $DEBUG $TSAN $TSAN_CRASH CRASH_TEST_KILL_ODD=1887 make J=1 whitebox_crash_test || $CONTRUN_NAME=tsan_whitebox_crash_test $TASK_CREATION_TOOL',
+                'user':'root',
+                $PARSER
+            },
+            $UPLOAD_DB_DIR,
+        ],
+        $REPORT
+    }
+]"
+
 #
 # RocksDB crash test with atomic flush under TSAN
 #
@@ -897,6 +1101,7 @@ run_regression()
   strip librocksdb.a
   send_size_to_ods static_lib_stripped $(stat --printf="%s" librocksdb.a)
 
+  make clean
   make -j$(nproc) shared_lib
   send_size_to_ods shared_lib $(stat --printf="%s" `readlink -f librocksdb.so`)
   strip `readlink -f librocksdb.so`
@@ -909,6 +1114,7 @@ run_regression()
   strip librocksdb.a
   send_size_to_ods static_lib_lite_stripped $(stat --printf="%s" librocksdb.a)
 
+  make clean
   make LITE=1 -j$(nproc) shared_lib
   send_size_to_ods shared_lib_lite $(stat --printf="%s" `readlink -f librocksdb.so`)
   strip `readlink -f librocksdb.so`
@@ -994,6 +1200,12 @@ case $1 in
   stress_crash)
     echo $STRESS_CRASH_TEST_COMMANDS
     ;;
+  blackbox_stress_crash)
+    echo $BLACKBOX_STRESS_CRASH_TEST_COMMANDS
+    ;;
+  whitebox_stress_crash)
+    echo $WHITEBOX_STRESS_CRASH_TEST_COMMANDS
+    ;;
   stress_crash_with_atomic_flush)
     echo $STRESS_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
     ;;
@@ -1009,6 +1221,12 @@ case $1 in
   asan_crash)
     echo $ASAN_CRASH_TEST_COMMANDS
     ;;
+  blackbox_asan_crash)
+    echo $ASAN_BLACKBOX_CRASH_TEST_COMMANDS
+    ;;
+  whitebox_asan_crash)
+    echo $ASAN_WHITEBOX_CRASH_TEST_COMMANDS
+    ;;
   asan_crash_with_atomic_flush)
     echo $ASAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
     ;;
@@ -1021,6 +1239,12 @@ case $1 in
   ubsan_crash)
     echo $UBSAN_CRASH_TEST_COMMANDS
     ;;
+  blackbox_ubsan_crash)
+    echo $UBSAN_BLACKBOX_CRASH_TEST_COMMANDS
+    ;;
+  whitebox_ubsan_crash)
+    echo $UBSAN_WHITEBOX_CRASH_TEST_COMMANDS
+    ;;
   ubsan_crash_with_atomic_flush)
     echo $UBSAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
     ;;
@@ -1036,6 +1260,12 @@ case $1 in
   tsan_crash)
     echo $TSAN_CRASH_TEST_COMMANDS
     ;;
+  blackbox_tsan_crash)
+    echo $TSAN_BLACKBOX_CRASH_TEST_COMMANDS
+    ;;
+  whitebox_tsan_crash)
+    echo $TSAN_WHITEBOX_CRASH_TEST_COMMANDS
+    ;;
   tsan_crash_with_atomic_flush)
     echo $TSAN_CRASH_TEST_WITH_ATOMIC_FLUSH_COMMANDS
     ;;

cache/cache.cc

@@ -10,7 +10,7 @@
 #include "rocksdb/cache.h"
 
 #include "cache/lru_cache.h"
-#include "options/options_helper.h"
+#include "rocksdb/utilities/options_type.h"
 #include "util/string_util.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -19,22 +19,19 @@ static std::unordered_map<std::string, OptionTypeInfo>
     lru_cache_options_type_info = {
         {"capacity",
          {offsetof(struct LRUCacheOptions, capacity), OptionType::kSizeT,
-          OptionVerificationType::kNormal, OptionTypeFlags::kMutable,
-          offsetof(struct LRUCacheOptions, capacity)}},
+          OptionVerificationType::kNormal, OptionTypeFlags::kMutable}},
         {"num_shard_bits",
          {offsetof(struct LRUCacheOptions, num_shard_bits), OptionType::kInt,
-          OptionVerificationType::kNormal, OptionTypeFlags::kMutable,
-          offsetof(struct LRUCacheOptions, num_shard_bits)}},
+          OptionVerificationType::kNormal, OptionTypeFlags::kMutable}},
         {"strict_capacity_limit",
          {offsetof(struct LRUCacheOptions, strict_capacity_limit),
           OptionType::kBoolean, OptionVerificationType::kNormal,
-          OptionTypeFlags::kMutable,
-          offsetof(struct LRUCacheOptions, strict_capacity_limit)}},
+          OptionTypeFlags::kMutable}},
         {"high_pri_pool_ratio",
          {offsetof(struct LRUCacheOptions, high_pri_pool_ratio),
           OptionType::kDouble, OptionVerificationType::kNormal,
-          OptionTypeFlags::kMutable,
-          offsetof(struct LRUCacheOptions, high_pri_pool_ratio)}}};
+          OptionTypeFlags::kMutable}},
+};
 #endif  // ROCKSDB_LITE
 
 Status Cache::CreateFromString(const ConfigOptions& config_options,

cache/cache_bench.cc

@@ -136,7 +136,7 @@ struct KeyGen {
     for (uint32_t i = 0; i < FLAGS_skew; ++i) {
       raw = std::min(raw, rnd.Next());
     }
-    uint64_t key = fastrange64(raw, max_key);
+    uint64_t key = FastRange64(raw, max_key);
     // Variable size and alignment
     size_t off = key % 8;
     key_data[0] = char{42};

cache/cache_test.cc

@@ -117,8 +117,8 @@ class CacheTest : public testing::TestWithParam<std::string> {
 
   void Insert(std::shared_ptr<Cache> cache, int key, int value,
               int charge = 1) {
-    cache->Insert(EncodeKey(key), EncodeValue(value), charge,
-                  &CacheTest::Deleter);
+    EXPECT_OK(cache->Insert(EncodeKey(key), EncodeValue(value), charge,
+                            &CacheTest::Deleter));
   }
 
   void Erase(std::shared_ptr<Cache> cache, int key) {
@@ -167,9 +167,10 @@ TEST_P(CacheTest, UsageTest) {
   for (int i = 1; i < 100; ++i) {
     std::string key(i, 'a');
     auto kv_size = key.size() + 5;
-    cache->Insert(key, reinterpret_cast<void*>(value), kv_size, dumbDeleter);
-    precise_cache->Insert(key, reinterpret_cast<void*>(value), kv_size,
-                          dumbDeleter);
+    ASSERT_OK(cache->Insert(key, reinterpret_cast<void*>(value), kv_size,
+                            dumbDeleter));
+    ASSERT_OK(precise_cache->Insert(key, reinterpret_cast<void*>(value),
+                                    kv_size, dumbDeleter));
     usage += kv_size;
     ASSERT_EQ(usage, cache->GetUsage());
     ASSERT_LT(usage, precise_cache->GetUsage());
@@ -183,10 +184,10 @@ TEST_P(CacheTest, UsageTest) {
   // make sure the cache will be overloaded
   for (uint64_t i = 1; i < kCapacity; ++i) {
     auto key = ToString(i);
-    cache->Insert(key, reinterpret_cast<void*>(value), key.size() + 5,
-                  dumbDeleter);
-    precise_cache->Insert(key, reinterpret_cast<void*>(value), key.size() + 5,
-                          dumbDeleter);
+    ASSERT_OK(cache->Insert(key, reinterpret_cast<void*>(value), key.size() + 5,
+                            dumbDeleter));
+    ASSERT_OK(precise_cache->Insert(key, reinterpret_cast<void*>(value),
+                                    key.size() + 5, dumbDeleter));
   }
 
   // the usage should be close to the capacity
@@ -215,11 +216,12 @@ TEST_P(CacheTest, PinnedUsageTest) {
     auto kv_size = key.size() + 5;
     Cache::Handle* handle;
     Cache::Handle* handle_in_precise_cache;
-    cache->Insert(key, reinterpret_cast<void*>(value), kv_size, dumbDeleter,
-                  &handle);
+    ASSERT_OK(cache->Insert(key, reinterpret_cast<void*>(value), kv_size,
+                            dumbDeleter, &handle));
     assert(handle);
-    precise_cache->Insert(key, reinterpret_cast<void*>(value), kv_size,
-                          dumbDeleter, &handle_in_precise_cache);
+    ASSERT_OK(precise_cache->Insert(key, reinterpret_cast<void*>(value),
+                                    kv_size, dumbDeleter,
+                                    &handle_in_precise_cache));
     assert(handle_in_precise_cache);
     pinned_usage += kv_size;
     ASSERT_EQ(pinned_usage, cache->GetPinnedUsage());
@@ -254,10 +256,10 @@ TEST_P(CacheTest, PinnedUsageTest) {
   // check that overloading the cache does not change the pinned usage
   for (uint64_t i = 1; i < 2 * kCapacity; ++i) {
     auto key = ToString(i);
-    cache->Insert(key, reinterpret_cast<void*>(value), key.size() + 5,
-                  dumbDeleter);
-    precise_cache->Insert(key, reinterpret_cast<void*>(value), key.size() + 5,
-                          dumbDeleter);
+    ASSERT_OK(cache->Insert(key, reinterpret_cast<void*>(value), key.size() + 5,
+                            dumbDeleter));
+    ASSERT_OK(precise_cache->Insert(key, reinterpret_cast<void*>(value),
+                                    key.size() + 5, dumbDeleter));
   }
   ASSERT_EQ(pinned_usage, cache->GetPinnedUsage());
   ASSERT_EQ(precise_cache_pinned_usage, precise_cache->GetPinnedUsage());

cache/lru_cache_test.cc

@@ -37,8 +37,9 @@ class LRUCacheTest : public testing::Test {
 
   void Insert(const std::string& key,
               Cache::Priority priority = Cache::Priority::LOW) {
-    cache_->Insert(key, 0 /*hash*/, nullptr /*value*/, 1 /*charge*/,
-                   nullptr /*deleter*/, nullptr /*handle*/, priority);
+    EXPECT_OK(cache_->Insert(key, 0 /*hash*/, nullptr /*value*/, 1 /*charge*/,
+                             nullptr /*deleter*/, nullptr /*handle*/,
+                             priority));
   }
 
   void Insert(char key, Cache::Priority priority = Cache::Priority::LOW) {

db/arena_wrapped_db_iter.cc

@@ -45,6 +45,7 @@ void ArenaWrappedDBIter::Init(Env* env, const ReadOptions& read_options,
                               true, max_sequential_skip_in_iteration,
                               read_callback, db_impl, cfd, allow_blob);
   sv_number_ = version_number;
+  read_options_ = read_options;
   allow_refresh_ = allow_refresh;
 }
 
@@ -98,8 +99,7 @@ ArenaWrappedDBIter* NewArenaWrappedDbIterator(
              max_sequential_skip_in_iterations, version_number, read_callback,
              db_impl, cfd, allow_blob, allow_refresh);
   if (db_impl != nullptr && cfd != nullptr && allow_refresh) {
-    iter->StoreRefreshInfo(read_options, db_impl, cfd, read_callback,
-                           allow_blob);
+    iter->StoreRefreshInfo(db_impl, cfd, read_callback, allow_blob);
   }
 
   return iter;

db/arena_wrapped_db_iter.h

@@ -41,6 +41,7 @@ class ArenaWrappedDBIter : public Iterator {
   virtual ReadRangeDelAggregator* GetRangeDelAggregator() {
     return db_iter_->GetRangeDelAggregator();
   }
+  const ReadOptions& GetReadOptions() { return read_options_; }
 
   // Set the internal iterator wrapped inside the DB Iterator. Usually it is
   // a merging iterator.
@@ -79,10 +80,8 @@ class ArenaWrappedDBIter : public Iterator {
 
   // Store some parameters so we can refresh the iterator at a later point
   // with these same params
-  void StoreRefreshInfo(const ReadOptions& read_options, DBImpl* db_impl,
-                        ColumnFamilyData* cfd, ReadCallback* read_callback,
-                        bool allow_blob) {
-    read_options_ = read_options;
+  void StoreRefreshInfo(DBImpl* db_impl, ColumnFamilyData* cfd,
+                        ReadCallback* read_callback, bool allow_blob) {
     db_impl_ = db_impl;
     cfd_ = cfd;
     read_callback_ = read_callback;

db/blob/blob_file_builder.cc

@@ -0,0 +1,316 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include "db/blob/blob_file_builder.h"
+
+#include <cassert>
+
+#include "db/blob/blob_file_addition.h"
+#include "db/blob/blob_index.h"
+#include "db/blob/blob_log_format.h"
+#include "db/blob/blob_log_writer.h"
+#include "db/version_set.h"
+#include "file/filename.h"
+#include "file/read_write_util.h"
+#include "file/writable_file_writer.h"
+#include "logging/logging.h"
+#include "options/cf_options.h"
+#include "rocksdb/slice.h"
+#include "rocksdb/status.h"
+#include "test_util/sync_point.h"
+#include "util/compression.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+BlobFileBuilder::BlobFileBuilder(
+    VersionSet* versions, Env* env, FileSystem* fs,
+    const ImmutableCFOptions* immutable_cf_options,
+    const MutableCFOptions* mutable_cf_options, const FileOptions* file_options,
+    int job_id, uint32_t column_family_id,
+    const std::string& column_family_name, Env::IOPriority io_priority,
+    Env::WriteLifeTimeHint write_hint,
+    std::vector<std::string>* blob_file_paths,
+    std::vector<BlobFileAddition>* blob_file_additions)
+    : BlobFileBuilder([versions]() { return versions->NewFileNumber(); }, env,
+                      fs, immutable_cf_options, mutable_cf_options,
+                      file_options, job_id, column_family_id,
+                      column_family_name, io_priority, write_hint,
+                      blob_file_paths, blob_file_additions) {}
+
+BlobFileBuilder::BlobFileBuilder(
+    std::function<uint64_t()> file_number_generator, Env* env, FileSystem* fs,
+    const ImmutableCFOptions* immutable_cf_options,
+    const MutableCFOptions* mutable_cf_options, const FileOptions* file_options,
+    int job_id, uint32_t column_family_id,
+    const std::string& column_family_name, Env::IOPriority io_priority,
+    Env::WriteLifeTimeHint write_hint,
+    std::vector<std::string>* blob_file_paths,
+    std::vector<BlobFileAddition>* blob_file_additions)
+    : file_number_generator_(std::move(file_number_generator)),
+      env_(env),
+      fs_(fs),
+      immutable_cf_options_(immutable_cf_options),
+      min_blob_size_(mutable_cf_options->min_blob_size),
+      blob_file_size_(mutable_cf_options->blob_file_size),
+      blob_compression_type_(mutable_cf_options->blob_compression_type),
+      file_options_(file_options),
+      job_id_(job_id),
+      column_family_id_(column_family_id),
+      column_family_name_(column_family_name),
+      io_priority_(io_priority),
+      write_hint_(write_hint),
+      blob_file_paths_(blob_file_paths),
+      blob_file_additions_(blob_file_additions),
+      blob_count_(0),
+      blob_bytes_(0) {
+  assert(file_number_generator_);
+  assert(env_);
+  assert(fs_);
+  assert(immutable_cf_options_);
+  assert(file_options_);
+  assert(blob_file_paths_);
+  assert(blob_file_paths_->empty());
+  assert(blob_file_additions_);
+  assert(blob_file_additions_->empty());
+}
+
+BlobFileBuilder::~BlobFileBuilder() = default;
+
+Status BlobFileBuilder::Add(const Slice& key, const Slice& value,
+                            std::string* blob_index) {
+  assert(blob_index);
+  assert(blob_index->empty());
+
+  if (value.size() < min_blob_size_) {
+    return Status::OK();
+  }
+
+  {
+    const Status s = OpenBlobFileIfNeeded();
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  Slice blob = value;
+  std::string compressed_blob;
+
+  {
+    const Status s = CompressBlobIfNeeded(&blob, &compressed_blob);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  uint64_t blob_file_number = 0;
+  uint64_t blob_offset = 0;
+
+  {
+    const Status s =
+        WriteBlobToFile(key, blob, &blob_file_number, &blob_offset);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  {
+    const Status s = CloseBlobFileIfNeeded();
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  BlobIndex::EncodeBlob(blob_index, blob_file_number, blob_offset, blob.size(),
+                        blob_compression_type_);
+
+  return Status::OK();
+}
+
+Status BlobFileBuilder::Finish() {
+  if (!IsBlobFileOpen()) {
+    return Status::OK();
+  }
+
+  return CloseBlobFile();
+}
+
+bool BlobFileBuilder::IsBlobFileOpen() const { return !!writer_; }
+
+Status BlobFileBuilder::OpenBlobFileIfNeeded() {
+  if (IsBlobFileOpen()) {
+    return Status::OK();
+  }
+
+  assert(!blob_count_);
+  assert(!blob_bytes_);
+
+  assert(file_number_generator_);
+  const uint64_t blob_file_number = file_number_generator_();
+
+  assert(immutable_cf_options_);
+  assert(!immutable_cf_options_->cf_paths.empty());
+  std::string blob_file_path = BlobFileName(
+      immutable_cf_options_->cf_paths.front().path, blob_file_number);
+
+  std::unique_ptr<FSWritableFile> file;
+
+  {
+    TEST_SYNC_POINT("BlobFileBuilder::OpenBlobFileIfNeeded:NewWritableFile");
+
+    assert(file_options_);
+    const Status s =
+        NewWritableFile(fs_, blob_file_path, &file, *file_options_);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  // Note: files get added to blob_file_paths_ right after the open, so they
+  // can be cleaned up upon failure. Contrast this with blob_file_additions_,
+  // which only contains successfully written files.
+  assert(blob_file_paths_);
+  blob_file_paths_->emplace_back(std::move(blob_file_path));
+
+  assert(file);
+  file->SetIOPriority(io_priority_);
+  file->SetWriteLifeTimeHint(write_hint_);
+
+  Statistics* const statistics = immutable_cf_options_->statistics;
+
+  std::unique_ptr<WritableFileWriter> file_writer(new WritableFileWriter(
+      std::move(file), blob_file_paths_->back(), *file_options_, env_,
+      nullptr /*IOTracer*/, statistics, immutable_cf_options_->listeners,
+      immutable_cf_options_->file_checksum_gen_factory));
+
+  std::unique_ptr<BlobLogWriter> blob_log_writer(
+      new BlobLogWriter(std::move(file_writer), env_, statistics,
+                        blob_file_number, immutable_cf_options_->use_fsync));
+
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+
+  BlobLogHeader header(column_family_id_, blob_compression_type_, has_ttl,
+                       expiration_range);
+
+  {
+    TEST_SYNC_POINT("BlobFileBuilder::OpenBlobFileIfNeeded:WriteHeader");
+
+    const Status s = blob_log_writer->WriteHeader(header);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  writer_ = std::move(blob_log_writer);
+
+  assert(IsBlobFileOpen());
+
+  return Status::OK();
+}
+
+Status BlobFileBuilder::CompressBlobIfNeeded(
+    Slice* blob, std::string* compressed_blob) const {
+  assert(blob);
+  assert(compressed_blob);
+  assert(compressed_blob->empty());
+
+  if (blob_compression_type_ == kNoCompression) {
+    return Status::OK();
+  }
+
+  CompressionOptions opts;
+  CompressionContext context(blob_compression_type_);
+  constexpr uint64_t sample_for_compression = 0;
+
+  CompressionInfo info(opts, context, CompressionDict::GetEmptyDict(),
+                       blob_compression_type_, sample_for_compression);
+
+  constexpr uint32_t compression_format_version = 2;
+
+  if (!CompressData(*blob, info, compression_format_version, compressed_blob)) {
+    return Status::Corruption("Error compressing blob");
+  }
+
+  *blob = Slice(*compressed_blob);
+
+  return Status::OK();
+}
+
+Status BlobFileBuilder::WriteBlobToFile(const Slice& key, const Slice& blob,
+                                        uint64_t* blob_file_number,
+                                        uint64_t* blob_offset) {
+  assert(IsBlobFileOpen());
+  assert(blob_file_number);
+  assert(blob_offset);
+
+  uint64_t key_offset = 0;
+
+  TEST_SYNC_POINT("BlobFileBuilder::WriteBlobToFile:AddRecord");
+
+  const Status s = writer_->AddRecord(key, blob, &key_offset, blob_offset);
+  if (!s.ok()) {
+    return s;
+  }
+
+  *blob_file_number = writer_->get_log_number();
+
+  ++blob_count_;
+  blob_bytes_ += BlobLogRecord::kHeaderSize + key.size() + blob.size();
+
+  return Status::OK();
+}
+
+Status BlobFileBuilder::CloseBlobFile() {
+  assert(IsBlobFileOpen());
+
+  BlobLogFooter footer;
+  footer.blob_count = blob_count_;
+
+  std::string checksum_method;
+  std::string checksum_value;
+
+  TEST_SYNC_POINT("BlobFileBuilder::WriteBlobToFile:AppendFooter");
+
+  const Status s =
+      writer_->AppendFooter(footer, &checksum_method, &checksum_value);
+  if (!s.ok()) {
+    return s;
+  }
+
+  const uint64_t blob_file_number = writer_->get_log_number();
+
+  assert(blob_file_additions_);
+  blob_file_additions_->emplace_back(blob_file_number, blob_count_, blob_bytes_,
+                                     std::move(checksum_method),
+                                     std::move(checksum_value));
+
+  assert(immutable_cf_options_);
+  ROCKS_LOG_INFO(immutable_cf_options_->info_log,
+                 "[%s] [JOB %d] Generated blob file #%" PRIu64 ": %" PRIu64
+                 " total blobs, %" PRIu64 " total bytes",
+                 column_family_name_.c_str(), job_id_, blob_file_number,
+                 blob_count_, blob_bytes_);
+
+  writer_.reset();
+  blob_count_ = 0;
+  blob_bytes_ = 0;
+
+  return Status::OK();
+}
+
+Status BlobFileBuilder::CloseBlobFileIfNeeded() {
+  assert(IsBlobFileOpen());
+
+  const WritableFileWriter* const file_writer = writer_->file();
+  assert(file_writer);
+
+  if (file_writer->GetFileSize() < blob_file_size_) {
+    return Status::OK();
+  }
+
+  return CloseBlobFile();
+}
+
+}  // namespace ROCKSDB_NAMESPACE

db/blob/blob_file_builder.h

@@ -0,0 +1,91 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+#pragma once
+
+#include <cinttypes>
+#include <functional>
+#include <memory>
+#include <string>
+#include <vector>
+
+#include "rocksdb/compression_type.h"
+#include "rocksdb/env.h"
+#include "rocksdb/rocksdb_namespace.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+class VersionSet;
+class FileSystem;
+struct ImmutableCFOptions;
+struct MutableCFOptions;
+struct FileOptions;
+class BlobFileAddition;
+class Status;
+class Slice;
+class BlobLogWriter;
+
+class BlobFileBuilder {
+ public:
+  BlobFileBuilder(VersionSet* versions, Env* env, FileSystem* fs,
+                  const ImmutableCFOptions* immutable_cf_options,
+                  const MutableCFOptions* mutable_cf_options,
+                  const FileOptions* file_options, int job_id,
+                  uint32_t column_family_id,
+                  const std::string& column_family_name,
+                  Env::IOPriority io_priority,
+                  Env::WriteLifeTimeHint write_hint,
+                  std::vector<std::string>* blob_file_paths,
+                  std::vector<BlobFileAddition>* blob_file_additions);
+
+  BlobFileBuilder(std::function<uint64_t()> file_number_generator, Env* env,
+                  FileSystem* fs,
+                  const ImmutableCFOptions* immutable_cf_options,
+                  const MutableCFOptions* mutable_cf_options,
+                  const FileOptions* file_options, int job_id,
+                  uint32_t column_family_id,
+                  const std::string& column_family_name,
+                  Env::IOPriority io_priority,
+                  Env::WriteLifeTimeHint write_hint,
+                  std::vector<std::string>* blob_file_paths,
+                  std::vector<BlobFileAddition>* blob_file_additions);
+
+  BlobFileBuilder(const BlobFileBuilder&) = delete;
+  BlobFileBuilder& operator=(const BlobFileBuilder&) = delete;
+
+  ~BlobFileBuilder();
+
+  Status Add(const Slice& key, const Slice& value, std::string* blob_index);
+  Status Finish();
+
+ private:
+  bool IsBlobFileOpen() const;
+  Status OpenBlobFileIfNeeded();
+  Status CompressBlobIfNeeded(Slice* blob, std::string* compressed_blob) const;
+  Status WriteBlobToFile(const Slice& key, const Slice& blob,
+                         uint64_t* blob_file_number, uint64_t* blob_offset);
+  Status CloseBlobFile();
+  Status CloseBlobFileIfNeeded();
+
+  std::function<uint64_t()> file_number_generator_;
+  Env* env_;
+  FileSystem* fs_;
+  const ImmutableCFOptions* immutable_cf_options_;
+  uint64_t min_blob_size_;
+  uint64_t blob_file_size_;
+  CompressionType blob_compression_type_;
+  const FileOptions* file_options_;
+  int job_id_;
+  uint32_t column_family_id_;
+  std::string column_family_name_;
+  Env::IOPriority io_priority_;
+  Env::WriteLifeTimeHint write_hint_;
+  std::vector<std::string>* blob_file_paths_;
+  std::vector<BlobFileAddition>* blob_file_additions_;
+  std::unique_ptr<BlobLogWriter> writer_;
+  uint64_t blob_count_;
+  uint64_t blob_bytes_;
+};
+
+}  // namespace ROCKSDB_NAMESPACE

db/blob/blob_file_builder_test.cc

@@ -0,0 +1,659 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include "db/blob/blob_file_builder.h"
+
+#include <cassert>
+#include <cinttypes>
+#include <string>
+#include <utility>
+#include <vector>
+
+#include "db/blob/blob_file_addition.h"
+#include "db/blob/blob_index.h"
+#include "db/blob/blob_log_format.h"
+#include "db/blob/blob_log_sequential_reader.h"
+#include "env/composite_env_wrapper.h"
+#include "env/mock_env.h"
+#include "file/filename.h"
+#include "file/random_access_file_reader.h"
+#include "options/cf_options.h"
+#include "rocksdb/env.h"
+#include "rocksdb/file_checksum.h"
+#include "rocksdb/options.h"
+#include "test_util/sync_point.h"
+#include "test_util/testharness.h"
+#include "util/compression.h"
+#include "utilities/fault_injection_env.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+class TestFileNumberGenerator {
+ public:
+  uint64_t operator()() { return ++next_file_number_; }
+
+ private:
+  uint64_t next_file_number_ = 1;
+};
+
+class BlobFileBuilderTest : public testing::Test {
+ protected:
+  BlobFileBuilderTest() : mock_env_(Env::Default()), fs_(&mock_env_) {}
+
+  void VerifyBlobFile(uint64_t blob_file_number,
+                      const std::string& blob_file_path,
+                      uint32_t column_family_id,
+                      CompressionType blob_compression_type,
+                      const std::vector<std::pair<std::string, std::string>>&
+                          expected_key_value_pairs,
+                      const std::vector<std::string>& blob_indexes) {
+    assert(expected_key_value_pairs.size() == blob_indexes.size());
+
+    std::unique_ptr<FSRandomAccessFile> file;
+    constexpr IODebugContext* dbg = nullptr;
+    ASSERT_OK(
+        fs_.NewRandomAccessFile(blob_file_path, file_options_, &file, dbg));
+
+    std::unique_ptr<RandomAccessFileReader> file_reader(
+        new RandomAccessFileReader(std::move(file), blob_file_path,
+                                   &mock_env_));
+
+    constexpr Statistics* statistics = nullptr;
+    BlobLogSequentialReader blob_log_reader(std::move(file_reader), &mock_env_,
+                                            statistics);
+
+    BlobLogHeader header;
+    ASSERT_OK(blob_log_reader.ReadHeader(&header));
+    ASSERT_EQ(header.version, kVersion1);
+    ASSERT_EQ(header.column_family_id, column_family_id);
+    ASSERT_EQ(header.compression, blob_compression_type);
+    ASSERT_FALSE(header.has_ttl);
+    ASSERT_EQ(header.expiration_range, ExpirationRange());
+
+    for (size_t i = 0; i < expected_key_value_pairs.size(); ++i) {
+      BlobLogRecord record;
+      uint64_t blob_offset = 0;
+
+      ASSERT_OK(blob_log_reader.ReadRecord(
+          &record, BlobLogSequentialReader::kReadHeaderKeyBlob, &blob_offset));
+
+      // Check the contents of the blob file
+      const auto& expected_key_value = expected_key_value_pairs[i];
+      const auto& key = expected_key_value.first;
+      const auto& value = expected_key_value.second;
+
+      ASSERT_EQ(record.key_size, key.size());
+      ASSERT_EQ(record.value_size, value.size());
+      ASSERT_EQ(record.expiration, 0);
+      ASSERT_EQ(record.key, key);
+      ASSERT_EQ(record.value, value);
+
+      // Make sure the blob reference returned by the builder points to the
+      // right place
+      BlobIndex blob_index;
+      ASSERT_OK(blob_index.DecodeFrom(blob_indexes[i]));
+      ASSERT_FALSE(blob_index.IsInlined());
+      ASSERT_FALSE(blob_index.HasTTL());
+      ASSERT_EQ(blob_index.file_number(), blob_file_number);
+      ASSERT_EQ(blob_index.offset(), blob_offset);
+      ASSERT_EQ(blob_index.size(), value.size());
+    }
+
+    BlobLogFooter footer;
+    ASSERT_OK(blob_log_reader.ReadFooter(&footer));
+    ASSERT_EQ(footer.blob_count, expected_key_value_pairs.size());
+    ASSERT_EQ(footer.expiration_range, ExpirationRange());
+  }
+
+  MockEnv mock_env_;
+  LegacyFileSystemWrapper fs_;
+  FileOptions file_options_;
+};
+
+TEST_F(BlobFileBuilderTest, BuildAndCheckOneFile) {
+  // Build a single blob file
+  constexpr size_t number_of_blobs = 10;
+  constexpr size_t key_size = 1;
+  constexpr size_t value_size = 4;
+  constexpr size_t value_offset = 1234;
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileBuilderTest_BuildAndCheckOneFile"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &mock_env_, &fs_,
+                          &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  std::vector<std::pair<std::string, std::string>> expected_key_value_pairs(
+      number_of_blobs);
+  std::vector<std::string> blob_indexes(number_of_blobs);
+
+  for (size_t i = 0; i < number_of_blobs; ++i) {
+    auto& expected_key_value = expected_key_value_pairs[i];
+
+    auto& key = expected_key_value.first;
+    key = std::to_string(i);
+    assert(key.size() == key_size);
+
+    auto& value = expected_key_value.second;
+    value = std::to_string(i + value_offset);
+    assert(value.size() == value_size);
+
+    auto& blob_index = blob_indexes[i];
+
+    ASSERT_OK(builder.Add(key, value, &blob_index));
+    ASSERT_FALSE(blob_index.empty());
+  }
+
+  ASSERT_OK(builder.Finish());
+
+  // Check the metadata generated
+  constexpr uint64_t blob_file_number = 2;
+
+  ASSERT_EQ(blob_file_paths.size(), 1);
+
+  const std::string& blob_file_path = blob_file_paths[0];
+
+  ASSERT_EQ(blob_file_path,
+            BlobFileName(immutable_cf_options.cf_paths.front().path,
+                         blob_file_number));
+
+  ASSERT_EQ(blob_file_additions.size(), 1);
+
+  const auto& blob_file_addition = blob_file_additions[0];
+
+  ASSERT_EQ(blob_file_addition.GetBlobFileNumber(), blob_file_number);
+  ASSERT_EQ(blob_file_addition.GetTotalBlobCount(), number_of_blobs);
+  ASSERT_EQ(
+      blob_file_addition.GetTotalBlobBytes(),
+      number_of_blobs * (BlobLogRecord::kHeaderSize + key_size + value_size));
+
+  // Verify the contents of the new blob file as well as the blob references
+  VerifyBlobFile(blob_file_number, blob_file_path, column_family_id,
+                 kNoCompression, expected_key_value_pairs, blob_indexes);
+}
+
+TEST_F(BlobFileBuilderTest, BuildAndCheckMultipleFiles) {
+  // Build multiple blob files: file size limit is set to the size of a single
+  // value, so each blob ends up in a file of its own
+  constexpr size_t number_of_blobs = 10;
+  constexpr size_t key_size = 1;
+  constexpr size_t value_size = 10;
+  constexpr size_t value_offset = 1234567890;
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileBuilderTest_BuildAndCheckMultipleFiles"),
+      0);
+  options.enable_blob_files = true;
+  options.blob_file_size = value_size;
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &mock_env_, &fs_,
+                          &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  std::vector<std::pair<std::string, std::string>> expected_key_value_pairs(
+      number_of_blobs);
+  std::vector<std::string> blob_indexes(number_of_blobs);
+
+  for (size_t i = 0; i < number_of_blobs; ++i) {
+    auto& expected_key_value = expected_key_value_pairs[i];
+
+    auto& key = expected_key_value.first;
+    key = std::to_string(i);
+    assert(key.size() == key_size);
+
+    auto& value = expected_key_value.second;
+    value = std::to_string(i + value_offset);
+    assert(value.size() == value_size);
+
+    auto& blob_index = blob_indexes[i];
+
+    ASSERT_OK(builder.Add(key, value, &blob_index));
+    ASSERT_FALSE(blob_index.empty());
+  }
+
+  ASSERT_OK(builder.Finish());
+
+  // Check the metadata generated
+  ASSERT_EQ(blob_file_paths.size(), number_of_blobs);
+  ASSERT_EQ(blob_file_additions.size(), number_of_blobs);
+
+  for (size_t i = 0; i < number_of_blobs; ++i) {
+    const uint64_t blob_file_number = i + 2;
+
+    ASSERT_EQ(blob_file_paths[i],
+              BlobFileName(immutable_cf_options.cf_paths.front().path,
+                           blob_file_number));
+
+    const auto& blob_file_addition = blob_file_additions[i];
+
+    ASSERT_EQ(blob_file_addition.GetBlobFileNumber(), blob_file_number);
+    ASSERT_EQ(blob_file_addition.GetTotalBlobCount(), 1);
+    ASSERT_EQ(blob_file_addition.GetTotalBlobBytes(),
+              BlobLogRecord::kHeaderSize + key_size + value_size);
+  }
+
+  // Verify the contents of the new blob files as well as the blob references
+  for (size_t i = 0; i < number_of_blobs; ++i) {
+    std::vector<std::pair<std::string, std::string>> expected_key_value_pair{
+        expected_key_value_pairs[i]};
+    std::vector<std::string> blob_index{blob_indexes[i]};
+
+    VerifyBlobFile(i + 2, blob_file_paths[i], column_family_id, kNoCompression,
+                   expected_key_value_pair, blob_index);
+  }
+}
+
+TEST_F(BlobFileBuilderTest, InlinedValues) {
+  // All values are below the min_blob_size threshold; no blob files get written
+  constexpr size_t number_of_blobs = 10;
+  constexpr size_t key_size = 1;
+  constexpr size_t value_size = 10;
+  constexpr size_t value_offset = 1234567890;
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileBuilderTest_InlinedValues"),
+      0);
+  options.enable_blob_files = true;
+  options.min_blob_size = 1024;
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &mock_env_, &fs_,
+                          &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  for (size_t i = 0; i < number_of_blobs; ++i) {
+    const std::string key = std::to_string(i);
+    assert(key.size() == key_size);
+
+    const std::string value = std::to_string(i + value_offset);
+    assert(value.size() == value_size);
+
+    std::string blob_index;
+    ASSERT_OK(builder.Add(key, value, &blob_index));
+    ASSERT_TRUE(blob_index.empty());
+  }
+
+  ASSERT_OK(builder.Finish());
+
+  // Check the metadata generated
+  ASSERT_TRUE(blob_file_paths.empty());
+  ASSERT_TRUE(blob_file_additions.empty());
+}
+
+TEST_F(BlobFileBuilderTest, Compression) {
+  // Build a blob file with a compressed blob
+  if (!Snappy_Supported()) {
+    return;
+  }
+
+  constexpr size_t key_size = 1;
+  constexpr size_t value_size = 100;
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileBuilderTest_Compression"), 0);
+  options.enable_blob_files = true;
+  options.blob_compression_type = kSnappyCompression;
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &mock_env_, &fs_,
+                          &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  const std::string key("1");
+  const std::string uncompressed_value(value_size, 'x');
+
+  std::string blob_index;
+
+  ASSERT_OK(builder.Add(key, uncompressed_value, &blob_index));
+  ASSERT_FALSE(blob_index.empty());
+
+  ASSERT_OK(builder.Finish());
+
+  // Check the metadata generated
+  constexpr uint64_t blob_file_number = 2;
+
+  ASSERT_EQ(blob_file_paths.size(), 1);
+
+  const std::string& blob_file_path = blob_file_paths[0];
+
+  ASSERT_EQ(blob_file_path,
+            BlobFileName(immutable_cf_options.cf_paths.front().path,
+                         blob_file_number));
+
+  ASSERT_EQ(blob_file_additions.size(), 1);
+
+  const auto& blob_file_addition = blob_file_additions[0];
+
+  ASSERT_EQ(blob_file_addition.GetBlobFileNumber(), blob_file_number);
+  ASSERT_EQ(blob_file_addition.GetTotalBlobCount(), 1);
+
+  CompressionOptions opts;
+  CompressionContext context(kSnappyCompression);
+  constexpr uint64_t sample_for_compression = 0;
+
+  CompressionInfo info(opts, context, CompressionDict::GetEmptyDict(),
+                       kSnappyCompression, sample_for_compression);
+
+  std::string compressed_value;
+  ASSERT_TRUE(Snappy_Compress(info, uncompressed_value.data(),
+                              uncompressed_value.size(), &compressed_value));
+
+  ASSERT_EQ(blob_file_addition.GetTotalBlobBytes(),
+            BlobLogRecord::kHeaderSize + key_size + compressed_value.size());
+
+  // Verify the contents of the new blob file as well as the blob reference
+  std::vector<std::pair<std::string, std::string>> expected_key_value_pairs{
+      {key, compressed_value}};
+  std::vector<std::string> blob_indexes{blob_index};
+
+  VerifyBlobFile(blob_file_number, blob_file_path, column_family_id,
+                 kSnappyCompression, expected_key_value_pairs, blob_indexes);
+}
+
+TEST_F(BlobFileBuilderTest, CompressionError) {
+  // Simulate an error during compression
+  if (!Snappy_Supported()) {
+    return;
+  }
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileBuilderTest_CompressionError"),
+      0);
+  options.enable_blob_files = true;
+  options.blob_compression_type = kSnappyCompression;
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &mock_env_, &fs_,
+                          &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  SyncPoint::GetInstance()->SetCallBack("CompressData:TamperWithReturnValue",
+                                        [](void* arg) {
+                                          bool* ret = static_cast<bool*>(arg);
+                                          *ret = false;
+                                        });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  constexpr char key[] = "1";
+  constexpr char value[] = "deadbeef";
+
+  std::string blob_index;
+
+  ASSERT_TRUE(builder.Add(key, value, &blob_index).IsCorruption());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+
+  constexpr uint64_t blob_file_number = 2;
+
+  ASSERT_EQ(blob_file_paths.size(), 1);
+  ASSERT_EQ(blob_file_paths[0],
+            BlobFileName(immutable_cf_options.cf_paths.front().path,
+                         blob_file_number));
+
+  ASSERT_TRUE(blob_file_additions.empty());
+}
+
+TEST_F(BlobFileBuilderTest, Checksum) {
+  // Build a blob file with checksum
+
+  class DummyFileChecksumGenerator : public FileChecksumGenerator {
+   public:
+    void Update(const char* /* data */, size_t /* n */) override {}
+
+    void Finalize() override {}
+
+    std::string GetChecksum() const override { return std::string("dummy"); }
+
+    const char* Name() const override { return "DummyFileChecksum"; }
+  };
+
+  class DummyFileChecksumGenFactory : public FileChecksumGenFactory {
+   public:
+    std::unique_ptr<FileChecksumGenerator> CreateFileChecksumGenerator(
+        const FileChecksumGenContext& /* context */) override {
+      return std::unique_ptr<FileChecksumGenerator>(
+          new DummyFileChecksumGenerator);
+    }
+
+    const char* Name() const override { return "DummyFileChecksumGenFactory"; }
+  };
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileBuilderTest_Checksum"), 0);
+  options.enable_blob_files = true;
+  options.file_checksum_gen_factory =
+      std::make_shared<DummyFileChecksumGenFactory>();
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &mock_env_, &fs_,
+                          &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  const std::string key("1");
+  const std::string value("deadbeef");
+
+  std::string blob_index;
+
+  ASSERT_OK(builder.Add(key, value, &blob_index));
+  ASSERT_FALSE(blob_index.empty());
+
+  ASSERT_OK(builder.Finish());
+
+  // Check the metadata generated
+  constexpr uint64_t blob_file_number = 2;
+
+  ASSERT_EQ(blob_file_paths.size(), 1);
+
+  const std::string& blob_file_path = blob_file_paths[0];
+
+  ASSERT_EQ(blob_file_path,
+            BlobFileName(immutable_cf_options.cf_paths.front().path,
+                         blob_file_number));
+
+  ASSERT_EQ(blob_file_additions.size(), 1);
+
+  const auto& blob_file_addition = blob_file_additions[0];
+
+  ASSERT_EQ(blob_file_addition.GetBlobFileNumber(), blob_file_number);
+  ASSERT_EQ(blob_file_addition.GetTotalBlobCount(), 1);
+  ASSERT_EQ(blob_file_addition.GetTotalBlobBytes(),
+            BlobLogRecord::kHeaderSize + key.size() + value.size());
+  ASSERT_EQ(blob_file_addition.GetChecksumMethod(), "DummyFileChecksum");
+  ASSERT_EQ(blob_file_addition.GetChecksumValue(), "dummy");
+
+  // Verify the contents of the new blob file as well as the blob reference
+  std::vector<std::pair<std::string, std::string>> expected_key_value_pairs{
+      {key, value}};
+  std::vector<std::string> blob_indexes{blob_index};
+
+  VerifyBlobFile(blob_file_number, blob_file_path, column_family_id,
+                 kNoCompression, expected_key_value_pairs, blob_indexes);
+}
+
+class BlobFileBuilderIOErrorTest
+    : public testing::Test,
+      public testing::WithParamInterface<std::string> {
+ protected:
+  BlobFileBuilderIOErrorTest()
+      : mock_env_(Env::Default()),
+        fault_injection_env_(&mock_env_),
+        fs_(&fault_injection_env_),
+        sync_point_(GetParam()) {}
+
+  MockEnv mock_env_;
+  FaultInjectionTestEnv fault_injection_env_;
+  LegacyFileSystemWrapper fs_;
+  FileOptions file_options_;
+  std::string sync_point_;
+};
+
+INSTANTIATE_TEST_CASE_P(
+    BlobFileBuilderTest, BlobFileBuilderIOErrorTest,
+    ::testing::ValuesIn(std::vector<std::string>{
+        "BlobFileBuilder::OpenBlobFileIfNeeded:NewWritableFile",
+        "BlobFileBuilder::OpenBlobFileIfNeeded:WriteHeader",
+        "BlobFileBuilder::WriteBlobToFile:AddRecord",
+        "BlobFileBuilder::WriteBlobToFile:AppendFooter"}));
+
+TEST_P(BlobFileBuilderIOErrorTest, IOError) {
+  // Simulate an I/O error during the specified step of Add()
+  // Note: blob_file_size will be set to value_size in order for the first blob
+  // to trigger close
+  constexpr size_t value_size = 8;
+
+  Options options;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&fault_injection_env_,
+                            "BlobFileBuilderIOErrorTest_IOError"),
+      0);
+  options.enable_blob_files = true;
+  options.blob_file_size = value_size;
+
+  ImmutableCFOptions immutable_cf_options(options);
+  MutableCFOptions mutable_cf_options(options);
+
+  constexpr int job_id = 1;
+  constexpr uint32_t column_family_id = 123;
+  constexpr char column_family_name[] = "foobar";
+  constexpr Env::IOPriority io_priority = Env::IO_HIGH;
+  constexpr Env::WriteLifeTimeHint write_hint = Env::WLTH_MEDIUM;
+
+  std::vector<std::string> blob_file_paths;
+  std::vector<BlobFileAddition> blob_file_additions;
+
+  BlobFileBuilder builder(TestFileNumberGenerator(), &fault_injection_env_,
+                          &fs_, &immutable_cf_options, &mutable_cf_options,
+                          &file_options_, job_id, column_family_id,
+                          column_family_name, io_priority, write_hint,
+                          &blob_file_paths, &blob_file_additions);
+
+  SyncPoint::GetInstance()->SetCallBack(sync_point_, [this](void* /* arg */) {
+    fault_injection_env_.SetFilesystemActive(false,
+                                             Status::IOError(sync_point_));
+  });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  constexpr char key[] = "1";
+  constexpr char value[] = "deadbeef";
+
+  std::string blob_index;
+
+  ASSERT_TRUE(builder.Add(key, value, &blob_index).IsIOError());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+
+  if (sync_point_ == "BlobFileBuilder::OpenBlobFileIfNeeded:NewWritableFile") {
+    ASSERT_TRUE(blob_file_paths.empty());
+  } else {
+    constexpr uint64_t blob_file_number = 2;
+
+    ASSERT_EQ(blob_file_paths.size(), 1);
+    ASSERT_EQ(blob_file_paths[0],
+              BlobFileName(immutable_cf_options.cf_paths.front().path,
+                           blob_file_number));
+  }
+
+  ASSERT_TRUE(blob_file_additions.empty());
+}
+
+}  // namespace ROCKSDB_NAMESPACE
+
+int main(int argc, char** argv) {
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

db/blob/blob_file_reader.cc

@@ -0,0 +1,423 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include "db/blob/blob_file_reader.h"
+
+#include <cassert>
+#include <string>
+
+#include "db/blob/blob_log_format.h"
+#include "file/filename.h"
+#include "options/cf_options.h"
+#include "rocksdb/file_system.h"
+#include "rocksdb/slice.h"
+#include "rocksdb/status.h"
+#include "test_util/sync_point.h"
+#include "util/compression.h"
+#include "util/crc32c.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+Status BlobFileReader::Create(
+    const ImmutableCFOptions& immutable_cf_options,
+    const FileOptions& file_options, uint32_t column_family_id,
+    HistogramImpl* blob_file_read_hist, uint64_t blob_file_number,
+    std::unique_ptr<BlobFileReader>* blob_file_reader) {
+  assert(blob_file_reader);
+  assert(!*blob_file_reader);
+
+  uint64_t file_size = 0;
+  std::unique_ptr<RandomAccessFileReader> file_reader;
+
+  {
+    const Status s =
+        OpenFile(immutable_cf_options, file_options, blob_file_read_hist,
+                 blob_file_number, &file_size, &file_reader);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  assert(file_reader);
+
+  CompressionType compression_type = kNoCompression;
+
+  {
+    const Status s =
+        ReadHeader(file_reader.get(), column_family_id, &compression_type);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  {
+    const Status s = ReadFooter(file_size, file_reader.get());
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  blob_file_reader->reset(
+      new BlobFileReader(std::move(file_reader), file_size, compression_type));
+
+  return Status::OK();
+}
+
+Status BlobFileReader::OpenFile(
+    const ImmutableCFOptions& immutable_cf_options,
+    const FileOptions& file_opts, HistogramImpl* blob_file_read_hist,
+    uint64_t blob_file_number, uint64_t* file_size,
+    std::unique_ptr<RandomAccessFileReader>* file_reader) {
+  assert(file_size);
+  assert(file_reader);
+
+  const auto& cf_paths = immutable_cf_options.cf_paths;
+  assert(!cf_paths.empty());
+
+  const std::string blob_file_path =
+      BlobFileName(cf_paths.front().path, blob_file_number);
+
+  FileSystem* const fs = immutable_cf_options.fs;
+  assert(fs);
+
+  constexpr IODebugContext* dbg = nullptr;
+
+  {
+    TEST_SYNC_POINT("BlobFileReader::OpenFile:GetFileSize");
+
+    const Status s =
+        fs->GetFileSize(blob_file_path, IOOptions(), file_size, dbg);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  if (*file_size < BlobLogHeader::kSize + BlobLogFooter::kSize) {
+    return Status::Corruption("Malformed blob file");
+  }
+
+  std::unique_ptr<FSRandomAccessFile> file;
+
+  {
+    TEST_SYNC_POINT("BlobFileReader::OpenFile:NewRandomAccessFile");
+
+    const Status s =
+        fs->NewRandomAccessFile(blob_file_path, file_opts, &file, dbg);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  assert(file);
+
+  if (immutable_cf_options.advise_random_on_open) {
+    file->Hint(FSRandomAccessFile::kRandom);
+  }
+
+  file_reader->reset(new RandomAccessFileReader(
+      std::move(file), blob_file_path, immutable_cf_options.env,
+      std::shared_ptr<IOTracer>(), immutable_cf_options.statistics,
+      BLOB_DB_BLOB_FILE_READ_MICROS, blob_file_read_hist,
+      immutable_cf_options.rate_limiter, immutable_cf_options.listeners));
+
+  return Status::OK();
+}
+
+Status BlobFileReader::ReadHeader(const RandomAccessFileReader* file_reader,
+                                  uint32_t column_family_id,
+                                  CompressionType* compression_type) {
+  assert(file_reader);
+  assert(compression_type);
+
+  Slice header_slice;
+  Buffer buf;
+  AlignedBuf aligned_buf;
+
+  {
+    TEST_SYNC_POINT("BlobFileReader::ReadHeader:ReadFromFile");
+
+    constexpr uint64_t read_offset = 0;
+    constexpr size_t read_size = BlobLogHeader::kSize;
+
+    const Status s = ReadFromFile(file_reader, read_offset, read_size,
+                                  &header_slice, &buf, &aligned_buf);
+    if (!s.ok()) {
+      return s;
+    }
+
+    TEST_SYNC_POINT_CALLBACK("BlobFileReader::ReadHeader:TamperWithResult",
+                             &header_slice);
+  }
+
+  BlobLogHeader header;
+
+  {
+    const Status s = header.DecodeFrom(header_slice);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  constexpr ExpirationRange no_expiration_range;
+
+  if (header.has_ttl || header.expiration_range != no_expiration_range) {
+    return Status::Corruption("Unexpected TTL blob file");
+  }
+
+  if (header.column_family_id != column_family_id) {
+    return Status::Corruption("Column family ID mismatch");
+  }
+
+  *compression_type = header.compression;
+
+  return Status::OK();
+}
+
+Status BlobFileReader::ReadFooter(uint64_t file_size,
+                                  const RandomAccessFileReader* file_reader) {
+  assert(file_size >= BlobLogHeader::kSize + BlobLogFooter::kSize);
+  assert(file_reader);
+
+  Slice footer_slice;
+  Buffer buf;
+  AlignedBuf aligned_buf;
+
+  {
+    TEST_SYNC_POINT("BlobFileReader::ReadFooter:ReadFromFile");
+
+    const uint64_t read_offset = file_size - BlobLogFooter::kSize;
+    constexpr size_t read_size = BlobLogFooter::kSize;
+
+    const Status s = ReadFromFile(file_reader, read_offset, read_size,
+                                  &footer_slice, &buf, &aligned_buf);
+    if (!s.ok()) {
+      return s;
+    }
+
+    TEST_SYNC_POINT_CALLBACK("BlobFileReader::ReadFooter:TamperWithResult",
+                             &footer_slice);
+  }
+
+  BlobLogFooter footer;
+
+  {
+    const Status s = footer.DecodeFrom(footer_slice);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  constexpr ExpirationRange no_expiration_range;
+
+  if (footer.expiration_range != no_expiration_range) {
+    return Status::Corruption("Unexpected TTL blob file");
+  }
+
+  return Status::OK();
+}
+
+Status BlobFileReader::ReadFromFile(const RandomAccessFileReader* file_reader,
+                                    uint64_t read_offset, size_t read_size,
+                                    Slice* slice, Buffer* buf,
+                                    AlignedBuf* aligned_buf) {
+  assert(slice);
+  assert(buf);
+  assert(aligned_buf);
+
+  assert(file_reader);
+
+  Status s;
+
+  if (file_reader->use_direct_io()) {
+    constexpr char* scratch = nullptr;
+
+    s = file_reader->Read(IOOptions(), read_offset, read_size, slice, scratch,
+                          aligned_buf);
+  } else {
+    buf->reset(new char[read_size]);
+    constexpr AlignedBuf* aligned_scratch = nullptr;
+
+    s = file_reader->Read(IOOptions(), read_offset, read_size, slice,
+                          buf->get(), aligned_scratch);
+  }
+
+  if (!s.ok()) {
+    return s;
+  }
+
+  if (slice->size() != read_size) {
+    return Status::Corruption("Failed to read data from blob file");
+  }
+
+  return Status::OK();
+}
+
+BlobFileReader::BlobFileReader(
+    std::unique_ptr<RandomAccessFileReader>&& file_reader, uint64_t file_size,
+    CompressionType compression_type)
+    : file_reader_(std::move(file_reader)),
+      file_size_(file_size),
+      compression_type_(compression_type) {
+  assert(file_reader_);
+}
+
+BlobFileReader::~BlobFileReader() = default;
+
+Status BlobFileReader::GetBlob(const ReadOptions& read_options,
+                               const Slice& user_key, uint64_t offset,
+                               uint64_t value_size,
+                               CompressionType compression_type,
+                               PinnableSlice* value) const {
+  assert(value);
+
+  const uint64_t key_size = user_key.size();
+
+  if (!IsValidBlobOffset(offset, key_size, value_size, file_size_)) {
+    return Status::Corruption("Invalid blob offset");
+  }
+
+  if (compression_type != compression_type_) {
+    return Status::Corruption("Compression type mismatch when reading blob");
+  }
+
+  // Note: if verify_checksum is set, we read the entire blob record to be able
+  // to perform the verification; otherwise, we just read the blob itself. Since
+  // the offset in BlobIndex actually points to the blob value, we need to make
+  // an adjustment in the former case.
+  const uint64_t adjustment =
+      read_options.verify_checksums
+          ? BlobLogRecord::CalculateAdjustmentForRecordHeader(key_size)
+          : 0;
+  assert(offset >= adjustment);
+
+  Slice record_slice;
+  Buffer buf;
+  AlignedBuf aligned_buf;
+
+  {
+    TEST_SYNC_POINT("BlobFileReader::GetBlob:ReadFromFile");
+
+    const uint64_t record_offset = offset - adjustment;
+    const uint64_t record_size = value_size + adjustment;
+
+    const Status s = ReadFromFile(file_reader_.get(), record_offset,
+                                  static_cast<size_t>(record_size),
+                                  &record_slice, &buf, &aligned_buf);
+    if (!s.ok()) {
+      return s;
+    }
+
+    TEST_SYNC_POINT_CALLBACK("BlobFileReader::GetBlob:TamperWithResult",
+                             &record_slice);
+  }
+
+  if (read_options.verify_checksums) {
+    const Status s = VerifyBlob(record_slice, user_key, value_size);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  const Slice value_slice(record_slice.data() + adjustment, value_size);
+
+  {
+    const Status s =
+        UncompressBlobIfNeeded(value_slice, compression_type, value);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  return Status::OK();
+}
+
+Status BlobFileReader::VerifyBlob(const Slice& record_slice,
+                                  const Slice& user_key, uint64_t value_size) {
+  BlobLogRecord record;
+
+  const Slice header_slice(record_slice.data(), BlobLogRecord::kHeaderSize);
+
+  {
+    const Status s = record.DecodeHeaderFrom(header_slice);
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  if (record.key_size != user_key.size()) {
+    return Status::Corruption("Key size mismatch when reading blob");
+  }
+
+  if (record.value_size != value_size) {
+    return Status::Corruption("Value size mismatch when reading blob");
+  }
+
+  record.key =
+      Slice(record_slice.data() + BlobLogRecord::kHeaderSize, record.key_size);
+  if (record.key != user_key) {
+    return Status::Corruption("Key mismatch when reading blob");
+  }
+
+  record.value = Slice(record.key.data() + record.key_size, value_size);
+
+  {
+    TEST_SYNC_POINT_CALLBACK("BlobFileReader::VerifyBlob:CheckBlobCRC",
+                             &record);
+
+    const Status s = record.CheckBlobCRC();
+    if (!s.ok()) {
+      return s;
+    }
+  }
+
+  return Status::OK();
+}
+
+Status BlobFileReader::UncompressBlobIfNeeded(const Slice& value_slice,
+                                              CompressionType compression_type,
+                                              PinnableSlice* value) {
+  assert(value);
+
+  if (compression_type == kNoCompression) {
+    SaveValue(value_slice, value);
+
+    return Status::OK();
+  }
+
+  UncompressionContext context(compression_type);
+  UncompressionInfo info(context, UncompressionDict::GetEmptyDict(),
+                         compression_type);
+
+  size_t uncompressed_size = 0;
+  constexpr uint32_t compression_format_version = 2;
+  constexpr MemoryAllocator* allocator = nullptr;
+
+  CacheAllocationPtr output =
+      UncompressData(info, value_slice.data(), value_slice.size(),
+                     &uncompressed_size, compression_format_version, allocator);
+
+  TEST_SYNC_POINT_CALLBACK(
+      "BlobFileReader::UncompressBlobIfNeeded:TamperWithResult", &output);
+
+  if (!output) {
+    return Status::Corruption("Unable to uncompress blob");
+  }
+
+  SaveValue(Slice(output.get(), uncompressed_size), value);
+
+  return Status::OK();
+}
+
+void BlobFileReader::SaveValue(const Slice& src, PinnableSlice* dst) {
+  assert(dst);
+
+  if (dst->IsPinned()) {
+    dst->Reset();
+  }
+
+  dst->PinSelf(src);
+}
+
+}  // namespace ROCKSDB_NAMESPACE

db/blob/blob_file_reader.h

@@ -0,0 +1,81 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#pragma once
+
+#include <cinttypes>
+#include <memory>
+
+#include "file/random_access_file_reader.h"
+#include "rocksdb/compression_type.h"
+#include "rocksdb/rocksdb_namespace.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+class Status;
+struct ImmutableCFOptions;
+struct FileOptions;
+class HistogramImpl;
+struct ReadOptions;
+class Slice;
+class PinnableSlice;
+
+class BlobFileReader {
+ public:
+  static Status Create(const ImmutableCFOptions& immutable_cf_options,
+                       const FileOptions& file_options,
+                       uint32_t column_family_id,
+                       HistogramImpl* blob_file_read_hist,
+                       uint64_t blob_file_number,
+                       std::unique_ptr<BlobFileReader>* reader);
+
+  BlobFileReader(const BlobFileReader&) = delete;
+  BlobFileReader& operator=(const BlobFileReader&) = delete;
+
+  ~BlobFileReader();
+
+  Status GetBlob(const ReadOptions& read_options, const Slice& user_key,
+                 uint64_t offset, uint64_t value_size,
+                 CompressionType compression_type, PinnableSlice* value) const;
+
+ private:
+  BlobFileReader(std::unique_ptr<RandomAccessFileReader>&& file_reader,
+                 uint64_t file_size, CompressionType compression_type);
+
+  static Status OpenFile(const ImmutableCFOptions& immutable_cf_options,
+                         const FileOptions& file_opts,
+                         HistogramImpl* blob_file_read_hist,
+                         uint64_t blob_file_number, uint64_t* file_size,
+                         std::unique_ptr<RandomAccessFileReader>* file_reader);
+
+  static Status ReadHeader(const RandomAccessFileReader* file_reader,
+                           uint32_t column_family_id,
+                           CompressionType* compression_type);
+
+  static Status ReadFooter(uint64_t file_size,
+                           const RandomAccessFileReader* file_reader);
+
+  using Buffer = std::unique_ptr<char[]>;
+
+  static Status ReadFromFile(const RandomAccessFileReader* file_reader,
+                             uint64_t read_offset, size_t read_size,
+                             Slice* slice, Buffer* buf,
+                             AlignedBuf* aligned_buf);
+
+  static Status VerifyBlob(const Slice& record_slice, const Slice& user_key,
+                           uint64_t value_size);
+
+  static Status UncompressBlobIfNeeded(const Slice& value_slice,
+                                       CompressionType compression_type,
+                                       PinnableSlice* value);
+
+  static void SaveValue(const Slice& src, PinnableSlice* dst);
+
+  std::unique_ptr<RandomAccessFileReader> file_reader_;
+  uint64_t file_size_;
+  CompressionType compression_type_;
+};
+
+}  // namespace ROCKSDB_NAMESPACE

db/blob/blob_file_reader_test.cc

@@ -0,0 +1,770 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include "db/blob/blob_file_reader.h"
+
+#include <cassert>
+#include <string>
+
+#include "db/blob/blob_log_format.h"
+#include "db/blob/blob_log_writer.h"
+#include "env/mock_env.h"
+#include "file/filename.h"
+#include "file/read_write_util.h"
+#include "file/writable_file_writer.h"
+#include "options/cf_options.h"
+#include "rocksdb/env.h"
+#include "rocksdb/file_system.h"
+#include "rocksdb/options.h"
+#include "test_util/testharness.h"
+#include "util/compression.h"
+#include "utilities/fault_injection_env.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+namespace {
+
+// Creates a test blob file with a single blob in it. Note: this method
+// makes it possible to test various corner cases by allowing the caller
+// to specify the contents of various blob file header/footer fields.
+void WriteBlobFile(const ImmutableCFOptions& immutable_cf_options,
+                   uint32_t column_family_id, bool has_ttl,
+                   const ExpirationRange& expiration_range_header,
+                   const ExpirationRange& expiration_range_footer,
+                   uint64_t blob_file_number, const Slice& key,
+                   const Slice& blob, CompressionType compression_type,
+                   uint64_t* blob_offset, uint64_t* blob_size) {
+  assert(!immutable_cf_options.cf_paths.empty());
+  assert(blob_offset);
+  assert(blob_size);
+
+  const std::string blob_file_path = BlobFileName(
+      immutable_cf_options.cf_paths.front().path, blob_file_number);
+
+  std::unique_ptr<FSWritableFile> file;
+  ASSERT_OK(NewWritableFile(immutable_cf_options.fs, blob_file_path, &file,
+                            FileOptions()));
+
+  std::unique_ptr<WritableFileWriter> file_writer(
+      new WritableFileWriter(std::move(file), blob_file_path, FileOptions(),
+                             immutable_cf_options.env));
+
+  constexpr Statistics* statistics = nullptr;
+  constexpr bool use_fsync = false;
+
+  BlobLogWriter blob_log_writer(std::move(file_writer),
+                                immutable_cf_options.env, statistics,
+                                blob_file_number, use_fsync);
+
+  BlobLogHeader header(column_family_id, compression_type, has_ttl,
+                       expiration_range_header);
+
+  ASSERT_OK(blob_log_writer.WriteHeader(header));
+
+  std::string compressed_blob;
+  Slice blob_to_write;
+
+  if (compression_type == kNoCompression) {
+    blob_to_write = blob;
+    *blob_size = blob.size();
+  } else {
+    CompressionOptions opts;
+    CompressionContext context(compression_type);
+    constexpr uint64_t sample_for_compression = 0;
+
+    CompressionInfo info(opts, context, CompressionDict::GetEmptyDict(),
+                         compression_type, sample_for_compression);
+
+    constexpr uint32_t compression_format_version = 2;
+
+    ASSERT_TRUE(
+        CompressData(blob, info, compression_format_version, &compressed_blob));
+
+    blob_to_write = compressed_blob;
+    *blob_size = compressed_blob.size();
+  }
+
+  uint64_t key_offset = 0;
+
+  ASSERT_OK(
+      blob_log_writer.AddRecord(key, blob_to_write, &key_offset, blob_offset));
+
+  BlobLogFooter footer;
+  footer.blob_count = 1;
+  footer.expiration_range = expiration_range_footer;
+
+  std::string checksum_method;
+  std::string checksum_value;
+
+  ASSERT_OK(
+      blob_log_writer.AppendFooter(footer, &checksum_method, &checksum_value));
+}
+
+}  // anonymous namespace
+
+class BlobFileReaderTest : public testing::Test {
+ protected:
+  BlobFileReaderTest() : mock_env_(Env::Default()) {}
+
+  MockEnv mock_env_;
+};
+
+TEST_F(BlobFileReaderTest, CreateReaderAndGetBlob) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileReaderTest_CreateReaderAndGetBlob"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kNoCompression, &blob_offset, &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_OK(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                   column_family_id, blob_file_read_hist,
+                                   blob_file_number, &reader));
+
+  // Make sure the blob can be retrieved with and without checksum verification
+  ReadOptions read_options;
+  read_options.verify_checksums = false;
+
+  {
+    PinnableSlice value;
+
+    ASSERT_OK(reader->GetBlob(read_options, key, blob_offset, blob_size,
+                              kNoCompression, &value));
+    ASSERT_EQ(value, blob);
+  }
+
+  read_options.verify_checksums = true;
+
+  {
+    PinnableSlice value;
+
+    ASSERT_OK(reader->GetBlob(read_options, key, blob_offset, blob_size,
+                              kNoCompression, &value));
+    ASSERT_EQ(value, blob);
+  }
+
+  // Invalid offset (too close to start of file)
+  {
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(read_options, key, blob_offset - 1, blob_size,
+                              kNoCompression, &value)
+                    .IsCorruption());
+  }
+
+  // Invalid offset (too close to end of file)
+  {
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(read_options, key, blob_offset + 1, blob_size,
+                              kNoCompression, &value)
+                    .IsCorruption());
+  }
+
+  // Incorrect compression type
+  {
+    PinnableSlice value;
+
+    ASSERT_TRUE(
+        reader
+            ->GetBlob(read_options, key, blob_offset, blob_size, kZSTD, &value)
+            .IsCorruption());
+  }
+
+  // Incorrect key size
+  {
+    constexpr char shorter_key[] = "k";
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(read_options, shorter_key,
+                              blob_offset - (sizeof(key) - sizeof(shorter_key)),
+                              blob_size, kNoCompression, &value)
+                    .IsCorruption());
+  }
+
+  // Incorrect key
+  {
+    constexpr char incorrect_key[] = "foo";
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(read_options, incorrect_key, blob_offset,
+                              blob_size, kNoCompression, &value)
+                    .IsCorruption());
+  }
+
+  // Incorrect value size
+  {
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(read_options, key, blob_offset, blob_size + 1,
+                              kNoCompression, &value)
+                    .IsCorruption());
+  }
+}
+
+TEST_F(BlobFileReaderTest, Malformed) {
+  // Write a blob file consisting of nothing but a header, and make sure we
+  // detect the error when we open it for reading
+
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileReaderTest_Malformed"), 0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr uint64_t blob_file_number = 1;
+
+  {
+    constexpr bool has_ttl = false;
+    constexpr ExpirationRange expiration_range;
+
+    const std::string blob_file_path = BlobFileName(
+        immutable_cf_options.cf_paths.front().path, blob_file_number);
+
+    std::unique_ptr<FSWritableFile> file;
+    ASSERT_OK(NewWritableFile(immutable_cf_options.fs, blob_file_path, &file,
+                              FileOptions()));
+
+    std::unique_ptr<WritableFileWriter> file_writer(
+        new WritableFileWriter(std::move(file), blob_file_path, FileOptions(),
+                               immutable_cf_options.env));
+
+    constexpr Statistics* statistics = nullptr;
+    constexpr bool use_fsync = false;
+
+    BlobLogWriter blob_log_writer(std::move(file_writer),
+                                  immutable_cf_options.env, statistics,
+                                  blob_file_number, use_fsync);
+
+    BlobLogHeader header(column_family_id, kNoCompression, has_ttl,
+                         expiration_range);
+
+    ASSERT_OK(blob_log_writer.WriteHeader(header));
+  }
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_TRUE(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                     column_family_id, blob_file_read_hist,
+                                     blob_file_number, &reader)
+                  .IsCorruption());
+}
+
+TEST_F(BlobFileReaderTest, TTL) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileReaderTest_TTL"), 0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = true;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kNoCompression, &blob_offset, &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_TRUE(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                     column_family_id, blob_file_read_hist,
+                                     blob_file_number, &reader)
+                  .IsCorruption());
+}
+
+TEST_F(BlobFileReaderTest, ExpirationRangeInHeader) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileReaderTest_ExpirationRangeInHeader"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  const ExpirationRange expiration_range_header(
+      1, 2);  // can be made constexpr when we adopt C++14
+  constexpr ExpirationRange expiration_range_footer;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range_header, expiration_range_footer,
+                blob_file_number, key, blob, kNoCompression, &blob_offset,
+                &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_TRUE(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                     column_family_id, blob_file_read_hist,
+                                     blob_file_number, &reader)
+                  .IsCorruption());
+}
+
+TEST_F(BlobFileReaderTest, ExpirationRangeInFooter) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileReaderTest_ExpirationRangeInFooter"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range_header;
+  const ExpirationRange expiration_range_footer(
+      1, 2);  // can be made constexpr when we adopt C++14
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range_header, expiration_range_footer,
+                blob_file_number, key, blob, kNoCompression, &blob_offset,
+                &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_TRUE(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                     column_family_id, blob_file_read_hist,
+                                     blob_file_number, &reader)
+                  .IsCorruption());
+}
+
+TEST_F(BlobFileReaderTest, IncorrectColumnFamily) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileReaderTest_IncorrectColumnFamily"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kNoCompression, &blob_offset, &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  constexpr uint32_t incorrect_column_family_id = 2;
+
+  ASSERT_TRUE(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                     incorrect_column_family_id,
+                                     blob_file_read_hist, blob_file_number,
+                                     &reader)
+                  .IsCorruption());
+}
+
+TEST_F(BlobFileReaderTest, BlobCRCError) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileReaderTest_BlobCRCError"), 0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kNoCompression, &blob_offset, &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_OK(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                   column_family_id, blob_file_read_hist,
+                                   blob_file_number, &reader));
+
+  SyncPoint::GetInstance()->SetCallBack(
+      "BlobFileReader::VerifyBlob:CheckBlobCRC", [](void* arg) {
+        BlobLogRecord* const record = static_cast<BlobLogRecord*>(arg);
+        assert(record);
+
+        record->blob_crc = 0xfaceb00c;
+      });
+
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  PinnableSlice value;
+
+  ASSERT_TRUE(reader
+                  ->GetBlob(ReadOptions(), key, blob_offset, blob_size,
+                            kNoCompression, &value)
+                  .IsCorruption());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+TEST_F(BlobFileReaderTest, Compression) {
+  if (!Snappy_Supported()) {
+    return;
+  }
+
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileReaderTest_Compression"), 0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kSnappyCompression, &blob_offset, &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_OK(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                   column_family_id, blob_file_read_hist,
+                                   blob_file_number, &reader));
+
+  // Make sure the blob can be retrieved with and without checksum verification
+  ReadOptions read_options;
+  read_options.verify_checksums = false;
+
+  {
+    PinnableSlice value;
+
+    ASSERT_OK(reader->GetBlob(read_options, key, blob_offset, blob_size,
+                              kSnappyCompression, &value));
+    ASSERT_EQ(value, blob);
+  }
+
+  read_options.verify_checksums = true;
+
+  {
+    PinnableSlice value;
+
+    ASSERT_OK(reader->GetBlob(read_options, key, blob_offset, blob_size,
+                              kSnappyCompression, &value));
+    ASSERT_EQ(value, blob);
+  }
+}
+
+TEST_F(BlobFileReaderTest, UncompressionError) {
+  if (!Snappy_Supported()) {
+    return;
+  }
+
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileReaderTest_UncompressionError"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kSnappyCompression, &blob_offset, &blob_size);
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  ASSERT_OK(BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                   column_family_id, blob_file_read_hist,
+                                   blob_file_number, &reader));
+
+  SyncPoint::GetInstance()->SetCallBack(
+      "BlobFileReader::UncompressBlobIfNeeded:TamperWithResult", [](void* arg) {
+        CacheAllocationPtr* const output =
+            static_cast<CacheAllocationPtr*>(arg);
+        assert(output);
+
+        output->reset();
+      });
+
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  PinnableSlice value;
+
+  ASSERT_TRUE(reader
+                  ->GetBlob(ReadOptions(), key, blob_offset, blob_size,
+                            kSnappyCompression, &value)
+                  .IsCorruption());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+class BlobFileReaderIOErrorTest
+    : public testing::Test,
+      public testing::WithParamInterface<std::string> {
+ protected:
+  BlobFileReaderIOErrorTest()
+      : mock_env_(Env::Default()),
+        fault_injection_env_(&mock_env_),
+        sync_point_(GetParam()) {}
+
+  MockEnv mock_env_;
+  FaultInjectionTestEnv fault_injection_env_;
+  std::string sync_point_;
+};
+
+INSTANTIATE_TEST_CASE_P(BlobFileReaderTest, BlobFileReaderIOErrorTest,
+                        ::testing::ValuesIn(std::vector<std::string>{
+                            "BlobFileReader::OpenFile:GetFileSize",
+                            "BlobFileReader::OpenFile:NewRandomAccessFile",
+                            "BlobFileReader::ReadHeader:ReadFromFile",
+                            "BlobFileReader::ReadFooter:ReadFromFile",
+                            "BlobFileReader::GetBlob:ReadFromFile"}));
+
+TEST_P(BlobFileReaderIOErrorTest, IOError) {
+  // Simulates an I/O error during the specified step
+
+  Options options;
+  options.env = &fault_injection_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&fault_injection_env_,
+                            "BlobFileReaderIOErrorTest_IOError"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kNoCompression, &blob_offset, &blob_size);
+
+  SyncPoint::GetInstance()->SetCallBack(sync_point_, [this](void* /* arg */) {
+    fault_injection_env_.SetFilesystemActive(false,
+                                             Status::IOError(sync_point_));
+  });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  const Status s = BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                          column_family_id, blob_file_read_hist,
+                                          blob_file_number, &reader);
+
+  const bool fail_during_create =
+      (sync_point_ != "BlobFileReader::GetBlob:ReadFromFile");
+
+  if (fail_during_create) {
+    ASSERT_TRUE(s.IsIOError());
+  } else {
+    ASSERT_OK(s);
+
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(ReadOptions(), key, blob_offset, blob_size,
+                              kNoCompression, &value)
+                    .IsIOError());
+  }
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+class BlobFileReaderDecodingErrorTest
+    : public testing::Test,
+      public testing::WithParamInterface<std::string> {
+ protected:
+  BlobFileReaderDecodingErrorTest()
+      : mock_env_(Env::Default()), sync_point_(GetParam()) {}
+
+  MockEnv mock_env_;
+  std::string sync_point_;
+};
+
+INSTANTIATE_TEST_CASE_P(BlobFileReaderTest, BlobFileReaderDecodingErrorTest,
+                        ::testing::ValuesIn(std::vector<std::string>{
+                            "BlobFileReader::ReadHeader:TamperWithResult",
+                            "BlobFileReader::ReadFooter:TamperWithResult",
+                            "BlobFileReader::GetBlob:TamperWithResult"}));
+
+TEST_P(BlobFileReaderDecodingErrorTest, DecodingError) {
+  Options options;
+  options.env = &mock_env_;
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileReaderDecodingErrorTest_DecodingError"),
+      0);
+  options.enable_blob_files = true;
+
+  ImmutableCFOptions immutable_cf_options(options);
+
+  constexpr uint32_t column_family_id = 1;
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+  constexpr uint64_t blob_file_number = 1;
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  uint64_t blob_offset = 0;
+  uint64_t blob_size = 0;
+
+  WriteBlobFile(immutable_cf_options, column_family_id, has_ttl,
+                expiration_range, expiration_range, blob_file_number, key, blob,
+                kNoCompression, &blob_offset, &blob_size);
+
+  SyncPoint::GetInstance()->SetCallBack(sync_point_, [](void* arg) {
+    Slice* const slice = static_cast<Slice*>(arg);
+    assert(slice);
+    assert(!slice->empty());
+
+    slice->remove_prefix(1);
+  });
+
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  const Status s = BlobFileReader::Create(immutable_cf_options, FileOptions(),
+                                          column_family_id, blob_file_read_hist,
+                                          blob_file_number, &reader);
+
+  const bool fail_during_create =
+      sync_point_ != "BlobFileReader::GetBlob:TamperWithResult";
+
+  if (fail_during_create) {
+    ASSERT_TRUE(s.IsCorruption());
+  } else {
+    ASSERT_OK(s);
+
+    PinnableSlice value;
+
+    ASSERT_TRUE(reader
+                    ->GetBlob(ReadOptions(), key, blob_offset, blob_size,
+                              kNoCompression, &value)
+                    .IsCorruption());
+  }
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+}  // namespace ROCKSDB_NAMESPACE
+
+int main(int argc, char** argv) {
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

db/blob/blob_index.h

@@ -3,13 +3,13 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 #pragma once
-#ifndef ROCKSDB_LITE
 
 #include <sstream>
 #include <string>
 
-#include "rocksdb/options.h"
+#include "rocksdb/compression_type.h"
 #include "util/coding.h"
+#include "util/compression.h"
 #include "util/string_util.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -83,6 +83,11 @@ class BlobIndex {
     return size_;
   }
 
+  CompressionType compression() const {
+    assert(!IsInlined());
+    return compression_;
+  }
+
   Status DecodeFrom(Slice slice) {
     static const std::string kErrorMessage = "Error while decoding blob index";
     assert(slice.size() > 0);
@@ -118,7 +123,8 @@ class BlobIndex {
       oss << "[inlined blob] value:" << value_.ToString(output_hex);
     } else {
       oss << "[blob ref] file:" << file_number_ << " offset:" << offset_
-          << " size:" << size_;
+          << " size:" << size_
+          << " compression: " << CompressionTypeToString(compression_);
     }
 
     if (HasTTL()) {
@@ -176,4 +182,3 @@ class BlobIndex {
 };
 
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/blob_log_format.cc

@@ -3,7 +3,6 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 //
-#ifndef ROCKSDB_LITE
 
 #include "db/blob/blob_log_format.h"
 
@@ -11,7 +10,6 @@
 #include "util/crc32c.h"
 
 namespace ROCKSDB_NAMESPACE {
-namespace blob_db {
 
 void BlobLogHeader::EncodeTo(std::string* dst) {
   assert(dst != nullptr);
@@ -97,6 +95,10 @@ Status BlobLogFooter::DecodeFrom(Slice src) {
   return Status::OK();
 }
 
+uint64_t BlobLogRecord::CalculateAdjustmentForRecordHeader(uint64_t key_size) {
+  return key_size + kHeaderSize;
+}
+
 void BlobLogRecord::EncodeHeaderTo(std::string* dst) {
   assert(dst != nullptr);
   dst->clear();
@@ -144,6 +146,4 @@ Status BlobLogRecord::CheckBlobCRC() const {
   return Status::OK();
 }
 
-}  // namespace blob_db
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/blob_log_format.h

@@ -7,8 +7,6 @@
 
 #pragma once
 
-#ifndef ROCKSDB_LITE
-
 #include <memory>
 #include <utility>
 
@@ -18,7 +16,6 @@
 #include "rocksdb/types.h"
 
 namespace ROCKSDB_NAMESPACE {
-namespace blob_db {
 
 constexpr uint32_t kMagicNumber = 2395959;  // 0x00248f37
 constexpr uint32_t kVersion1 = 1;
@@ -107,6 +104,11 @@ struct BlobLogRecord {
   // header include fields up to blob CRC
   static constexpr size_t kHeaderSize = 32;
 
+  // Note that the offset field of BlobIndex actually points to the blob value
+  // as opposed to the start of the blob record. The following method can
+  // be used to calculate the adjustment needed to read the blob record header.
+  static uint64_t CalculateAdjustmentForRecordHeader(uint64_t key_size);
+
   uint64_t key_size = 0;
   uint64_t value_size = 0;
   uint64_t expiration = 0;
@@ -126,6 +128,19 @@ struct BlobLogRecord {
   Status CheckBlobCRC() const;
 };
 
-}  // namespace blob_db
+// Checks whether a blob offset is potentially valid or not.
+inline bool IsValidBlobOffset(uint64_t value_offset, uint64_t key_size,
+                              uint64_t value_size, uint64_t file_size) {
+  if (value_offset <
+      BlobLogHeader::kSize + BlobLogRecord::kHeaderSize + key_size) {
+    return false;
+  }
+
+  if (value_offset + value_size + BlobLogFooter::kSize > file_size) {
+    return false;
+  }
+
+  return true;
+}
+
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/blob_log_sequential_reader.cc

@@ -3,9 +3,8 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 //
-#ifndef ROCKSDB_LITE
 
-#include "db/blob/blob_log_reader.h"
+#include "db/blob/blob_log_sequential_reader.h"
 
 #include <algorithm>
 
@@ -14,17 +13,22 @@
 #include "util/stop_watch.h"
 
 namespace ROCKSDB_NAMESPACE {
-namespace blob_db {
 
-Reader::Reader(std::unique_ptr<RandomAccessFileReader>&& file_reader, Env* env,
-               Statistics* statistics)
+BlobLogSequentialReader::BlobLogSequentialReader(
+    std::unique_ptr<RandomAccessFileReader>&& file_reader, Env* env,
+    Statistics* statistics)
     : file_(std::move(file_reader)),
       env_(env),
       statistics_(statistics),
-      buffer_(),
       next_byte_(0) {}
 
-Status Reader::ReadSlice(uint64_t size, Slice* slice, char* buf) {
+BlobLogSequentialReader::~BlobLogSequentialReader() = default;
+
+Status BlobLogSequentialReader::ReadSlice(uint64_t size, Slice* slice,
+                                          char* buf) {
+  assert(slice);
+  assert(file_);
+
   StopWatch read_sw(env_, statistics_, BLOB_DB_BLOB_FILE_READ_MICROS);
   Status s = file_->Read(IOOptions(), next_byte_, static_cast<size_t>(size),
                          slice, buf, nullptr);
@@ -39,9 +43,13 @@ Status Reader::ReadSlice(uint64_t size, Slice* slice, char* buf) {
   return s;
 }
 
-Status Reader::ReadHeader(BlobLogHeader* header) {
-  assert(file_.get() != nullptr);
+Status BlobLogSequentialReader::ReadHeader(BlobLogHeader* header) {
+  assert(header);
   assert(next_byte_ == 0);
+
+  static_assert(BlobLogHeader::kSize <= sizeof(header_buf_),
+                "Buffer is smaller than BlobLogHeader::kSize");
+
   Status s = ReadSlice(BlobLogHeader::kSize, &buffer_, header_buf_);
   if (!s.ok()) {
     return s;
@@ -54,8 +62,13 @@ Status Reader::ReadHeader(BlobLogHeader* header) {
   return header->DecodeFrom(buffer_);
 }
 
-Status Reader::ReadRecord(BlobLogRecord* record, ReadLevel level,
-                          uint64_t* blob_offset) {
+Status BlobLogSequentialReader::ReadRecord(BlobLogRecord* record,
+                                           ReadLevel level,
+                                           uint64_t* blob_offset) {
+  assert(record);
+  static_assert(BlobLogRecord::kHeaderSize <= sizeof(header_buf_),
+                "Buffer is smaller than BlobLogRecord::kHeaderSize");
+
   Status s = ReadSlice(BlobLogRecord::kHeaderSize, &buffer_, header_buf_);
   if (!s.ok()) {
     return s;
@@ -101,6 +114,21 @@ Status Reader::ReadRecord(BlobLogRecord* record, ReadLevel level,
   return s;
 }
 
-}  // namespace blob_db
+Status BlobLogSequentialReader::ReadFooter(BlobLogFooter* footer) {
+  assert(footer);
+  static_assert(BlobLogFooter::kSize <= sizeof(header_buf_),
+                "Buffer is smaller than BlobLogFooter::kSize");
+
+  Status s = ReadSlice(BlobLogFooter::kSize, &buffer_, header_buf_);
+  if (!s.ok()) {
+    return s;
+  }
+
+  if (buffer_.size() != BlobLogFooter::kSize) {
+    return Status::Corruption("EOF reached before file footer");
+  }
+
+  return footer->DecodeFrom(buffer_);
+}
+
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/blob_log_sequential_reader.h

@@ -5,32 +5,27 @@
 //
 #pragma once
 
-#ifndef ROCKSDB_LITE
-
 #include <memory>
-#include <string>
 
 #include "db/blob/blob_log_format.h"
-#include "file/random_access_file_reader.h"
-#include "rocksdb/env.h"
 #include "rocksdb/slice.h"
-#include "rocksdb/statistics.h"
-#include "rocksdb/status.h"
 
 namespace ROCKSDB_NAMESPACE {
 
-class SequentialFileReader;
-class Logger;
-
-namespace blob_db {
+class RandomAccessFileReader;
+class Env;
+class Statistics;
+class Status;
 
 /**
- * Reader is a general purpose log stream reader implementation. The actual job
- * of reading from the device is implemented by the SequentialFile interface.
+ * BlobLogSequentialReader is a general purpose log stream reader
+ * implementation. The actual job of reading from the device is implemented by
+ * the RandomAccessFileReader interface.
  *
- * Please see Writer for details on the file and record layout.
+ * Please see BlobLogWriter for details on the file and record layout.
  */
-class Reader {
+
+class BlobLogSequentialReader {
  public:
   enum ReadLevel {
     kReadHeader,
@@ -38,27 +33,28 @@ class Reader {
     kReadHeaderKeyBlob,
   };
 
-  // Create a reader that will return log records from "*file".
-  // "*file" must remain live while this Reader is in use.
-  Reader(std::unique_ptr<RandomAccessFileReader>&& file_reader, Env* env,
-         Statistics* statistics);
-  // No copying allowed
-  Reader(const Reader&) = delete;
-  Reader& operator=(const Reader&) = delete;
+  // Create a reader that will return log records from "*file_reader".
+  BlobLogSequentialReader(std::unique_ptr<RandomAccessFileReader>&& file_reader,
+                          Env* env, Statistics* statistics);
 
-  ~Reader() = default;
+  // No copying allowed
+  BlobLogSequentialReader(const BlobLogSequentialReader&) = delete;
+  BlobLogSequentialReader& operator=(const BlobLogSequentialReader&) = delete;
+
+  ~BlobLogSequentialReader();
 
   Status ReadHeader(BlobLogHeader* header);
 
   // Read the next record into *record.  Returns true if read
-  // successfully, false if we hit end of the input.  May use
-  // "*scratch" as temporary storage.  The contents filled in *record
-  // will only be valid until the next mutating operation on this
-  // reader or the next mutation to *scratch.
+  // successfully, false if we hit end of the input. The contents filled in
+  // *record will only be valid until the next mutating operation on this
+  // reader.
   // If blob_offset is non-null, return offset of the blob through it.
   Status ReadRecord(BlobLogRecord* record, ReadLevel level = kReadHeader,
                     uint64_t* blob_offset = nullptr);
 
+  Status ReadFooter(BlobLogFooter* footer);
+
   void ResetNextByte() { next_byte_ = 0; }
 
   uint64_t GetNextByte() const { return next_byte_; }
@@ -73,10 +69,8 @@ class Reader {
   Slice buffer_;
   char header_buf_[BlobLogRecord::kHeaderSize];
 
-  // which byte to read next. For asserting proper usage
+  // which byte to read next
   uint64_t next_byte_;
 };
 
-}  // namespace blob_db
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/blob_log_writer.cc

@@ -2,7 +2,6 @@
 //  This source code is licensed under both the GPLv2 (found in the
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
-#ifndef ROCKSDB_LITE
 
 #include "db/blob/blob_log_writer.h"
 
@@ -13,33 +12,35 @@
 #include "file/writable_file_writer.h"
 #include "monitoring/statistics.h"
 #include "rocksdb/env.h"
+#include "test_util/sync_point.h"
 #include "util/coding.h"
 #include "util/stop_watch.h"
 
 namespace ROCKSDB_NAMESPACE {
-namespace blob_db {
 
-Writer::Writer(std::unique_ptr<WritableFileWriter>&& dest, Env* env,
-               Statistics* statistics, uint64_t log_number, uint64_t bpsync,
-               bool use_fs, uint64_t boffset)
+BlobLogWriter::BlobLogWriter(std::unique_ptr<WritableFileWriter>&& dest,
+                             Env* env, Statistics* statistics,
+                             uint64_t log_number, bool use_fs, uint64_t boffset)
     : dest_(std::move(dest)),
       env_(env),
       statistics_(statistics),
       log_number_(log_number),
       block_offset_(boffset),
-      bytes_per_sync_(bpsync),
-      next_sync_offset_(0),
       use_fsync_(use_fs),
       last_elem_type_(kEtNone) {}
 
-Status Writer::Sync() {
+BlobLogWriter::~BlobLogWriter() = default;
+
+Status BlobLogWriter::Sync() {
+  TEST_SYNC_POINT("BlobLogWriter::Sync");
+
   StopWatch sync_sw(env_, statistics_, BLOB_DB_BLOB_FILE_SYNC_MICROS);
   Status s = dest_->Sync(use_fsync_);
   RecordTick(statistics_, BLOB_DB_BLOB_FILE_SYNCED);
   return s;
 }
 
-Status Writer::WriteHeader(BlobLogHeader& header) {
+Status BlobLogWriter::WriteHeader(BlobLogHeader& header) {
   assert(block_offset_ == 0);
   assert(last_elem_type_ == kEtNone);
   std::string str;
@@ -56,7 +57,9 @@ Status Writer::WriteHeader(BlobLogHeader& header) {
   return s;
 }
 
-Status Writer::AppendFooter(BlobLogFooter& footer) {
+Status BlobLogWriter::AppendFooter(BlobLogFooter& footer,
+                                   std::string* checksum_method,
+                                   std::string* checksum_value) {
   assert(block_offset_ != 0);
   assert(last_elem_type_ == kEtFileHdr || last_elem_type_ == kEtRecord);
 
@@ -66,7 +69,34 @@ Status Writer::AppendFooter(BlobLogFooter& footer) {
   Status s = dest_->Append(Slice(str));
   if (s.ok()) {
     block_offset_ += str.size();
-    s = dest_->Close();
+
+    s = Sync();
+
+    if (s.ok()) {
+      s = dest_->Close();
+
+      if (s.ok()) {
+        assert(!!checksum_method == !!checksum_value);
+
+        if (checksum_method) {
+          assert(checksum_method->empty());
+
+          std::string method = dest_->GetFileChecksumFuncName();
+          if (method != kUnknownFileChecksumFuncName) {
+            *checksum_method = std::move(method);
+          }
+        }
+        if (checksum_value) {
+          assert(checksum_value->empty());
+
+          std::string value = dest_->GetFileChecksum();
+          if (value != kUnknownFileChecksum) {
+            *checksum_value = std::move(value);
+          }
+        }
+      }
+    }
+
     dest_.reset();
   }
 
@@ -76,9 +106,9 @@ Status Writer::AppendFooter(BlobLogFooter& footer) {
   return s;
 }
 
-Status Writer::AddRecord(const Slice& key, const Slice& val,
-                         uint64_t expiration, uint64_t* key_offset,
-                         uint64_t* blob_offset) {
+Status BlobLogWriter::AddRecord(const Slice& key, const Slice& val,
+                                uint64_t expiration, uint64_t* key_offset,
+                                uint64_t* blob_offset) {
   assert(block_offset_ != 0);
   assert(last_elem_type_ == kEtFileHdr || last_elem_type_ == kEtRecord);
 
@@ -89,8 +119,8 @@ Status Writer::AddRecord(const Slice& key, const Slice& val,
   return s;
 }
 
-Status Writer::AddRecord(const Slice& key, const Slice& val,
-                         uint64_t* key_offset, uint64_t* blob_offset) {
+Status BlobLogWriter::AddRecord(const Slice& key, const Slice& val,
+                                uint64_t* key_offset, uint64_t* blob_offset) {
   assert(block_offset_ != 0);
   assert(last_elem_type_ == kEtFileHdr || last_elem_type_ == kEtRecord);
 
@@ -101,8 +131,8 @@ Status Writer::AddRecord(const Slice& key, const Slice& val,
   return s;
 }
 
-void Writer::ConstructBlobHeader(std::string* buf, const Slice& key,
-                                 const Slice& val, uint64_t expiration) {
+void BlobLogWriter::ConstructBlobHeader(std::string* buf, const Slice& key,
+                                        const Slice& val, uint64_t expiration) {
   BlobLogRecord record;
   record.key = key;
   record.value = val;
@@ -110,9 +140,10 @@ void Writer::ConstructBlobHeader(std::string* buf, const Slice& key,
   record.EncodeHeaderTo(buf);
 }
 
-Status Writer::EmitPhysicalRecord(const std::string& headerbuf,
-                                  const Slice& key, const Slice& val,
-                                  uint64_t* key_offset, uint64_t* blob_offset) {
+Status BlobLogWriter::EmitPhysicalRecord(const std::string& headerbuf,
+                                         const Slice& key, const Slice& val,
+                                         uint64_t* key_offset,
+                                         uint64_t* blob_offset) {
   StopWatch write_sw(env_, statistics_, BLOB_DB_BLOB_FILE_WRITE_MICROS);
   Status s = dest_->Append(Slice(headerbuf));
   if (s.ok()) {
@@ -134,6 +165,4 @@ Status Writer::EmitPhysicalRecord(const std::string& headerbuf,
   return s;
 }
 
-}  // namespace blob_db
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/blob_log_writer.h

@@ -4,8 +4,6 @@
 //  (found in the LICENSE.Apache file in the root directory).
 #pragma once
 
-#ifndef ROCKSDB_LITE
-
 #include <cstdint>
 #include <memory>
 #include <string>
@@ -21,29 +19,27 @@ namespace ROCKSDB_NAMESPACE {
 
 class WritableFileWriter;
 
-namespace blob_db {
-
 /**
- * Writer is the blob log stream writer. It provides an append-only
+ * BlobLogWriter is the blob log stream writer. It provides an append-only
  * abstraction for writing blob data.
  *
  *
  * Look at blob_db_format.h to see the details of the record formats.
  */
 
-class Writer {
+class BlobLogWriter {
  public:
   // Create a writer that will append data to "*dest".
   // "*dest" must be initially empty.
-  // "*dest" must remain live while this Writer is in use.
-  Writer(std::unique_ptr<WritableFileWriter>&& dest, Env* env,
-         Statistics* statistics, uint64_t log_number, uint64_t bpsync,
-         bool use_fsync, uint64_t boffset = 0);
+  // "*dest" must remain live while this BlobLogWriter is in use.
+  BlobLogWriter(std::unique_ptr<WritableFileWriter>&& dest, Env* env,
+                Statistics* statistics, uint64_t log_number, bool use_fsync,
+                uint64_t boffset = 0);
   // No copying allowed
-  Writer(const Writer&) = delete;
-  Writer& operator=(const Writer&) = delete;
+  BlobLogWriter(const BlobLogWriter&) = delete;
+  BlobLogWriter& operator=(const BlobLogWriter&) = delete;
 
-  ~Writer() = default;
+  ~BlobLogWriter();
 
   static void ConstructBlobHeader(std::string* buf, const Slice& key,
                                   const Slice& val, uint64_t expiration);
@@ -58,7 +54,8 @@ class Writer {
                             const Slice& val, uint64_t* key_offset,
                             uint64_t* blob_offset);
 
-  Status AppendFooter(BlobLogFooter& footer);
+  Status AppendFooter(BlobLogFooter& footer, std::string* checksum_method,
+                      std::string* checksum_value);
 
   Status WriteHeader(BlobLogHeader& header);
 
@@ -68,20 +65,14 @@ class Writer {
 
   uint64_t get_log_number() const { return log_number_; }
 
-  bool ShouldSync() const { return block_offset_ > next_sync_offset_; }
-
   Status Sync();
 
-  void ResetSyncPointer() { next_sync_offset_ += bytes_per_sync_; }
-
  private:
   std::unique_ptr<WritableFileWriter> dest_;
   Env* env_;
   Statistics* statistics_;
   uint64_t log_number_;
   uint64_t block_offset_;  // Current offset in block
-  uint64_t bytes_per_sync_;
-  uint64_t next_sync_offset_;
   bool use_fsync_;
 
  public:
@@ -89,6 +80,4 @@ class Writer {
   ElemType last_elem_type_;
 };
 
-}  // namespace blob_db
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/db_blob_index_test.cc

@@ -41,12 +41,13 @@ class DBBlobIndexTest : public DBTestBase {
                                        Tier::kImmutableMemtables,
                                        Tier::kL0SstFile, Tier::kLnSstFile};
 
-  DBBlobIndexTest() : DBTestBase("/db_blob_index_test") {}
+  DBBlobIndexTest()
+      : DBTestBase("/db_blob_index_test", /*env_do_fsync=*/true) {}
 
   ColumnFamilyHandle* cfh() { return dbfull()->DefaultColumnFamily(); }
 
   ColumnFamilyData* cfd() {
-    return reinterpret_cast<ColumnFamilyHandleImpl*>(cfh())->cfd();
+    return static_cast_with_check<ColumnFamilyHandleImpl>(cfh())->cfd();
   }
 
   Status PutBlobIndex(WriteBatch* batch, const Slice& key,

db/builder.cc

@@ -13,11 +13,13 @@
 #include <deque>
 #include <vector>
 
+#include "db/blob/blob_file_builder.h"
 #include "db/compaction/compaction_iterator.h"
 #include "db/dbformat.h"
 #include "db/event_helpers.h"
 #include "db/internal_stats.h"
 #include "db/merge_helper.h"
+#include "db/output_validator.h"
 #include "db/range_del_aggregator.h"
 #include "db/table_cache.h"
 #include "db/version_edit.h"
@@ -67,13 +69,14 @@ TableBuilder* NewTableBuilder(
 }
 
 Status BuildTable(
-    const std::string& dbname, Env* env, FileSystem* fs,
+    const std::string& dbname, VersionSet* versions, Env* env, FileSystem* fs,
     const ImmutableCFOptions& ioptions,
     const MutableCFOptions& mutable_cf_options, const FileOptions& file_options,
     TableCache* table_cache, InternalIterator* iter,
     std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
         range_del_iters,
-    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
+    FileMetaData* meta, std::vector<BlobFileAddition>* blob_file_additions,
+    const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
     uint32_t column_family_id, const std::string& column_family_name,
@@ -83,7 +86,8 @@ Status BuildTable(
     uint64_t sample_for_compression, const CompressionOptions& compression_opts,
     bool paranoid_file_checks, InternalStats* internal_stats,
     TableFileCreationReason reason, IOStatus* io_status,
-    EventLogger* event_logger, int job_id, const Env::IOPriority io_priority,
+    const std::shared_ptr<IOTracer>& io_tracer, EventLogger* event_logger,
+    int job_id, const Env::IOPriority io_priority,
     TableProperties* table_properties, int level, const uint64_t creation_time,
     const uint64_t oldest_key_time, Env::WriteLifeTimeHint write_hint,
     const uint64_t file_creation_time, const std::string& db_id,
@@ -93,6 +97,11 @@ Status BuildTable(
          column_family_name.empty());
   // Reports the IOStats for flush for every following bytes.
   const size_t kReportFlushIOStatsEvery = 1048576;
+  OutputValidator output_validator(
+      internal_comparator,
+      /*enable_order_check=*/
+      mutable_cf_options.check_flush_compaction_key_order,
+      /*enable_hash=*/paranoid_file_checks);
   Status s;
   meta->fd.file_size = 0;
   iter->SeekToFirst();
@@ -104,12 +113,14 @@ Status BuildTable(
 
   std::string fname = TableFileName(ioptions.cf_paths, meta->fd.GetNumber(),
                                     meta->fd.GetPathId());
+  std::vector<std::string> blob_file_paths;
+  std::string file_checksum = kUnknownFileChecksum;
+  std::string file_checksum_func_name = kUnknownFileChecksumFuncName;
 #ifndef ROCKSDB_LITE
   EventHelpers::NotifyTableFileCreationStarted(
       ioptions.listeners, dbname, column_family_name, fname, job_id, reason);
 #endif  // !ROCKSDB_LITE
   TableProperties tp;
-
   if (iter->Valid() || !range_del_agg->IsEmpty()) {
     TableBuilder* builder;
     std::unique_ptr<WritableFileWriter> file_writer;
@@ -124,19 +135,26 @@ Status BuildTable(
       bool use_direct_writes = file_options.use_direct_writes;
       TEST_SYNC_POINT_CALLBACK("BuildTable:create_file", &use_direct_writes);
 #endif  // !NDEBUG
-      s = NewWritableFile(fs, fname, &file, file_options);
+      IOStatus io_s = NewWritableFile(fs, fname, &file, file_options);
+      assert(s.ok());
+      s = io_s;
+      if (io_status->ok()) {
+        *io_status = io_s;
+      }
       if (!s.ok()) {
         EventHelpers::LogAndNotifyTableFileCreationFinished(
             event_logger, ioptions.listeners, dbname, column_family_name, fname,
-            job_id, meta->fd, kInvalidBlobFileNumber, tp, reason, s);
+            job_id, meta->fd, kInvalidBlobFileNumber, tp, reason, s,
+            file_checksum, file_checksum_func_name);
         return s;
       }
       file->SetIOPriority(io_priority);
       file->SetWriteLifeTimeHint(write_hint);
 
       file_writer.reset(new WritableFileWriter(
-          std::move(file), fname, file_options, env, ioptions.statistics,
-          ioptions.listeners, ioptions.file_checksum_gen_factory));
+          std::move(file), fname, file_options, env, io_tracer,
+          ioptions.statistics, ioptions.listeners,
+          ioptions.file_checksum_gen_factory));
 
       builder = NewTableBuilder(
           ioptions, mutable_cf_options, internal_comparator,
@@ -153,16 +171,32 @@ Status BuildTable(
                       snapshots.empty() ? 0 : snapshots.back(),
                       snapshot_checker);
 
+    std::unique_ptr<BlobFileBuilder> blob_file_builder(
+        (mutable_cf_options.enable_blob_files && blob_file_additions)
+            ? new BlobFileBuilder(versions, env, fs, &ioptions,
+                                  &mutable_cf_options, &file_options, job_id,
+                                  column_family_id, column_family_name,
+                                  io_priority, write_hint, &blob_file_paths,
+                                  blob_file_additions)
+            : nullptr);
+
     CompactionIterator c_iter(
         iter, internal_comparator.user_comparator(), &merge, kMaxSequenceNumber,
         &snapshots, earliest_write_conflict_snapshot, snapshot_checker, env,
         ShouldReportDetailedTime(env, ioptions.statistics),
-        true /* internal key corruption is not ok */, range_del_agg.get());
+        true /* internal key corruption is not ok */, range_del_agg.get(),
+        blob_file_builder.get(), ioptions.allow_data_in_errors);
+
     c_iter.SeekToFirst();
     for (; c_iter.Valid(); c_iter.Next()) {
       const Slice& key = c_iter.key();
       const Slice& value = c_iter.value();
       const ParsedInternalKey& ikey = c_iter.ikey();
+      // Generate a rolling 64-bit hash of the key and values
+      s = output_validator.Add(key, value);
+      if (!s.ok()) {
+        break;
+      }
       builder->Add(key, value);
       meta->UpdateBoundaries(key, value, ikey.sequence, ikey.type);
 
@@ -173,27 +207,37 @@ Status BuildTable(
             ThreadStatus::FLUSH_BYTES_WRITTEN, IOSTATS(bytes_written));
       }
     }
+    if (!s.ok()) {
+      c_iter.status().PermitUncheckedError();
+    } else if (!c_iter.status().ok()) {
+      s = c_iter.status();
+    }
+    if (s.ok()) {
+      auto range_del_it = range_del_agg->NewIterator();
+      for (range_del_it->SeekToFirst(); range_del_it->Valid();
+           range_del_it->Next()) {
+        auto tombstone = range_del_it->Tombstone();
+        auto kv = tombstone.Serialize();
+        builder->Add(kv.first.Encode(), kv.second);
+        meta->UpdateBoundariesForRange(kv.first, tombstone.SerializeEndKey(),
+                                       tombstone.seq_, internal_comparator);
+      }
 
-    auto range_del_it = range_del_agg->NewIterator();
-    for (range_del_it->SeekToFirst(); range_del_it->Valid();
-         range_del_it->Next()) {
-      auto tombstone = range_del_it->Tombstone();
-      auto kv = tombstone.Serialize();
-      builder->Add(kv.first.Encode(), kv.second);
-      meta->UpdateBoundariesForRange(kv.first, tombstone.SerializeEndKey(),
-                                     tombstone.seq_, internal_comparator);
+      if (blob_file_builder) {
+        s = blob_file_builder->Finish();
+      }
     }
 
-    // Finish and check for builder errors
-    bool empty = builder->IsEmpty();
-    s = c_iter.status();
     TEST_SYNC_POINT("BuildTable:BeforeFinishBuildTable");
+    const bool empty = builder->IsEmpty();
     if (!s.ok() || empty) {
       builder->Abandon();
     } else {
       s = builder->Finish();
     }
-    *io_status = builder->io_status();
+    if (io_status->ok()) {
+      *io_status = builder->io_status();
+    }
 
     if (s.ok() && !empty) {
       uint64_t file_size = builder->FileSize();
@@ -208,20 +252,24 @@ Status BuildTable(
     delete builder;
 
     // Finish and check for file errors
+    TEST_SYNC_POINT("BuildTable:BeforeSyncTable");
     if (s.ok() && !empty) {
       StopWatch sw(env, ioptions.statistics, TABLE_SYNC_MICROS);
       *io_status = file_writer->Sync(ioptions.use_fsync);
     }
-    if (io_status->ok() && !empty) {
+    TEST_SYNC_POINT("BuildTable:BeforeCloseTableFile");
+    if (s.ok() && io_status->ok() && !empty) {
       *io_status = file_writer->Close();
     }
-    if (io_status->ok() && !empty) {
+    if (s.ok() && io_status->ok() && !empty) {
       // Add the checksum information to file metadata.
       meta->file_checksum = file_writer->GetFileChecksum();
       meta->file_checksum_func_name = file_writer->GetFileChecksumFuncName();
+      file_checksum = meta->file_checksum;
+      file_checksum_func_name = meta->file_checksum_func_name;
     }
 
-    if (!io_status->ok()) {
+    if (s.ok()) {
       s = *io_status;
     }
 
@@ -234,8 +282,9 @@ Status BuildTable(
       // No matter whether use_direct_io_for_flush_and_compaction is true,
       // we will regrad this verification as user reads since the goal is
       // to cache it here for further user reads
+      ReadOptions read_options;
       std::unique_ptr<InternalIterator> it(table_cache->NewIterator(
-          ReadOptions(), file_options, internal_comparator, *meta,
+          read_options, file_options, internal_comparator, *meta,
           nullptr /* range_del_agg */,
           mutable_cf_options.prefix_extractor.get(), nullptr,
           (internal_stats == nullptr) ? nullptr
@@ -248,9 +297,17 @@ Status BuildTable(
           /*allow_unprepared_value*/ false));
       s = it->status();
       if (s.ok() && paranoid_file_checks) {
+        OutputValidator file_validator(internal_comparator,
+                                       /*enable_order_check=*/true,
+                                       /*enable_hash=*/true);
         for (it->SeekToFirst(); it->Valid(); it->Next()) {
+          // Generate a rolling 64-bit hash of the key and values
+          file_validator.Add(it->key(), it->value()).PermitUncheckedError();
         }
         s = it->status();
+        if (s.ok() && !output_validator.CompareValidator(file_validator)) {
+          s = Status::Corruption("Paranoid checksums do not match");
+        }
       }
     }
   }
@@ -261,7 +318,21 @@ Status BuildTable(
   }
 
   if (!s.ok() || meta->fd.GetFileSize() == 0) {
-    fs->DeleteFile(fname, IOOptions(), nullptr);
+    constexpr IODebugContext* dbg = nullptr;
+
+    Status ignored = fs->DeleteFile(fname, IOOptions(), dbg);
+    ignored.PermitUncheckedError();
+
+    assert(blob_file_additions || blob_file_paths.empty());
+
+    if (blob_file_additions) {
+      for (const std::string& blob_file_path : blob_file_paths) {
+        ignored = fs->DeleteFile(blob_file_path, IOOptions(), dbg);
+        ignored.PermitUncheckedError();
+      }
+
+      blob_file_additions->clear();
+    }
   }
 
   if (meta->fd.GetFileSize() == 0) {
@@ -270,7 +341,8 @@ Status BuildTable(
   // Output to event logger and fire events.
   EventHelpers::LogAndNotifyTableFileCreationFinished(
       event_logger, ioptions.listeners, dbname, column_family_name, fname,
-      job_id, meta->fd, meta->oldest_blob_file_number, tp, reason, s);
+      job_id, meta->fd, meta->oldest_blob_file_number, tp, reason, s,
+      file_checksum, file_checksum_func_name);
 
   return s;
 }

db/builder.h

@@ -27,8 +27,10 @@ namespace ROCKSDB_NAMESPACE {
 struct Options;
 struct FileMetaData;
 
+class VersionSet;
 class Env;
 struct EnvOptions;
+class BlobFileAddition;
 class Iterator;
 class SnapshotChecker;
 class TableCache;
@@ -63,13 +65,14 @@ TableBuilder* NewTableBuilder(
 // @param column_family_name Name of the column family that is also identified
 //    by column_family_id, or empty string if unknown.
 extern Status BuildTable(
-    const std::string& dbname, Env* env, FileSystem* fs,
+    const std::string& dbname, VersionSet* versions, Env* env, FileSystem* fs,
     const ImmutableCFOptions& options,
     const MutableCFOptions& mutable_cf_options, const FileOptions& file_options,
     TableCache* table_cache, InternalIterator* iter,
     std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
         range_del_iters,
-    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
+    FileMetaData* meta, std::vector<BlobFileAddition>* blob_file_additions,
+    const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
     uint32_t column_family_id, const std::string& column_family_name,
@@ -79,7 +82,8 @@ extern Status BuildTable(
     const uint64_t sample_for_compression,
     const CompressionOptions& compression_opts, bool paranoid_file_checks,
     InternalStats* internal_stats, TableFileCreationReason reason,
-    IOStatus* io_status, EventLogger* event_logger = nullptr, int job_id = 0,
+    IOStatus* io_status, const std::shared_ptr<IOTracer>& io_tracer,
+    EventLogger* event_logger = nullptr, int job_id = 0,
     const Env::IOPriority io_priority = Env::IO_HIGH,
     TableProperties* table_properties = nullptr, int level = -1,
     const uint64_t creation_time = 0, const uint64_t oldest_key_time = 0,

db/c.cc

@@ -504,13 +504,13 @@ rocksdb_t* rocksdb_open_with_ttl(
   return result;
 }
 
-rocksdb_t* rocksdb_open_for_read_only(
-    const rocksdb_options_t* options,
-    const char* name,
-    unsigned char error_if_log_file_exist,
-    char** errptr) {
+rocksdb_t* rocksdb_open_for_read_only(const rocksdb_options_t* options,
+                                      const char* name,
+                                      unsigned char error_if_wal_file_exists,
+                                      char** errptr) {
   DB* db;
-  if (SaveError(errptr, DB::OpenForReadOnly(options->rep, std::string(name), &db, error_if_log_file_exist))) {
+  if (SaveError(errptr, DB::OpenForReadOnly(options->rep, std::string(name),
+                                            &db, error_if_wal_file_exists))) {
     return nullptr;
   }
   rocksdb_t* result = new rocksdb_t;
@@ -595,6 +595,15 @@ void rocksdb_backup_engine_restore_db_from_latest_backup(
                                                        restore_options->rep));
 }
 
+void rocksdb_backup_engine_restore_db_from_backup(
+    rocksdb_backup_engine_t* be, const char* db_dir, const char* wal_dir,
+    const rocksdb_restore_options_t* restore_options, const uint32_t backup_id,
+    char** errptr) {
+  SaveError(errptr, be->rep->RestoreDBFromBackup(backup_id, std::string(db_dir),
+                                                 std::string(wal_dir),
+                                                 restore_options->rep));
+}
+
 const rocksdb_backup_engine_info_t* rocksdb_backup_engine_get_backup_info(
     rocksdb_backup_engine_t* be) {
   rocksdb_backup_engine_info_t* result = new rocksdb_backup_engine_info_t;
@@ -698,12 +707,47 @@ rocksdb_t* rocksdb_open_column_families(
   return result;
 }
 
+rocksdb_t* rocksdb_open_column_families_with_ttl(
+    const rocksdb_options_t* db_options, const char* name,
+    int num_column_families, const char* const* column_family_names,
+    const rocksdb_options_t* const* column_family_options,
+    rocksdb_column_family_handle_t** column_family_handles, const int* ttls,
+    char** errptr) {
+  std::vector<int32_t> ttls_vec;
+  std::vector<ColumnFamilyDescriptor> column_families;
+  for (int i = 0; i < num_column_families; i++) {
+    ttls_vec.push_back(ttls[i]);
+
+    column_families.push_back(ColumnFamilyDescriptor(
+        std::string(column_family_names[i]),
+        ColumnFamilyOptions(column_family_options[i]->rep)));
+  }
+
+  ROCKSDB_NAMESPACE::DBWithTTL* db;
+  std::vector<ColumnFamilyHandle*> handles;
+  if (SaveError(errptr, ROCKSDB_NAMESPACE::DBWithTTL::Open(
+                            DBOptions(db_options->rep), std::string(name),
+                            column_families, &handles, &db, ttls_vec))) {
+    return nullptr;
+  }
+
+  for (size_t i = 0; i < handles.size(); i++) {
+    rocksdb_column_family_handle_t* c_handle =
+        new rocksdb_column_family_handle_t;
+    c_handle->rep = handles[i];
+    column_family_handles[i] = c_handle;
+  }
+  rocksdb_t* result = new rocksdb_t;
+  result->rep = db;
+  return result;
+}
+
 rocksdb_t* rocksdb_open_for_read_only_column_families(
     const rocksdb_options_t* db_options, const char* name,
     int num_column_families, const char* const* column_family_names,
     const rocksdb_options_t* const* column_family_options,
     rocksdb_column_family_handle_t** column_family_handles,
-    unsigned char error_if_log_file_exist, char** errptr) {
+    unsigned char error_if_wal_file_exists, char** errptr) {
   std::vector<ColumnFamilyDescriptor> column_families;
   for (int i = 0; i < num_column_families; i++) {
     column_families.push_back(ColumnFamilyDescriptor(
@@ -713,8 +757,10 @@ rocksdb_t* rocksdb_open_for_read_only_column_families(
 
   DB* db;
   std::vector<ColumnFamilyHandle*> handles;
-  if (SaveError(errptr, DB::OpenForReadOnly(DBOptions(db_options->rep),
-          std::string(name), column_families, &handles, &db, error_if_log_file_exist))) {
+  if (SaveError(errptr,
+                DB::OpenForReadOnly(DBOptions(db_options->rep),
+                                    std::string(name), column_families,
+                                    &handles, &db, error_if_wal_file_exists))) {
     return nullptr;
   }
 
@@ -796,6 +842,18 @@ rocksdb_column_family_handle_t* rocksdb_create_column_family(
   return handle;
 }
 
+rocksdb_column_family_handle_t* rocksdb_create_column_family_with_ttl(
+    rocksdb_t* db, const rocksdb_options_t* column_family_options,
+    const char* column_family_name, int ttl, char** errptr) {
+  ROCKSDB_NAMESPACE::DBWithTTL* db_with_ttl =
+      static_cast<ROCKSDB_NAMESPACE::DBWithTTL*>(db->rep);
+  rocksdb_column_family_handle_t* handle = new rocksdb_column_family_handle_t;
+  SaveError(errptr, db_with_ttl->CreateColumnFamilyWithTtl(
+                        ColumnFamilyOptions(column_family_options->rep),
+                        std::string(column_family_name), &(handle->rep), ttl));
+  return handle;
+}
+
 void rocksdb_drop_column_family(
     rocksdb_t* db,
     rocksdb_column_family_handle_t* handle,
@@ -996,6 +1054,55 @@ void rocksdb_multi_get_cf(
   }
 }
 
+unsigned char rocksdb_key_may_exist(rocksdb_t* db,
+                                    const rocksdb_readoptions_t* options,
+                                    const char* key, size_t key_len,
+                                    char** value, size_t* val_len,
+                                    const char* timestamp, size_t timestamp_len,
+                                    unsigned char* value_found) {
+  std::string tmp;
+  std::string time;
+  if (timestamp) {
+    time.assign(timestamp, timestamp_len);
+  }
+  bool found = false;
+  const bool result = db->rep->KeyMayExist(options->rep, Slice(key, key_len),
+                                           &tmp, timestamp ? &time : nullptr,
+                                           value_found ? &found : nullptr);
+  if (value_found) {
+    *value_found = found;
+    if (found) {
+      *val_len = tmp.size();
+      *value = CopyString(tmp);
+    }
+  }
+  return result;
+}
+
+unsigned char rocksdb_key_may_exist_cf(
+    rocksdb_t* db, const rocksdb_readoptions_t* options,
+    rocksdb_column_family_handle_t* column_family, const char* key,
+    size_t key_len, char** value, size_t* val_len, const char* timestamp,
+    size_t timestamp_len, unsigned char* value_found) {
+  std::string tmp;
+  std::string time;
+  if (timestamp) {
+    time.assign(timestamp, timestamp_len);
+  }
+  bool found = false;
+  const bool result = db->rep->KeyMayExist(
+      options->rep, column_family->rep, Slice(key, key_len), &tmp,
+      timestamp ? &time : nullptr, value_found ? &found : nullptr);
+  if (value_found) {
+    *value_found = found;
+    if (found) {
+      *val_len = tmp.size();
+      *value = CopyString(tmp);
+    }
+  }
+  return result;
+}
+
 rocksdb_iterator_t* rocksdb_create_iterator(
     rocksdb_t* db,
     const rocksdb_readoptions_t* options) {
@@ -2401,6 +2508,10 @@ void rocksdb_options_set_max_compaction_bytes(rocksdb_options_t* opt,
   opt->rep.max_compaction_bytes = n;
 }
 
+uint64_t rocksdb_options_get_max_compaction_bytes(rocksdb_options_t* opt) {
+  return opt->rep.max_compaction_bytes;
+}
+
 void rocksdb_options_set_max_bytes_for_level_multiplier_additional(
     rocksdb_options_t* opt, int* level_values, size_t num_levels) {
   opt->rep.max_bytes_for_level_multiplier_additional.resize(num_levels);
@@ -2476,14 +2587,26 @@ void rocksdb_options_set_wal_recovery_mode(rocksdb_options_t* opt,int mode) {
   opt->rep.wal_recovery_mode = static_cast<WALRecoveryMode>(mode);
 }
 
+int rocksdb_options_get_wal_recovery_mode(rocksdb_options_t* opt) {
+  return static_cast<int>(opt->rep.wal_recovery_mode);
+}
+
 void rocksdb_options_set_compression(rocksdb_options_t* opt, int t) {
   opt->rep.compression = static_cast<CompressionType>(t);
 }
 
+int rocksdb_options_get_compression(rocksdb_options_t* opt) {
+  return opt->rep.compression;
+}
+
 void rocksdb_options_set_bottommost_compression(rocksdb_options_t* opt, int t) {
   opt->rep.bottommost_compression = static_cast<CompressionType>(t);
 }
 
+int rocksdb_options_get_bottommost_compression(rocksdb_options_t* opt) {
+  return opt->rep.bottommost_compression;
+}
+
 void rocksdb_options_set_compression_per_level(rocksdb_options_t* opt,
                                                int* level_values,
                                                size_t num_levels) {
@@ -2537,6 +2660,10 @@ void rocksdb_options_set_use_fsync(
   opt->rep.use_fsync = use_fsync;
 }
 
+int rocksdb_options_get_use_fsync(rocksdb_options_t* opt) {
+  return opt->rep.use_fsync;
+}
+
 void rocksdb_options_set_db_log_dir(
     rocksdb_options_t* opt, const char* db_log_dir) {
   opt->rep.db_log_dir = db_log_dir;
@@ -2551,16 +2678,28 @@ void rocksdb_options_set_WAL_ttl_seconds(rocksdb_options_t* opt, uint64_t ttl) {
   opt->rep.WAL_ttl_seconds = ttl;
 }
 
+uint64_t rocksdb_options_get_WAL_ttl_seconds(rocksdb_options_t* opt) {
+  return opt->rep.WAL_ttl_seconds;
+}
+
 void rocksdb_options_set_WAL_size_limit_MB(
     rocksdb_options_t* opt, uint64_t limit) {
   opt->rep.WAL_size_limit_MB = limit;
 }
 
+uint64_t rocksdb_options_get_WAL_size_limit_MB(rocksdb_options_t* opt) {
+  return opt->rep.WAL_size_limit_MB;
+}
+
 void rocksdb_options_set_manifest_preallocation_size(
     rocksdb_options_t* opt, size_t v) {
   opt->rep.manifest_preallocation_size = v;
 }
 
+size_t rocksdb_options_get_manifest_preallocation_size(rocksdb_options_t* opt) {
+  return opt->rep.manifest_preallocation_size;
+}
+
 // noop
 void rocksdb_options_set_purge_redundant_kvs_while_flush(
     rocksdb_options_t* /*opt*/, unsigned char /*v*/) {}
@@ -2570,41 +2709,86 @@ void rocksdb_options_set_use_direct_reads(rocksdb_options_t* opt,
   opt->rep.use_direct_reads = v;
 }
 
+unsigned char rocksdb_options_get_use_direct_reads(rocksdb_options_t* opt) {
+  return opt->rep.use_direct_reads;
+}
+
 void rocksdb_options_set_use_direct_io_for_flush_and_compaction(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.use_direct_io_for_flush_and_compaction = v;
 }
 
+unsigned char rocksdb_options_get_use_direct_io_for_flush_and_compaction(
+    rocksdb_options_t* opt) {
+  return opt->rep.use_direct_io_for_flush_and_compaction;
+}
+
 void rocksdb_options_set_allow_mmap_reads(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.allow_mmap_reads = v;
 }
 
+unsigned char rocksdb_options_get_allow_mmap_reads(rocksdb_options_t* opt) {
+  return opt->rep.allow_mmap_reads;
+}
+
 void rocksdb_options_set_allow_mmap_writes(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.allow_mmap_writes = v;
 }
 
+unsigned char rocksdb_options_get_allow_mmap_writes(rocksdb_options_t* opt) {
+  return opt->rep.allow_mmap_writes;
+}
+
 void rocksdb_options_set_is_fd_close_on_exec(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.is_fd_close_on_exec = v;
 }
 
+unsigned char rocksdb_options_get_is_fd_close_on_exec(rocksdb_options_t* opt) {
+  return opt->rep.is_fd_close_on_exec;
+}
+
 void rocksdb_options_set_skip_log_error_on_recovery(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.skip_log_error_on_recovery = v;
 }
 
+unsigned char rocksdb_options_get_skip_log_error_on_recovery(
+    rocksdb_options_t* opt) {
+  return opt->rep.skip_log_error_on_recovery;
+}
+
 void rocksdb_options_set_stats_dump_period_sec(
     rocksdb_options_t* opt, unsigned int v) {
   opt->rep.stats_dump_period_sec = v;
 }
 
+unsigned int rocksdb_options_get_stats_dump_period_sec(rocksdb_options_t* opt) {
+  return opt->rep.stats_dump_period_sec;
+}
+
+void rocksdb_options_set_stats_persist_period_sec(rocksdb_options_t* opt,
+                                                  unsigned int v) {
+  opt->rep.stats_persist_period_sec = v;
+}
+
+unsigned int rocksdb_options_get_stats_persist_period_sec(
+    rocksdb_options_t* opt) {
+  return opt->rep.stats_persist_period_sec;
+}
+
 void rocksdb_options_set_advise_random_on_open(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.advise_random_on_open = v;
 }
 
+unsigned char rocksdb_options_get_advise_random_on_open(
+    rocksdb_options_t* opt) {
+  return opt->rep.advise_random_on_open;
+}
+
 void rocksdb_options_set_access_hint_on_compaction_start(
     rocksdb_options_t* opt, int v) {
   switch(v) {
@@ -2627,41 +2811,78 @@ void rocksdb_options_set_access_hint_on_compaction_start(
   }
 }
 
+int rocksdb_options_get_access_hint_on_compaction_start(
+    rocksdb_options_t* opt) {
+  return opt->rep.access_hint_on_compaction_start;
+}
+
 void rocksdb_options_set_use_adaptive_mutex(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.use_adaptive_mutex = v;
 }
 
+unsigned char rocksdb_options_get_use_adaptive_mutex(rocksdb_options_t* opt) {
+  return opt->rep.use_adaptive_mutex;
+}
+
 void rocksdb_options_set_wal_bytes_per_sync(
     rocksdb_options_t* opt, uint64_t v) {
   opt->rep.wal_bytes_per_sync = v;
 }
 
+uint64_t rocksdb_options_get_wal_bytes_per_sync(rocksdb_options_t* opt) {
+  return opt->rep.wal_bytes_per_sync;
+}
+
 void rocksdb_options_set_bytes_per_sync(
     rocksdb_options_t* opt, uint64_t v) {
   opt->rep.bytes_per_sync = v;
 }
 
+uint64_t rocksdb_options_get_bytes_per_sync(rocksdb_options_t* opt) {
+  return opt->rep.bytes_per_sync;
+}
+
 void rocksdb_options_set_writable_file_max_buffer_size(rocksdb_options_t* opt,
                                                        uint64_t v) {
   opt->rep.writable_file_max_buffer_size = static_cast<size_t>(v);
 }
 
+uint64_t rocksdb_options_get_writable_file_max_buffer_size(
+    rocksdb_options_t* opt) {
+  return opt->rep.writable_file_max_buffer_size;
+}
+
 void rocksdb_options_set_allow_concurrent_memtable_write(rocksdb_options_t* opt,
                                                          unsigned char v) {
   opt->rep.allow_concurrent_memtable_write = v;
 }
 
+unsigned char rocksdb_options_get_allow_concurrent_memtable_write(
+    rocksdb_options_t* opt) {
+  return opt->rep.allow_concurrent_memtable_write;
+}
+
 void rocksdb_options_set_enable_write_thread_adaptive_yield(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.enable_write_thread_adaptive_yield = v;
 }
 
+unsigned char rocksdb_options_get_enable_write_thread_adaptive_yield(
+    rocksdb_options_t* opt) {
+  return opt->rep.enable_write_thread_adaptive_yield;
+}
+
 void rocksdb_options_set_max_sequential_skip_in_iterations(
     rocksdb_options_t* opt, uint64_t v) {
   opt->rep.max_sequential_skip_in_iterations = v;
 }
 
+uint64_t rocksdb_options_get_max_sequential_skip_in_iterations(
+    rocksdb_options_t* opt) {
+  return opt->rep.max_sequential_skip_in_iterations;
+}
+
 void rocksdb_options_set_max_write_buffer_number(rocksdb_options_t* opt, int n) {
   opt->rep.max_write_buffer_number = n;
 }
@@ -2731,67 +2952,130 @@ void rocksdb_options_set_max_background_jobs(rocksdb_options_t* opt, int n) {
   opt->rep.max_background_jobs = n;
 }
 
+int rocksdb_options_get_max_background_jobs(rocksdb_options_t* opt) {
+  return opt->rep.max_background_jobs;
+}
+
 void rocksdb_options_set_max_background_compactions(rocksdb_options_t* opt, int n) {
   opt->rep.max_background_compactions = n;
 }
 
+int rocksdb_options_get_max_background_compactions(rocksdb_options_t* opt) {
+  return opt->rep.max_background_compactions;
+}
+
 void rocksdb_options_set_base_background_compactions(rocksdb_options_t* opt,
                                                      int n) {
   opt->rep.base_background_compactions = n;
 }
 
+int rocksdb_options_get_base_background_compactions(rocksdb_options_t* opt) {
+  return opt->rep.base_background_compactions;
+}
+
 void rocksdb_options_set_max_background_flushes(rocksdb_options_t* opt, int n) {
   opt->rep.max_background_flushes = n;
 }
 
+int rocksdb_options_get_max_background_flushes(rocksdb_options_t* opt) {
+  return opt->rep.max_background_flushes;
+}
+
 void rocksdb_options_set_max_log_file_size(rocksdb_options_t* opt, size_t v) {
   opt->rep.max_log_file_size = v;
 }
 
+size_t rocksdb_options_get_max_log_file_size(rocksdb_options_t* opt) {
+  return opt->rep.max_log_file_size;
+}
+
 void rocksdb_options_set_log_file_time_to_roll(rocksdb_options_t* opt, size_t v) {
   opt->rep.log_file_time_to_roll = v;
 }
 
+size_t rocksdb_options_get_log_file_time_to_roll(rocksdb_options_t* opt) {
+  return opt->rep.log_file_time_to_roll;
+}
+
 void rocksdb_options_set_keep_log_file_num(rocksdb_options_t* opt, size_t v) {
   opt->rep.keep_log_file_num = v;
 }
 
+size_t rocksdb_options_get_keep_log_file_num(rocksdb_options_t* opt) {
+  return opt->rep.keep_log_file_num;
+}
+
 void rocksdb_options_set_recycle_log_file_num(rocksdb_options_t* opt,
                                               size_t v) {
   opt->rep.recycle_log_file_num = v;
 }
 
+size_t rocksdb_options_get_recycle_log_file_num(rocksdb_options_t* opt) {
+  return opt->rep.recycle_log_file_num;
+}
+
 void rocksdb_options_set_soft_rate_limit(rocksdb_options_t* opt, double v) {
   opt->rep.soft_rate_limit = v;
 }
 
+double rocksdb_options_get_soft_rate_limit(rocksdb_options_t* opt) {
+  return opt->rep.soft_rate_limit;
+}
+
 void rocksdb_options_set_hard_rate_limit(rocksdb_options_t* opt, double v) {
   opt->rep.hard_rate_limit = v;
 }
 
+double rocksdb_options_get_hard_rate_limit(rocksdb_options_t* opt) {
+  return opt->rep.hard_rate_limit;
+}
+
 void rocksdb_options_set_soft_pending_compaction_bytes_limit(rocksdb_options_t* opt, size_t v) {
   opt->rep.soft_pending_compaction_bytes_limit = v;
 }
 
+size_t rocksdb_options_get_soft_pending_compaction_bytes_limit(
+    rocksdb_options_t* opt) {
+  return opt->rep.soft_pending_compaction_bytes_limit;
+}
+
 void rocksdb_options_set_hard_pending_compaction_bytes_limit(rocksdb_options_t* opt, size_t v) {
   opt->rep.hard_pending_compaction_bytes_limit = v;
 }
 
+size_t rocksdb_options_get_hard_pending_compaction_bytes_limit(
+    rocksdb_options_t* opt) {
+  return opt->rep.hard_pending_compaction_bytes_limit;
+}
+
 void rocksdb_options_set_rate_limit_delay_max_milliseconds(
     rocksdb_options_t* opt, unsigned int v) {
   opt->rep.rate_limit_delay_max_milliseconds = v;
 }
 
+unsigned int rocksdb_options_get_rate_limit_delay_max_milliseconds(
+    rocksdb_options_t* opt) {
+  return opt->rep.rate_limit_delay_max_milliseconds;
+}
+
 void rocksdb_options_set_max_manifest_file_size(
     rocksdb_options_t* opt, size_t v) {
   opt->rep.max_manifest_file_size = v;
 }
 
+size_t rocksdb_options_get_max_manifest_file_size(rocksdb_options_t* opt) {
+  return opt->rep.max_manifest_file_size;
+}
+
 void rocksdb_options_set_table_cache_numshardbits(
     rocksdb_options_t* opt, int v) {
   opt->rep.table_cache_numshardbits = v;
 }
 
+int rocksdb_options_get_table_cache_numshardbits(rocksdb_options_t* opt) {
+  return opt->rep.table_cache_numshardbits;
+}
+
 void rocksdb_options_set_table_cache_remove_scan_count_limit(
     rocksdb_options_t* /*opt*/, int /*v*/) {
   // this option is deprecated
@@ -2802,19 +3086,38 @@ void rocksdb_options_set_arena_block_size(
   opt->rep.arena_block_size = v;
 }
 
+size_t rocksdb_options_get_arena_block_size(rocksdb_options_t* opt) {
+  return opt->rep.arena_block_size;
+}
+
 void rocksdb_options_set_disable_auto_compactions(rocksdb_options_t* opt, int disable) {
   opt->rep.disable_auto_compactions = disable;
 }
 
+unsigned char rocksdb_options_get_disable_auto_compactions(
+    rocksdb_options_t* opt) {
+  return opt->rep.disable_auto_compactions;
+}
+
 void rocksdb_options_set_optimize_filters_for_hits(rocksdb_options_t* opt, int v) {
   opt->rep.optimize_filters_for_hits = v;
 }
 
+unsigned char rocksdb_options_get_optimize_filters_for_hits(
+    rocksdb_options_t* opt) {
+  return opt->rep.optimize_filters_for_hits;
+}
+
 void rocksdb_options_set_delete_obsolete_files_period_micros(
     rocksdb_options_t* opt, uint64_t v) {
   opt->rep.delete_obsolete_files_period_micros = v;
 }
 
+uint64_t rocksdb_options_get_delete_obsolete_files_period_micros(
+    rocksdb_options_t* opt) {
+  return opt->rep.delete_obsolete_files_period_micros;
+}
+
 void rocksdb_options_prepare_for_bulk_load(rocksdb_options_t* opt) {
   opt->rep.PrepareForBulkLoad();
 }
@@ -2828,11 +3131,20 @@ void rocksdb_options_set_memtable_prefix_bloom_size_ratio(
   opt->rep.memtable_prefix_bloom_size_ratio = v;
 }
 
+double rocksdb_options_get_memtable_prefix_bloom_size_ratio(
+    rocksdb_options_t* opt) {
+  return opt->rep.memtable_prefix_bloom_size_ratio;
+}
+
 void rocksdb_options_set_memtable_huge_page_size(rocksdb_options_t* opt,
                                                  size_t v) {
   opt->rep.memtable_huge_page_size = v;
 }
 
+size_t rocksdb_options_get_memtable_huge_page_size(rocksdb_options_t* opt) {
+  return opt->rep.memtable_huge_page_size;
+}
+
 void rocksdb_options_set_hash_skip_list_rep(
     rocksdb_options_t *opt, size_t bucket_count,
     int32_t skiplist_height, int32_t skiplist_branching_factor) {
@@ -2867,31 +3179,56 @@ void rocksdb_options_set_max_successive_merges(
   opt->rep.max_successive_merges = v;
 }
 
+size_t rocksdb_options_get_max_successive_merges(rocksdb_options_t* opt) {
+  return opt->rep.max_successive_merges;
+}
+
 void rocksdb_options_set_bloom_locality(
     rocksdb_options_t* opt, uint32_t v) {
   opt->rep.bloom_locality = v;
 }
 
+uint32_t rocksdb_options_get_bloom_locality(rocksdb_options_t* opt) {
+  return opt->rep.bloom_locality;
+}
+
 void rocksdb_options_set_inplace_update_support(
     rocksdb_options_t* opt, unsigned char v) {
   opt->rep.inplace_update_support = v;
 }
 
+unsigned char rocksdb_options_get_inplace_update_support(
+    rocksdb_options_t* opt) {
+  return opt->rep.inplace_update_support;
+}
+
 void rocksdb_options_set_inplace_update_num_locks(
     rocksdb_options_t* opt, size_t v) {
   opt->rep.inplace_update_num_locks = v;
 }
 
+size_t rocksdb_options_get_inplace_update_num_locks(rocksdb_options_t* opt) {
+  return opt->rep.inplace_update_num_locks;
+}
+
 void rocksdb_options_set_report_bg_io_stats(
     rocksdb_options_t* opt, int v) {
   opt->rep.report_bg_io_stats = v;
 }
 
+unsigned char rocksdb_options_get_report_bg_io_stats(rocksdb_options_t* opt) {
+  return opt->rep.report_bg_io_stats;
+}
+
 void rocksdb_options_set_compaction_style(rocksdb_options_t *opt, int style) {
   opt->rep.compaction_style =
       static_cast<ROCKSDB_NAMESPACE::CompactionStyle>(style);
 }
 
+int rocksdb_options_get_compaction_style(rocksdb_options_t* opt) {
+  return opt->rep.compaction_style;
+}
+
 void rocksdb_options_set_universal_compaction_options(rocksdb_options_t *opt, rocksdb_universal_compaction_options_t *uco) {
   opt->rep.compaction_options_universal = *(uco->rep);
 }
@@ -2921,6 +3258,10 @@ void rocksdb_options_set_atomic_flush(rocksdb_options_t* opt,
   opt->rep.atomic_flush = atomic_flush;
 }
 
+unsigned char rocksdb_options_get_atomic_flush(rocksdb_options_t* opt) {
+  return opt->rep.atomic_flush;
+}
+
 rocksdb_ratelimiter_t* rocksdb_ratelimiter_create(
     int64_t rate_bytes_per_sec,
     int64_t refill_period_us,
@@ -3320,11 +3661,20 @@ void rocksdb_readoptions_set_verify_checksums(
   opt->rep.verify_checksums = v;
 }
 
+unsigned char rocksdb_readoptions_get_verify_checksums(
+    rocksdb_readoptions_t* opt) {
+  return opt->rep.verify_checksums;
+}
+
 void rocksdb_readoptions_set_fill_cache(
     rocksdb_readoptions_t* opt, unsigned char v) {
   opt->rep.fill_cache = v;
 }
 
+unsigned char rocksdb_readoptions_get_fill_cache(rocksdb_readoptions_t* opt) {
+  return opt->rep.fill_cache;
+}
+
 void rocksdb_readoptions_set_snapshot(
     rocksdb_readoptions_t* opt,
     const rocksdb_snapshot_t* snap) {
@@ -3361,11 +3711,19 @@ void rocksdb_readoptions_set_read_tier(
   opt->rep.read_tier = static_cast<ROCKSDB_NAMESPACE::ReadTier>(v);
 }
 
+int rocksdb_readoptions_get_read_tier(rocksdb_readoptions_t* opt) {
+  return static_cast<int>(opt->rep.read_tier);
+}
+
 void rocksdb_readoptions_set_tailing(
     rocksdb_readoptions_t* opt, unsigned char v) {
   opt->rep.tailing = v;
 }
 
+unsigned char rocksdb_readoptions_get_tailing(rocksdb_readoptions_t* opt) {
+  return opt->rep.tailing;
+}
+
 void rocksdb_readoptions_set_managed(
     rocksdb_readoptions_t* opt, unsigned char v) {
   opt->rep.managed = v;
@@ -3376,37 +3734,70 @@ void rocksdb_readoptions_set_readahead_size(
   opt->rep.readahead_size = v;
 }
 
+size_t rocksdb_readoptions_get_readahead_size(rocksdb_readoptions_t* opt) {
+  return opt->rep.readahead_size;
+}
+
 void rocksdb_readoptions_set_prefix_same_as_start(
     rocksdb_readoptions_t* opt, unsigned char v) {
   opt->rep.prefix_same_as_start = v;
 }
 
+unsigned char rocksdb_readoptions_get_prefix_same_as_start(
+    rocksdb_readoptions_t* opt) {
+  return opt->rep.prefix_same_as_start;
+}
+
 void rocksdb_readoptions_set_pin_data(rocksdb_readoptions_t* opt,
                                       unsigned char v) {
   opt->rep.pin_data = v;
 }
 
+unsigned char rocksdb_readoptions_get_pin_data(rocksdb_readoptions_t* opt) {
+  return opt->rep.pin_data;
+}
+
 void rocksdb_readoptions_set_total_order_seek(rocksdb_readoptions_t* opt,
                                               unsigned char v) {
   opt->rep.total_order_seek = v;
 }
 
+unsigned char rocksdb_readoptions_get_total_order_seek(
+    rocksdb_readoptions_t* opt) {
+  return opt->rep.total_order_seek;
+}
+
 void rocksdb_readoptions_set_max_skippable_internal_keys(
     rocksdb_readoptions_t* opt,
     uint64_t v) {
   opt->rep.max_skippable_internal_keys = v;
 }
 
+uint64_t rocksdb_readoptions_get_max_skippable_internal_keys(
+    rocksdb_readoptions_t* opt) {
+  return opt->rep.max_skippable_internal_keys;
+}
+
 void rocksdb_readoptions_set_background_purge_on_iterator_cleanup(
     rocksdb_readoptions_t* opt, unsigned char v) {
   opt->rep.background_purge_on_iterator_cleanup = v;
 }
 
+unsigned char rocksdb_readoptions_get_background_purge_on_iterator_cleanup(
+    rocksdb_readoptions_t* opt) {
+  return opt->rep.background_purge_on_iterator_cleanup;
+}
+
 void rocksdb_readoptions_set_ignore_range_deletions(
     rocksdb_readoptions_t* opt, unsigned char v) {
   opt->rep.ignore_range_deletions = v;
 }
 
+unsigned char rocksdb_readoptions_get_ignore_range_deletions(
+    rocksdb_readoptions_t* opt) {
+  return opt->rep.ignore_range_deletions;
+}
+
 rocksdb_writeoptions_t* rocksdb_writeoptions_create() {
   return new rocksdb_writeoptions_t;
 }
@@ -3420,33 +3811,61 @@ void rocksdb_writeoptions_set_sync(
   opt->rep.sync = v;
 }
 
+unsigned char rocksdb_writeoptions_get_sync(rocksdb_writeoptions_t* opt) {
+  return opt->rep.sync;
+}
+
 void rocksdb_writeoptions_disable_WAL(rocksdb_writeoptions_t* opt, int disable) {
   opt->rep.disableWAL = disable;
 }
 
+unsigned char rocksdb_writeoptions_get_disable_WAL(
+    rocksdb_writeoptions_t* opt) {
+  return opt->rep.disableWAL;
+}
+
 void rocksdb_writeoptions_set_ignore_missing_column_families(
     rocksdb_writeoptions_t* opt,
     unsigned char v) {
   opt->rep.ignore_missing_column_families = v;
 }
 
+unsigned char rocksdb_writeoptions_get_ignore_missing_column_families(
+    rocksdb_writeoptions_t* opt) {
+  return opt->rep.ignore_missing_column_families;
+}
+
 void rocksdb_writeoptions_set_no_slowdown(
     rocksdb_writeoptions_t* opt,
     unsigned char v) {
   opt->rep.no_slowdown = v;
 }
 
+unsigned char rocksdb_writeoptions_get_no_slowdown(
+    rocksdb_writeoptions_t* opt) {
+  return opt->rep.no_slowdown;
+}
+
 void rocksdb_writeoptions_set_low_pri(
     rocksdb_writeoptions_t* opt,
     unsigned char v) {
   opt->rep.low_pri = v;
 }
 
+unsigned char rocksdb_writeoptions_get_low_pri(rocksdb_writeoptions_t* opt) {
+  return opt->rep.low_pri;
+}
+
 void rocksdb_writeoptions_set_memtable_insert_hint_per_batch(
     rocksdb_writeoptions_t* opt, unsigned char v) {
   opt->rep.memtable_insert_hint_per_batch = v;
 }
 
+unsigned char rocksdb_writeoptions_get_memtable_insert_hint_per_batch(
+    rocksdb_writeoptions_t* opt) {
+  return opt->rep.memtable_insert_hint_per_batch;
+}
+
 rocksdb_compactoptions_t* rocksdb_compactoptions_create() {
   return new rocksdb_compactoptions_t;
 }
@@ -3460,21 +3879,40 @@ void rocksdb_compactoptions_set_bottommost_level_compaction(
   opt->rep.bottommost_level_compaction = static_cast<BottommostLevelCompaction>(v);
 }
 
+unsigned char rocksdb_compactoptions_get_bottommost_level_compaction(
+    rocksdb_compactoptions_t* opt) {
+  return static_cast<unsigned char>(opt->rep.bottommost_level_compaction);
+}
+
 void rocksdb_compactoptions_set_exclusive_manual_compaction(
     rocksdb_compactoptions_t* opt, unsigned char v) {
   opt->rep.exclusive_manual_compaction = v;
 }
 
+unsigned char rocksdb_compactoptions_get_exclusive_manual_compaction(
+    rocksdb_compactoptions_t* opt) {
+  return opt->rep.exclusive_manual_compaction;
+}
+
 void rocksdb_compactoptions_set_change_level(rocksdb_compactoptions_t* opt,
                                              unsigned char v) {
   opt->rep.change_level = v;
 }
 
+unsigned char rocksdb_compactoptions_get_change_level(
+    rocksdb_compactoptions_t* opt) {
+  return opt->rep.change_level;
+}
+
 void rocksdb_compactoptions_set_target_level(rocksdb_compactoptions_t* opt,
                                              int n) {
   opt->rep.target_level = n;
 }
 
+int rocksdb_compactoptions_get_target_level(rocksdb_compactoptions_t* opt) {
+  return opt->rep.target_level;
+}
+
 rocksdb_flushoptions_t* rocksdb_flushoptions_create() {
   return new rocksdb_flushoptions_t;
 }
@@ -3488,6 +3926,10 @@ void rocksdb_flushoptions_set_wait(
   opt->rep.wait = v;
 }
 
+unsigned char rocksdb_flushoptions_get_wait(rocksdb_flushoptions_t* opt) {
+  return opt->rep.wait;
+}
+
 rocksdb_cache_t* rocksdb_cache_create_lru(size_t capacity) {
   rocksdb_cache_t* c = new rocksdb_cache_t;
   c->rep = NewLRUCache(capacity);
@@ -3502,6 +3944,10 @@ void rocksdb_cache_set_capacity(rocksdb_cache_t* cache, size_t capacity) {
   cache->rep->SetCapacity(capacity);
 }
 
+size_t rocksdb_cache_get_capacity(rocksdb_cache_t* cache) {
+  return cache->rep->GetCapacity();
+}
+
 size_t rocksdb_cache_get_usage(rocksdb_cache_t* cache) {
   return cache->rep->GetUsage();
 }
@@ -3539,10 +3985,20 @@ void rocksdb_env_set_background_threads(rocksdb_env_t* env, int n) {
   env->rep->SetBackgroundThreads(n);
 }
 
+void rocksdb_env_set_bottom_priority_background_threads(rocksdb_env_t* env,
+                                                        int n) {
+  env->rep->SetBackgroundThreads(n, Env::BOTTOM);
+}
+
 void rocksdb_env_set_high_priority_background_threads(rocksdb_env_t* env, int n) {
   env->rep->SetBackgroundThreads(n, Env::HIGH);
 }
 
+void rocksdb_env_set_low_priority_background_threads(rocksdb_env_t* env,
+                                                     int n) {
+  env->rep->SetBackgroundThreads(n, Env::LOW);
+}
+
 void rocksdb_env_join_all_threads(rocksdb_env_t* env) {
   env->rep->WaitForJoin();
 }

db/c_test.c

@@ -58,7 +58,11 @@ static void StartPhase(const char* name) {
 static const char* GetTempDir(void) {
     const char* ret = getenv("TEST_TMPDIR");
     if (ret == NULL || ret[0] == '\0')
-        ret = "/tmp";
+#ifdef OS_WIN
+      ret = getenv("TEMP");
+#else
+      ret = "/tmp";
+#endif
     return ret;
 }
 #ifdef _MSC_VER
@@ -1296,6 +1300,29 @@ int main(int argc, char** argv) {
       Free(&vals[i]);
     }
 
+    {
+      unsigned char value_found = 0;
+
+      CheckCondition(!rocksdb_key_may_exist(db, roptions, "invalid_key", 11,
+                                            NULL, NULL, NULL, 0, NULL));
+      CheckCondition(!rocksdb_key_may_exist(db, roptions, "invalid_key", 11,
+                                            &vals[0], &vals_sizes[0], NULL, 0,
+                                            &value_found));
+      if (value_found) {
+        Free(&vals[0]);
+      }
+
+      CheckCondition(!rocksdb_key_may_exist_cf(db, roptions, handles[1],
+                                               "invalid_key", 11, NULL, NULL,
+                                               NULL, 0, NULL));
+      CheckCondition(!rocksdb_key_may_exist_cf(db, roptions, handles[1],
+                                               "invalid_key", 11, &vals[0],
+                                               &vals_sizes[0], NULL, 0, NULL));
+      if (value_found) {
+        Free(&vals[0]);
+      }
+    }
+
     rocksdb_iterator_t* iter = rocksdb_create_iterator_cf(db, roptions, handles[1]);
     CheckCondition(!rocksdb_iter_valid(iter));
     rocksdb_iter_seek_to_first(iter);
@@ -1564,6 +1591,173 @@ int main(int argc, char** argv) {
     rocksdb_options_set_max_subcompactions(o, 123456);
     CheckCondition(123456 == rocksdb_options_get_max_subcompactions(o));
 
+    rocksdb_options_set_max_background_jobs(o, 2);
+    CheckCondition(2 == rocksdb_options_get_max_background_jobs(o));
+
+    rocksdb_options_set_max_background_compactions(o, 3);
+    CheckCondition(3 == rocksdb_options_get_max_background_compactions(o));
+
+    rocksdb_options_set_base_background_compactions(o, 4);
+    CheckCondition(4 == rocksdb_options_get_base_background_compactions(o));
+
+    rocksdb_options_set_max_background_flushes(o, 5);
+    CheckCondition(5 == rocksdb_options_get_max_background_flushes(o));
+
+    rocksdb_options_set_max_log_file_size(o, 6);
+    CheckCondition(6 == rocksdb_options_get_max_log_file_size(o));
+
+    rocksdb_options_set_log_file_time_to_roll(o, 7);
+    CheckCondition(7 == rocksdb_options_get_log_file_time_to_roll(o));
+
+    rocksdb_options_set_keep_log_file_num(o, 8);
+    CheckCondition(8 == rocksdb_options_get_keep_log_file_num(o));
+
+    rocksdb_options_set_recycle_log_file_num(o, 9);
+    CheckCondition(9 == rocksdb_options_get_recycle_log_file_num(o));
+
+    rocksdb_options_set_soft_rate_limit(o, 2.0);
+    CheckCondition(2.0 == rocksdb_options_get_soft_rate_limit(o));
+
+    rocksdb_options_set_hard_rate_limit(o, 4.0);
+    CheckCondition(4.0 == rocksdb_options_get_hard_rate_limit(o));
+
+    rocksdb_options_set_soft_pending_compaction_bytes_limit(o, 10);
+    CheckCondition(10 ==
+                   rocksdb_options_get_soft_pending_compaction_bytes_limit(o));
+
+    rocksdb_options_set_hard_pending_compaction_bytes_limit(o, 11);
+    CheckCondition(11 ==
+                   rocksdb_options_get_hard_pending_compaction_bytes_limit(o));
+
+    rocksdb_options_set_rate_limit_delay_max_milliseconds(o, 1);
+    CheckCondition(1 ==
+                   rocksdb_options_get_rate_limit_delay_max_milliseconds(o));
+
+    rocksdb_options_set_max_manifest_file_size(o, 12);
+    CheckCondition(12 == rocksdb_options_get_max_manifest_file_size(o));
+
+    rocksdb_options_set_table_cache_numshardbits(o, 13);
+    CheckCondition(13 == rocksdb_options_get_table_cache_numshardbits(o));
+
+    rocksdb_options_set_arena_block_size(o, 14);
+    CheckCondition(14 == rocksdb_options_get_arena_block_size(o));
+
+    rocksdb_options_set_use_fsync(o, 1);
+    CheckCondition(1 == rocksdb_options_get_use_fsync(o));
+
+    rocksdb_options_set_WAL_ttl_seconds(o, 15);
+    CheckCondition(15 == rocksdb_options_get_WAL_ttl_seconds(o));
+
+    rocksdb_options_set_WAL_size_limit_MB(o, 16);
+    CheckCondition(16 == rocksdb_options_get_WAL_size_limit_MB(o));
+
+    rocksdb_options_set_manifest_preallocation_size(o, 17);
+    CheckCondition(17 == rocksdb_options_get_manifest_preallocation_size(o));
+
+    rocksdb_options_set_allow_mmap_reads(o, 1);
+    CheckCondition(1 == rocksdb_options_get_allow_mmap_reads(o));
+
+    rocksdb_options_set_allow_mmap_writes(o, 1);
+    CheckCondition(1 == rocksdb_options_get_allow_mmap_writes(o));
+
+    rocksdb_options_set_use_direct_reads(o, 1);
+    CheckCondition(1 == rocksdb_options_get_use_direct_reads(o));
+
+    rocksdb_options_set_use_direct_io_for_flush_and_compaction(o, 1);
+    CheckCondition(
+        1 == rocksdb_options_get_use_direct_io_for_flush_and_compaction(o));
+
+    rocksdb_options_set_is_fd_close_on_exec(o, 1);
+    CheckCondition(1 == rocksdb_options_get_is_fd_close_on_exec(o));
+
+    rocksdb_options_set_skip_log_error_on_recovery(o, 1);
+    CheckCondition(1 == rocksdb_options_get_skip_log_error_on_recovery(o));
+
+    rocksdb_options_set_stats_dump_period_sec(o, 18);
+    CheckCondition(18 == rocksdb_options_get_stats_dump_period_sec(o));
+
+    rocksdb_options_set_stats_persist_period_sec(o, 5);
+    CheckCondition(5 == rocksdb_options_get_stats_persist_period_sec(o));
+
+    rocksdb_options_set_advise_random_on_open(o, 1);
+    CheckCondition(1 == rocksdb_options_get_advise_random_on_open(o));
+
+    rocksdb_options_set_access_hint_on_compaction_start(o, 3);
+    CheckCondition(3 == rocksdb_options_get_access_hint_on_compaction_start(o));
+
+    rocksdb_options_set_use_adaptive_mutex(o, 1);
+    CheckCondition(1 == rocksdb_options_get_use_adaptive_mutex(o));
+
+    rocksdb_options_set_bytes_per_sync(o, 19);
+    CheckCondition(19 == rocksdb_options_get_bytes_per_sync(o));
+
+    rocksdb_options_set_wal_bytes_per_sync(o, 20);
+    CheckCondition(20 == rocksdb_options_get_wal_bytes_per_sync(o));
+
+    rocksdb_options_set_writable_file_max_buffer_size(o, 21);
+    CheckCondition(21 == rocksdb_options_get_writable_file_max_buffer_size(o));
+
+    rocksdb_options_set_allow_concurrent_memtable_write(o, 1);
+    CheckCondition(1 == rocksdb_options_get_allow_concurrent_memtable_write(o));
+
+    rocksdb_options_set_enable_write_thread_adaptive_yield(o, 1);
+    CheckCondition(1 ==
+                   rocksdb_options_get_enable_write_thread_adaptive_yield(o));
+
+    rocksdb_options_set_max_sequential_skip_in_iterations(o, 22);
+    CheckCondition(22 ==
+                   rocksdb_options_get_max_sequential_skip_in_iterations(o));
+
+    rocksdb_options_set_disable_auto_compactions(o, 1);
+    CheckCondition(1 == rocksdb_options_get_disable_auto_compactions(o));
+
+    rocksdb_options_set_optimize_filters_for_hits(o, 1);
+    CheckCondition(1 == rocksdb_options_get_optimize_filters_for_hits(o));
+
+    rocksdb_options_set_delete_obsolete_files_period_micros(o, 23);
+    CheckCondition(23 ==
+                   rocksdb_options_get_delete_obsolete_files_period_micros(o));
+
+    rocksdb_options_set_memtable_prefix_bloom_size_ratio(o, 2.0);
+    CheckCondition(2.0 ==
+                   rocksdb_options_get_memtable_prefix_bloom_size_ratio(o));
+
+    rocksdb_options_set_max_compaction_bytes(o, 24);
+    CheckCondition(24 == rocksdb_options_get_max_compaction_bytes(o));
+
+    rocksdb_options_set_memtable_huge_page_size(o, 25);
+    CheckCondition(25 == rocksdb_options_get_memtable_huge_page_size(o));
+
+    rocksdb_options_set_max_successive_merges(o, 26);
+    CheckCondition(26 == rocksdb_options_get_max_successive_merges(o));
+
+    rocksdb_options_set_bloom_locality(o, 27);
+    CheckCondition(27 == rocksdb_options_get_bloom_locality(o));
+
+    rocksdb_options_set_inplace_update_support(o, 1);
+    CheckCondition(1 == rocksdb_options_get_inplace_update_support(o));
+
+    rocksdb_options_set_inplace_update_num_locks(o, 28);
+    CheckCondition(28 == rocksdb_options_get_inplace_update_num_locks(o));
+
+    rocksdb_options_set_report_bg_io_stats(o, 1);
+    CheckCondition(1 == rocksdb_options_get_report_bg_io_stats(o));
+
+    rocksdb_options_set_wal_recovery_mode(o, 2);
+    CheckCondition(2 == rocksdb_options_get_wal_recovery_mode(o));
+
+    rocksdb_options_set_compression(o, 5);
+    CheckCondition(5 == rocksdb_options_get_compression(o));
+
+    rocksdb_options_set_bottommost_compression(o, 4);
+    CheckCondition(4 == rocksdb_options_get_bottommost_compression(o));
+
+    rocksdb_options_set_compaction_style(o, 2);
+    CheckCondition(2 == rocksdb_options_get_compaction_style(o));
+
+    rocksdb_options_set_atomic_flush(o, 1);
+    CheckCondition(1 == rocksdb_options_get_atomic_flush(o));
+
     // Create a copy that should be equal to the original.
     rocksdb_options_t* copy;
     copy = rocksdb_options_create_copy(o);
@@ -1607,6 +1801,70 @@ int main(int argc, char** argv) {
     CheckCondition(1 == rocksdb_options_get_enable_pipelined_write(copy));
     CheckCondition(1 == rocksdb_options_get_unordered_write(copy));
     CheckCondition(123456 == rocksdb_options_get_max_subcompactions(copy));
+    CheckCondition(2 == rocksdb_options_get_max_background_jobs(copy));
+    CheckCondition(3 == rocksdb_options_get_max_background_compactions(copy));
+    CheckCondition(4 == rocksdb_options_get_base_background_compactions(copy));
+    CheckCondition(5 == rocksdb_options_get_max_background_flushes(copy));
+    CheckCondition(6 == rocksdb_options_get_max_log_file_size(copy));
+    CheckCondition(7 == rocksdb_options_get_log_file_time_to_roll(copy));
+    CheckCondition(8 == rocksdb_options_get_keep_log_file_num(copy));
+    CheckCondition(9 == rocksdb_options_get_recycle_log_file_num(copy));
+    CheckCondition(2.0 == rocksdb_options_get_soft_rate_limit(copy));
+    CheckCondition(4.0 == rocksdb_options_get_hard_rate_limit(copy));
+    CheckCondition(
+        10 == rocksdb_options_get_soft_pending_compaction_bytes_limit(copy));
+    CheckCondition(
+        11 == rocksdb_options_get_hard_pending_compaction_bytes_limit(copy));
+    CheckCondition(1 ==
+                   rocksdb_options_get_rate_limit_delay_max_milliseconds(copy));
+    CheckCondition(12 == rocksdb_options_get_max_manifest_file_size(copy));
+    CheckCondition(13 == rocksdb_options_get_table_cache_numshardbits(copy));
+    CheckCondition(14 == rocksdb_options_get_arena_block_size(copy));
+    CheckCondition(1 == rocksdb_options_get_use_fsync(copy));
+    CheckCondition(15 == rocksdb_options_get_WAL_ttl_seconds(copy));
+    CheckCondition(16 == rocksdb_options_get_WAL_size_limit_MB(copy));
+    CheckCondition(17 == rocksdb_options_get_manifest_preallocation_size(copy));
+    CheckCondition(1 == rocksdb_options_get_allow_mmap_reads(copy));
+    CheckCondition(1 == rocksdb_options_get_allow_mmap_writes(copy));
+    CheckCondition(1 == rocksdb_options_get_use_direct_reads(copy));
+    CheckCondition(
+        1 == rocksdb_options_get_use_direct_io_for_flush_and_compaction(copy));
+    CheckCondition(1 == rocksdb_options_get_is_fd_close_on_exec(copy));
+    CheckCondition(1 == rocksdb_options_get_skip_log_error_on_recovery(copy));
+    CheckCondition(18 == rocksdb_options_get_stats_dump_period_sec(copy));
+    CheckCondition(5 == rocksdb_options_get_stats_persist_period_sec(copy));
+    CheckCondition(1 == rocksdb_options_get_advise_random_on_open(copy));
+    CheckCondition(3 ==
+                   rocksdb_options_get_access_hint_on_compaction_start(copy));
+    CheckCondition(1 == rocksdb_options_get_use_adaptive_mutex(copy));
+    CheckCondition(19 == rocksdb_options_get_bytes_per_sync(copy));
+    CheckCondition(20 == rocksdb_options_get_wal_bytes_per_sync(copy));
+    CheckCondition(21 ==
+                   rocksdb_options_get_writable_file_max_buffer_size(copy));
+    CheckCondition(1 ==
+                   rocksdb_options_get_allow_concurrent_memtable_write(copy));
+    CheckCondition(
+        1 == rocksdb_options_get_enable_write_thread_adaptive_yield(copy));
+    CheckCondition(22 ==
+                   rocksdb_options_get_max_sequential_skip_in_iterations(copy));
+    CheckCondition(1 == rocksdb_options_get_disable_auto_compactions(copy));
+    CheckCondition(1 == rocksdb_options_get_optimize_filters_for_hits(copy));
+    CheckCondition(
+        23 == rocksdb_options_get_delete_obsolete_files_period_micros(copy));
+    CheckCondition(2.0 ==
+                   rocksdb_options_get_memtable_prefix_bloom_size_ratio(copy));
+    CheckCondition(24 == rocksdb_options_get_max_compaction_bytes(copy));
+    CheckCondition(25 == rocksdb_options_get_memtable_huge_page_size(copy));
+    CheckCondition(26 == rocksdb_options_get_max_successive_merges(copy));
+    CheckCondition(27 == rocksdb_options_get_bloom_locality(copy));
+    CheckCondition(1 == rocksdb_options_get_inplace_update_support(copy));
+    CheckCondition(28 == rocksdb_options_get_inplace_update_num_locks(copy));
+    CheckCondition(1 == rocksdb_options_get_report_bg_io_stats(copy));
+    CheckCondition(2 == rocksdb_options_get_wal_recovery_mode(copy));
+    CheckCondition(5 == rocksdb_options_get_compression(copy));
+    CheckCondition(4 == rocksdb_options_get_bottommost_compression(copy));
+    CheckCondition(2 == rocksdb_options_get_compaction_style(copy));
+    CheckCondition(1 == rocksdb_options_get_atomic_flush(copy));
 
     // Copies should be independent.
     rocksdb_options_set_allow_ingest_behind(copy, 0);
@@ -1745,10 +2003,358 @@ int main(int argc, char** argv) {
     CheckCondition(90001 == rocksdb_options_get_max_subcompactions(copy));
     CheckCondition(123456 == rocksdb_options_get_max_subcompactions(o));
 
+    rocksdb_options_set_max_background_jobs(copy, 12);
+    CheckCondition(12 == rocksdb_options_get_max_background_jobs(copy));
+    CheckCondition(2 == rocksdb_options_get_max_background_jobs(o));
+
+    rocksdb_options_set_max_background_compactions(copy, 13);
+    CheckCondition(13 == rocksdb_options_get_max_background_compactions(copy));
+    CheckCondition(3 == rocksdb_options_get_max_background_compactions(o));
+
+    rocksdb_options_set_base_background_compactions(copy, 14);
+    CheckCondition(14 == rocksdb_options_get_base_background_compactions(copy));
+    CheckCondition(4 == rocksdb_options_get_base_background_compactions(o));
+
+    rocksdb_options_set_max_background_flushes(copy, 15);
+    CheckCondition(15 == rocksdb_options_get_max_background_flushes(copy));
+    CheckCondition(5 == rocksdb_options_get_max_background_flushes(o));
+
+    rocksdb_options_set_max_log_file_size(copy, 16);
+    CheckCondition(16 == rocksdb_options_get_max_log_file_size(copy));
+    CheckCondition(6 == rocksdb_options_get_max_log_file_size(o));
+
+    rocksdb_options_set_log_file_time_to_roll(copy, 17);
+    CheckCondition(17 == rocksdb_options_get_log_file_time_to_roll(copy));
+    CheckCondition(7 == rocksdb_options_get_log_file_time_to_roll(o));
+
+    rocksdb_options_set_keep_log_file_num(copy, 18);
+    CheckCondition(18 == rocksdb_options_get_keep_log_file_num(copy));
+    CheckCondition(8 == rocksdb_options_get_keep_log_file_num(o));
+
+    rocksdb_options_set_recycle_log_file_num(copy, 19);
+    CheckCondition(19 == rocksdb_options_get_recycle_log_file_num(copy));
+    CheckCondition(9 == rocksdb_options_get_recycle_log_file_num(o));
+
+    rocksdb_options_set_soft_rate_limit(copy, 4.0);
+    CheckCondition(4.0 == rocksdb_options_get_soft_rate_limit(copy));
+    CheckCondition(2.0 == rocksdb_options_get_soft_rate_limit(o));
+
+    rocksdb_options_set_hard_rate_limit(copy, 2.0);
+    CheckCondition(2.0 == rocksdb_options_get_hard_rate_limit(copy));
+    CheckCondition(4.0 == rocksdb_options_get_hard_rate_limit(o));
+
+    rocksdb_options_set_soft_pending_compaction_bytes_limit(copy, 110);
+    CheckCondition(
+        110 == rocksdb_options_get_soft_pending_compaction_bytes_limit(copy));
+    CheckCondition(10 ==
+                   rocksdb_options_get_soft_pending_compaction_bytes_limit(o));
+
+    rocksdb_options_set_hard_pending_compaction_bytes_limit(copy, 111);
+    CheckCondition(
+        111 == rocksdb_options_get_hard_pending_compaction_bytes_limit(copy));
+    CheckCondition(11 ==
+                   rocksdb_options_get_hard_pending_compaction_bytes_limit(o));
+
+    rocksdb_options_set_rate_limit_delay_max_milliseconds(copy, 0);
+    CheckCondition(0 ==
+                   rocksdb_options_get_rate_limit_delay_max_milliseconds(copy));
+    CheckCondition(1 ==
+                   rocksdb_options_get_rate_limit_delay_max_milliseconds(o));
+
+    rocksdb_options_set_max_manifest_file_size(copy, 112);
+    CheckCondition(112 == rocksdb_options_get_max_manifest_file_size(copy));
+    CheckCondition(12 == rocksdb_options_get_max_manifest_file_size(o));
+
+    rocksdb_options_set_table_cache_numshardbits(copy, 113);
+    CheckCondition(113 == rocksdb_options_get_table_cache_numshardbits(copy));
+    CheckCondition(13 == rocksdb_options_get_table_cache_numshardbits(o));
+
+    rocksdb_options_set_arena_block_size(copy, 114);
+    CheckCondition(114 == rocksdb_options_get_arena_block_size(copy));
+    CheckCondition(14 == rocksdb_options_get_arena_block_size(o));
+
+    rocksdb_options_set_use_fsync(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_use_fsync(copy));
+    CheckCondition(1 == rocksdb_options_get_use_fsync(o));
+
+    rocksdb_options_set_WAL_ttl_seconds(copy, 115);
+    CheckCondition(115 == rocksdb_options_get_WAL_ttl_seconds(copy));
+    CheckCondition(15 == rocksdb_options_get_WAL_ttl_seconds(o));
+
+    rocksdb_options_set_WAL_size_limit_MB(copy, 116);
+    CheckCondition(116 == rocksdb_options_get_WAL_size_limit_MB(copy));
+    CheckCondition(16 == rocksdb_options_get_WAL_size_limit_MB(o));
+
+    rocksdb_options_set_manifest_preallocation_size(copy, 117);
+    CheckCondition(117 ==
+                   rocksdb_options_get_manifest_preallocation_size(copy));
+    CheckCondition(17 == rocksdb_options_get_manifest_preallocation_size(o));
+
+    rocksdb_options_set_allow_mmap_reads(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_allow_mmap_reads(copy));
+    CheckCondition(1 == rocksdb_options_get_allow_mmap_reads(o));
+
+    rocksdb_options_set_allow_mmap_writes(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_allow_mmap_writes(copy));
+    CheckCondition(1 == rocksdb_options_get_allow_mmap_writes(o));
+
+    rocksdb_options_set_use_direct_reads(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_use_direct_reads(copy));
+    CheckCondition(1 == rocksdb_options_get_use_direct_reads(o));
+
+    rocksdb_options_set_use_direct_io_for_flush_and_compaction(copy, 0);
+    CheckCondition(
+        0 == rocksdb_options_get_use_direct_io_for_flush_and_compaction(copy));
+    CheckCondition(
+        1 == rocksdb_options_get_use_direct_io_for_flush_and_compaction(o));
+
+    rocksdb_options_set_is_fd_close_on_exec(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_is_fd_close_on_exec(copy));
+    CheckCondition(1 == rocksdb_options_get_is_fd_close_on_exec(o));
+
+    rocksdb_options_set_skip_log_error_on_recovery(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_skip_log_error_on_recovery(copy));
+    CheckCondition(1 == rocksdb_options_get_skip_log_error_on_recovery(o));
+
+    rocksdb_options_set_stats_dump_period_sec(copy, 218);
+    CheckCondition(218 == rocksdb_options_get_stats_dump_period_sec(copy));
+    CheckCondition(18 == rocksdb_options_get_stats_dump_period_sec(o));
+
+    rocksdb_options_set_stats_persist_period_sec(copy, 600);
+    CheckCondition(600 == rocksdb_options_get_stats_persist_period_sec(copy));
+    CheckCondition(5 == rocksdb_options_get_stats_persist_period_sec(o));
+
+    rocksdb_options_set_advise_random_on_open(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_advise_random_on_open(copy));
+    CheckCondition(1 == rocksdb_options_get_advise_random_on_open(o));
+
+    rocksdb_options_set_access_hint_on_compaction_start(copy, 2);
+    CheckCondition(2 ==
+                   rocksdb_options_get_access_hint_on_compaction_start(copy));
+    CheckCondition(3 == rocksdb_options_get_access_hint_on_compaction_start(o));
+
+    rocksdb_options_set_use_adaptive_mutex(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_use_adaptive_mutex(copy));
+    CheckCondition(1 == rocksdb_options_get_use_adaptive_mutex(o));
+
+    rocksdb_options_set_bytes_per_sync(copy, 219);
+    CheckCondition(219 == rocksdb_options_get_bytes_per_sync(copy));
+    CheckCondition(19 == rocksdb_options_get_bytes_per_sync(o));
+
+    rocksdb_options_set_wal_bytes_per_sync(copy, 120);
+    CheckCondition(120 == rocksdb_options_get_wal_bytes_per_sync(copy));
+    CheckCondition(20 == rocksdb_options_get_wal_bytes_per_sync(o));
+
+    rocksdb_options_set_writable_file_max_buffer_size(copy, 121);
+    CheckCondition(121 ==
+                   rocksdb_options_get_writable_file_max_buffer_size(copy));
+    CheckCondition(21 == rocksdb_options_get_writable_file_max_buffer_size(o));
+
+    rocksdb_options_set_allow_concurrent_memtable_write(copy, 0);
+    CheckCondition(0 ==
+                   rocksdb_options_get_allow_concurrent_memtable_write(copy));
+    CheckCondition(1 == rocksdb_options_get_allow_concurrent_memtable_write(o));
+
+    rocksdb_options_set_enable_write_thread_adaptive_yield(copy, 0);
+    CheckCondition(
+        0 == rocksdb_options_get_enable_write_thread_adaptive_yield(copy));
+    CheckCondition(1 ==
+                   rocksdb_options_get_enable_write_thread_adaptive_yield(o));
+
+    rocksdb_options_set_max_sequential_skip_in_iterations(copy, 122);
+    CheckCondition(122 ==
+                   rocksdb_options_get_max_sequential_skip_in_iterations(copy));
+    CheckCondition(22 ==
+                   rocksdb_options_get_max_sequential_skip_in_iterations(o));
+
+    rocksdb_options_set_disable_auto_compactions(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_disable_auto_compactions(copy));
+    CheckCondition(1 == rocksdb_options_get_disable_auto_compactions(o));
+
+    rocksdb_options_set_optimize_filters_for_hits(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_optimize_filters_for_hits(copy));
+    CheckCondition(1 == rocksdb_options_get_optimize_filters_for_hits(o));
+
+    rocksdb_options_set_delete_obsolete_files_period_micros(copy, 123);
+    CheckCondition(
+        123 == rocksdb_options_get_delete_obsolete_files_period_micros(copy));
+    CheckCondition(23 ==
+                   rocksdb_options_get_delete_obsolete_files_period_micros(o));
+
+    rocksdb_options_set_memtable_prefix_bloom_size_ratio(copy, 4.0);
+    CheckCondition(4.0 ==
+                   rocksdb_options_get_memtable_prefix_bloom_size_ratio(copy));
+    CheckCondition(2.0 ==
+                   rocksdb_options_get_memtable_prefix_bloom_size_ratio(o));
+
+    rocksdb_options_set_max_compaction_bytes(copy, 124);
+    CheckCondition(124 == rocksdb_options_get_max_compaction_bytes(copy));
+    CheckCondition(24 == rocksdb_options_get_max_compaction_bytes(o));
+
+    rocksdb_options_set_memtable_huge_page_size(copy, 125);
+    CheckCondition(125 == rocksdb_options_get_memtable_huge_page_size(copy));
+    CheckCondition(25 == rocksdb_options_get_memtable_huge_page_size(o));
+
+    rocksdb_options_set_max_successive_merges(copy, 126);
+    CheckCondition(126 == rocksdb_options_get_max_successive_merges(copy));
+    CheckCondition(26 == rocksdb_options_get_max_successive_merges(o));
+
+    rocksdb_options_set_bloom_locality(copy, 127);
+    CheckCondition(127 == rocksdb_options_get_bloom_locality(copy));
+    CheckCondition(27 == rocksdb_options_get_bloom_locality(o));
+
+    rocksdb_options_set_inplace_update_support(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_inplace_update_support(copy));
+    CheckCondition(1 == rocksdb_options_get_inplace_update_support(o));
+
+    rocksdb_options_set_inplace_update_num_locks(copy, 128);
+    CheckCondition(128 == rocksdb_options_get_inplace_update_num_locks(copy));
+    CheckCondition(28 == rocksdb_options_get_inplace_update_num_locks(o));
+
+    rocksdb_options_set_report_bg_io_stats(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_report_bg_io_stats(copy));
+    CheckCondition(1 == rocksdb_options_get_report_bg_io_stats(o));
+
+    rocksdb_options_set_wal_recovery_mode(copy, 1);
+    CheckCondition(1 == rocksdb_options_get_wal_recovery_mode(copy));
+    CheckCondition(2 == rocksdb_options_get_wal_recovery_mode(o));
+
+    rocksdb_options_set_compression(copy, 4);
+    CheckCondition(4 == rocksdb_options_get_compression(copy));
+    CheckCondition(5 == rocksdb_options_get_compression(o));
+
+    rocksdb_options_set_bottommost_compression(copy, 3);
+    CheckCondition(3 == rocksdb_options_get_bottommost_compression(copy));
+    CheckCondition(4 == rocksdb_options_get_bottommost_compression(o));
+
+    rocksdb_options_set_compaction_style(copy, 1);
+    CheckCondition(1 == rocksdb_options_get_compaction_style(copy));
+    CheckCondition(2 == rocksdb_options_get_compaction_style(o));
+
+    rocksdb_options_set_atomic_flush(copy, 0);
+    CheckCondition(0 == rocksdb_options_get_atomic_flush(copy));
+    CheckCondition(1 == rocksdb_options_get_atomic_flush(o));
+
     rocksdb_options_destroy(copy);
     rocksdb_options_destroy(o);
   }
 
+  StartPhase("read_options");
+  {
+    rocksdb_readoptions_t* ro;
+    ro = rocksdb_readoptions_create();
+
+    rocksdb_readoptions_set_verify_checksums(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_verify_checksums(ro));
+
+    rocksdb_readoptions_set_fill_cache(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_fill_cache(ro));
+
+    rocksdb_readoptions_set_read_tier(ro, 2);
+    CheckCondition(2 == rocksdb_readoptions_get_read_tier(ro));
+
+    rocksdb_readoptions_set_tailing(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_tailing(ro));
+
+    rocksdb_readoptions_set_readahead_size(ro, 100);
+    CheckCondition(100 == rocksdb_readoptions_get_readahead_size(ro));
+
+    rocksdb_readoptions_set_prefix_same_as_start(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_prefix_same_as_start(ro));
+
+    rocksdb_readoptions_set_pin_data(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_pin_data(ro));
+
+    rocksdb_readoptions_set_total_order_seek(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_total_order_seek(ro));
+
+    rocksdb_readoptions_set_max_skippable_internal_keys(ro, 200);
+    CheckCondition(200 ==
+                   rocksdb_readoptions_get_max_skippable_internal_keys(ro));
+
+    rocksdb_readoptions_set_background_purge_on_iterator_cleanup(ro, 1);
+    CheckCondition(
+        1 == rocksdb_readoptions_get_background_purge_on_iterator_cleanup(ro));
+
+    rocksdb_readoptions_set_ignore_range_deletions(ro, 1);
+    CheckCondition(1 == rocksdb_readoptions_get_ignore_range_deletions(ro));
+
+    rocksdb_readoptions_destroy(ro);
+  }
+
+  StartPhase("write_options");
+  {
+    rocksdb_writeoptions_t* wo;
+    wo = rocksdb_writeoptions_create();
+
+    rocksdb_writeoptions_set_sync(wo, 1);
+    CheckCondition(1 == rocksdb_writeoptions_get_sync(wo));
+
+    rocksdb_writeoptions_disable_WAL(wo, 1);
+    CheckCondition(1 == rocksdb_writeoptions_get_disable_WAL(wo));
+
+    rocksdb_writeoptions_set_ignore_missing_column_families(wo, 1);
+    CheckCondition(1 ==
+                   rocksdb_writeoptions_get_ignore_missing_column_families(wo));
+
+    rocksdb_writeoptions_set_no_slowdown(wo, 1);
+    CheckCondition(1 == rocksdb_writeoptions_get_no_slowdown(wo));
+
+    rocksdb_writeoptions_set_low_pri(wo, 1);
+    CheckCondition(1 == rocksdb_writeoptions_get_low_pri(wo));
+
+    rocksdb_writeoptions_set_memtable_insert_hint_per_batch(wo, 1);
+    CheckCondition(1 ==
+                   rocksdb_writeoptions_get_memtable_insert_hint_per_batch(wo));
+
+    rocksdb_writeoptions_destroy(wo);
+  }
+
+  StartPhase("compact_options");
+  {
+    rocksdb_compactoptions_t* co;
+    co = rocksdb_compactoptions_create();
+
+    rocksdb_compactoptions_set_exclusive_manual_compaction(co, 1);
+    CheckCondition(1 ==
+                   rocksdb_compactoptions_get_exclusive_manual_compaction(co));
+
+    rocksdb_compactoptions_set_bottommost_level_compaction(co, 1);
+    CheckCondition(1 ==
+                   rocksdb_compactoptions_get_bottommost_level_compaction(co));
+
+    rocksdb_compactoptions_set_change_level(co, 1);
+    CheckCondition(1 == rocksdb_compactoptions_get_change_level(co));
+
+    rocksdb_compactoptions_set_target_level(co, 1);
+    CheckCondition(1 == rocksdb_compactoptions_get_target_level(co));
+
+    rocksdb_compactoptions_destroy(co);
+  }
+
+  StartPhase("flush_options");
+  {
+    rocksdb_flushoptions_t* fo;
+    fo = rocksdb_flushoptions_create();
+
+    rocksdb_flushoptions_set_wait(fo, 1);
+    CheckCondition(1 == rocksdb_flushoptions_get_wait(fo));
+
+    rocksdb_flushoptions_destroy(fo);
+  }
+
+  StartPhase("cache_options");
+  {
+    rocksdb_cache_t* co;
+    co = rocksdb_cache_create_lru(100);
+    CheckCondition(100 == rocksdb_cache_get_capacity(co));
+
+    rocksdb_cache_set_capacity(co, 200);
+    CheckCondition(200 == rocksdb_cache_get_capacity(co));
+
+    rocksdb_cache_destroy(co);
+  }
+
   StartPhase("iterate_upper_bound");
   {
     // Create new empty database

db/column_family.cc

@@ -12,6 +12,7 @@
 #include <algorithm>
 #include <cinttypes>
 #include <limits>
+#include <sstream>
 #include <string>
 #include <vector>
 
@@ -31,9 +32,10 @@
 #include "monitoring/thread_status_util.h"
 #include "options/options_helper.h"
 #include "port/port.h"
-#include "table/block_based/block_based_table_factory.h"
+#include "rocksdb/table.h"
 #include "table/merging_iterator.h"
 #include "util/autovector.h"
+#include "util/cast_util.h"
 #include "util/compression.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -147,6 +149,16 @@ Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options) {
           "should be nonzero if we're using zstd's dictionary generator.");
     }
   }
+
+  if (!CompressionTypeSupported(cf_options.blob_compression_type)) {
+    std::ostringstream oss;
+    oss << "The specified blob compression type "
+        << CompressionTypeToString(cf_options.blob_compression_type)
+        << " is not available.";
+
+    return Status::InvalidArgument(oss.str());
+  }
+
   return Status::OK();
 }
 
@@ -345,8 +357,8 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
     result.max_compaction_bytes = result.target_file_size_base * 25;
   }
 
-  bool is_block_based_table =
-      (result.table_factory->Name() == BlockBasedTableFactory().Name());
+  bool is_block_based_table = (result.table_factory->IsInstanceOf(
+      TableFactory::kBlockBasedTableName()));
 
   const uint64_t kAdjustedTtl = 30 * 24 * 60 * 60;
   if (result.ttl == kDefaultTtl) {
@@ -487,7 +499,8 @@ ColumnFamilyData::ColumnFamilyData(
     Cache* _table_cache, WriteBufferManager* write_buffer_manager,
     const ColumnFamilyOptions& cf_options, const ImmutableDBOptions& db_options,
     const FileOptions& file_options, ColumnFamilySet* column_family_set,
-    BlockCacheTracer* const block_cache_tracer)
+    BlockCacheTracer* const block_cache_tracer,
+    const std::shared_ptr<IOTracer>& io_tracer)
     : id_(id),
       name_(name),
       dummy_versions_(_dummy_versions),
@@ -545,7 +558,7 @@ ColumnFamilyData::ColumnFamilyData(
     internal_stats_.reset(
         new InternalStats(ioptions_.num_levels, db_options.env, this));
     table_cache_.reset(new TableCache(ioptions_, file_options, _table_cache,
-                                      block_cache_tracer));
+                                      block_cache_tracer, io_tracer));
     if (ioptions_.compaction_style == kCompactionStyleLevel) {
       compaction_picker_.reset(
           new LevelCompactionPicker(ioptions_, &internal_comparator_));
@@ -1042,17 +1055,19 @@ void ColumnFamilyData::CreateNewMemtable(
 }
 
 bool ColumnFamilyData::NeedsCompaction() const {
-  return compaction_picker_->NeedsCompaction(current_->storage_info());
+  return !mutable_cf_options_.disable_auto_compactions &&
+         compaction_picker_->NeedsCompaction(current_->storage_info());
 }
 
 Compaction* ColumnFamilyData::PickCompaction(
-    const MutableCFOptions& mutable_options, LogBuffer* log_buffer) {
+    const MutableCFOptions& mutable_options,
+    const MutableDBOptions& mutable_db_options, LogBuffer* log_buffer) {
   SequenceNumber earliest_mem_seqno =
       std::min(mem_->GetEarliestSequenceNumber(),
                imm_.current()->GetEarliestSequenceNumber(false));
   auto* result = compaction_picker_->PickCompaction(
-      GetName(), mutable_options, current_->storage_info(), log_buffer,
-      earliest_mem_seqno);
+      GetName(), mutable_options, mutable_db_options, current_->storage_info(),
+      log_buffer, earliest_mem_seqno);
   if (result != nullptr) {
     result->SetInputVersion(current_);
   }
@@ -1087,10 +1102,12 @@ Status ColumnFamilyData::RangesOverlapWithMemtables(
       super_version->mem->NewRangeTombstoneIterator(read_opts, read_seq);
   range_del_agg.AddTombstones(
       std::unique_ptr<FragmentedRangeTombstoneIterator>(active_range_del_iter));
-  super_version->imm->AddRangeTombstoneIterators(read_opts, nullptr /* arena */,
-                                                 &range_del_agg);
-
   Status status;
+  status = super_version->imm->AddRangeTombstoneIterators(
+      read_opts, nullptr /* arena */, &range_del_agg);
+  // AddRangeTombstoneIterators always return Status::OK.
+  assert(status.ok());
+
   for (size_t i = 0; i < ranges.size() && status.ok() && !*overlap; ++i) {
     auto* vstorage = super_version->current->storage_info();
     auto* ucmp = vstorage->InternalComparator()->user_comparator();
@@ -1101,7 +1118,8 @@ Status ColumnFamilyData::RangesOverlapWithMemtables(
     ParsedInternalKey seek_result;
     if (status.ok()) {
       if (memtable_iter->Valid() &&
-          !ParseInternalKey(memtable_iter->key(), &seek_result)) {
+          ParseInternalKey(memtable_iter->key(), &seek_result) !=
+              Status::OK()) {
         status = Status::Corruption("DB have corrupted keys");
       }
     }
@@ -1122,14 +1140,16 @@ const int ColumnFamilyData::kCompactAllLevels = -1;
 const int ColumnFamilyData::kCompactToBaseLevel = -2;
 
 Compaction* ColumnFamilyData::CompactRange(
-    const MutableCFOptions& mutable_cf_options, int input_level,
+    const MutableCFOptions& mutable_cf_options,
+    const MutableDBOptions& mutable_db_options, int input_level,
     int output_level, const CompactRangeOptions& compact_range_options,
     const InternalKey* begin, const InternalKey* end,
     InternalKey** compaction_end, bool* conflict,
     uint64_t max_file_num_to_ignore) {
   auto* result = compaction_picker_->CompactRange(
-      GetName(), mutable_cf_options, current_->storage_info(), input_level,
-      output_level, compact_range_options, begin, end, compaction_end, conflict,
+      GetName(), mutable_cf_options, mutable_db_options,
+      current_->storage_info(), input_level, output_level,
+      compact_range_options, begin, end, compaction_end, conflict,
       max_file_num_to_ignore);
   if (result != nullptr) {
     result->SetInputVersion(current_);
@@ -1299,7 +1319,8 @@ Status ColumnFamilyData::ValidateOptions(
   }
 
   if (cf_options.ttl > 0 && cf_options.ttl != kDefaultTtl) {
-    if (cf_options.table_factory->Name() != BlockBasedTableFactory().Name()) {
+    if (!cf_options.table_factory->IsInstanceOf(
+            TableFactory::kBlockBasedTableName())) {
       return Status::NotSupported(
           "TTL is only supported in Block-Based Table format. ");
     }
@@ -1307,7 +1328,8 @@ Status ColumnFamilyData::ValidateOptions(
 
   if (cf_options.periodic_compaction_seconds > 0 &&
       cf_options.periodic_compaction_seconds != kDefaultPeriodicCompSecs) {
-    if (cf_options.table_factory->Name() != BlockBasedTableFactory().Name()) {
+    if (!cf_options.table_factory->IsInstanceOf(
+            TableFactory::kBlockBasedTableName())) {
       return Status::NotSupported(
           "Periodic Compaction is only supported in "
           "Block-Based Table format. ");
@@ -1399,12 +1421,13 @@ ColumnFamilySet::ColumnFamilySet(const std::string& dbname,
                                  Cache* table_cache,
                                  WriteBufferManager* _write_buffer_manager,
                                  WriteController* _write_controller,
-                                 BlockCacheTracer* const block_cache_tracer)
+                                 BlockCacheTracer* const block_cache_tracer,
+                                 const std::shared_ptr<IOTracer>& io_tracer)
     : max_column_family_(0),
       dummy_cfd_(new ColumnFamilyData(
           ColumnFamilyData::kDummyColumnFamilyDataId, "", nullptr, nullptr,
           nullptr, ColumnFamilyOptions(), *db_options, file_options, nullptr,
-          block_cache_tracer)),
+          block_cache_tracer, io_tracer)),
       default_cfd_cache_(nullptr),
       db_name_(dbname),
       db_options_(db_options),
@@ -1412,7 +1435,8 @@ ColumnFamilySet::ColumnFamilySet(const std::string& dbname,
       table_cache_(table_cache),
       write_buffer_manager_(_write_buffer_manager),
       write_controller_(_write_controller),
-      block_cache_tracer_(block_cache_tracer) {
+      block_cache_tracer_(block_cache_tracer),
+      io_tracer_(io_tracer) {
   // initialize linked list
   dummy_cfd_->prev_ = dummy_cfd_;
   dummy_cfd_->next_ = dummy_cfd_;
@@ -1478,7 +1502,7 @@ ColumnFamilyData* ColumnFamilySet::CreateColumnFamily(
   assert(column_families_.find(name) == column_families_.end());
   ColumnFamilyData* new_cfd = new ColumnFamilyData(
       id, name, dummy_versions, table_cache_, write_buffer_manager_, options,
-      *db_options_, file_options_, this, block_cache_tracer_);
+      *db_options_, file_options_, this, block_cache_tracer_, io_tracer_);
   column_families_.insert({name, id});
   column_family_data_.insert({id, new_cfd});
   max_column_family_ = std::max(max_column_family_, id);
@@ -1546,7 +1570,7 @@ ColumnFamilyHandle* ColumnFamilyMemTablesImpl::GetColumnFamilyHandle() {
 uint32_t GetColumnFamilyID(ColumnFamilyHandle* column_family) {
   uint32_t column_family_id = 0;
   if (column_family != nullptr) {
-    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+    auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
     column_family_id = cfh->GetID();
   }
   return column_family_id;

db/column_family.h

@@ -387,6 +387,7 @@ class ColumnFamilyData {
   bool NeedsCompaction() const;
   // REQUIRES: DB mutex held
   Compaction* PickCompaction(const MutableCFOptions& mutable_options,
+                             const MutableDBOptions& mutable_db_options,
                              LogBuffer* log_buffer);
 
   // Check if the passed range overlap with any running compactions.
@@ -412,6 +413,7 @@ class ColumnFamilyData {
   static const int kCompactToBaseLevel;
   // REQUIRES: DB mutex held
   Compaction* CompactRange(const MutableCFOptions& mutable_cf_options,
+                           const MutableDBOptions& mutable_db_options,
                            int input_level, int output_level,
                            const CompactRangeOptions& compact_range_options,
                            const InternalKey* begin, const InternalKey* end,
@@ -516,7 +518,8 @@ class ColumnFamilyData {
                    const ImmutableDBOptions& db_options,
                    const FileOptions& file_options,
                    ColumnFamilySet* column_family_set,
-                   BlockCacheTracer* const block_cache_tracer);
+                   BlockCacheTracer* const block_cache_tracer,
+                   const std::shared_ptr<IOTracer>& io_tracer);
 
   std::vector<std::string> GetDbPaths() const;
 
@@ -649,7 +652,8 @@ class ColumnFamilySet {
                   const FileOptions& file_options, Cache* table_cache,
                   WriteBufferManager* _write_buffer_manager,
                   WriteController* _write_controller,
-                  BlockCacheTracer* const block_cache_tracer);
+                  BlockCacheTracer* const block_cache_tracer,
+                  const std::shared_ptr<IOTracer>& io_tracer);
   ~ColumnFamilySet();
 
   ColumnFamilyData* GetDefault() const;
@@ -713,6 +717,7 @@ class ColumnFamilySet {
   WriteBufferManager* write_buffer_manager_;
   WriteController* write_controller_;
   BlockCacheTracer* const block_cache_tracer_;
+  std::shared_ptr<IOTracer> io_tracer_;
 };
 
 // We use ColumnFamilyMemTablesImpl to provide WriteBatch a way to access

db/column_family_test.cc

@@ -22,26 +22,18 @@
 #include "rocksdb/env.h"
 #include "rocksdb/iterator.h"
 #include "rocksdb/utilities/object_registry.h"
-#include "test_util/fault_injection_test_env.h"
 #include "test_util/sync_point.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"
 #include "util/coding.h"
 #include "util/string_util.h"
+#include "utilities/fault_injection_env.h"
 #include "utilities/merge_operators.h"
 
 namespace ROCKSDB_NAMESPACE {
 
 static const int kValueSize = 1000;
 
-namespace {
-std::string RandomString(Random* rnd, int len) {
-  std::string r;
-  test::RandomString(rnd, len, &r);
-  return r;
-}
-}  // anonymous namespace
-
 // counts how many operations were performed
 class EnvCounter : public EnvWrapper {
  public:
@@ -87,7 +79,12 @@ class ColumnFamilyTestBase : public testing::Test {
     std::vector<ColumnFamilyDescriptor> column_families;
     for (auto h : handles_) {
       ColumnFamilyDescriptor cfdescriptor;
-      h->GetDescriptor(&cfdescriptor);
+      Status s = h->GetDescriptor(&cfdescriptor);
+#ifdef ROCKSDB_LITE
+      EXPECT_TRUE(s.IsNotSupported());
+#else
+      EXPECT_OK(s);
+#endif  // ROCKSDB_LITE
       column_families.push_back(cfdescriptor);
     }
     Close();
@@ -109,11 +106,11 @@ class ColumnFamilyTestBase : public testing::Test {
       // preserves the implementation that was in place when all of the
       // magic values in this file were picked.
       *storage = std::string(kValueSize, ' ');
-      return Slice(*storage);
     } else {
       Random r(k);
-      return test::RandomString(&r, kValueSize, storage);
+      *storage = r.RandomString(kValueSize);
     }
+    return Slice(*storage);
   }
 
   void Build(int base, int n, int flush_every = 0) {
@@ -122,7 +119,7 @@ class ColumnFamilyTestBase : public testing::Test {
 
     for (int i = 0; i < n; i++) {
       if (flush_every != 0 && i != 0 && i % flush_every == 0) {
-        DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
+        DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
         dbi->TEST_FlushMemTable();
       }
 
@@ -176,7 +173,7 @@ class ColumnFamilyTestBase : public testing::Test {
   void Close() {
     for (auto h : handles_) {
       if (h) {
-        db_->DestroyColumnFamilyHandle(h);
+        ASSERT_OK(db_->DestroyColumnFamilyHandle(h));
       }
     }
     handles_.clear();
@@ -227,7 +224,7 @@ class ColumnFamilyTestBase : public testing::Test {
     Open({"default"});
   }
 
-  DBImpl* dbfull() { return reinterpret_cast<DBImpl*>(db_); }
+  DBImpl* dbfull() { return static_cast_with_check<DBImpl>(db_); }
 
   int GetProperty(int cf, std::string property) {
     std::string value;
@@ -287,9 +284,11 @@ class ColumnFamilyTestBase : public testing::Test {
       // Verify the CF options of the returned CF handle.
       ColumnFamilyDescriptor desc;
       ASSERT_OK(handles_[cfi]->GetDescriptor(&desc));
-      RocksDBOptionsParser::VerifyCFOptions(ConfigOptions(), desc.options,
-                                            current_cf_opt);
-
+      // Need to sanitize the default column family options before comparing
+      // them.
+      ASSERT_OK(RocksDBOptionsParser::VerifyCFOptions(
+          ConfigOptions(), desc.options,
+          SanitizeOptions(dbfull()->immutable_db_options(), current_cf_opt)));
 #endif  // !ROCKSDB_LITE
       cfi++;
     }
@@ -315,7 +314,7 @@ class ColumnFamilyTestBase : public testing::Test {
   void DropColumnFamilies(const std::vector<int>& cfs) {
     for (auto cf : cfs) {
       ASSERT_OK(db_->DropColumnFamily(handles_[cf]));
-      db_->DestroyColumnFamilyHandle(handles_[cf]);
+      ASSERT_OK(db_->DestroyColumnFamilyHandle(handles_[cf]));
       handles_[cf] = nullptr;
       names_[cf] = "";
     }
@@ -329,14 +328,14 @@ class ColumnFamilyTestBase : public testing::Test {
       // 10 bytes for key, rest is value
       if (!save) {
         ASSERT_OK(Put(cf, test::RandomKey(&rnd_, 11),
-                      RandomString(&rnd_, key_value_size - 10)));
+                      rnd_.RandomString(key_value_size - 10)));
       } else {
         std::string key = test::RandomKey(&rnd_, 11);
         keys_[cf].insert(key);
-        ASSERT_OK(Put(cf, key, RandomString(&rnd_, key_value_size - 10)));
+        ASSERT_OK(Put(cf, key, rnd_.RandomString(key_value_size - 10)));
       }
     }
-    db_->FlushWAL(false);
+    ASSERT_OK(db_->FlushWAL(/*sync=*/false));
   }
 
 #ifndef ROCKSDB_LITE  // TEST functions in DB are not supported in lite
@@ -570,7 +569,7 @@ TEST_P(ColumnFamilyTest, DontReuseColumnFamilyID) {
     Open();
     CreateColumnFamilies({"one", "two", "three"});
     for (size_t i = 0; i < handles_.size(); ++i) {
-      auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(handles_[i]);
+      auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(handles_[i]);
       ASSERT_EQ(i, cfh->GetID());
     }
     if (iter == 1) {
@@ -586,7 +585,7 @@ TEST_P(ColumnFamilyTest, DontReuseColumnFamilyID) {
     CreateColumnFamilies({"three2"});
     // ID 3 that was used for dropped column family "three" should not be
     // reused
-    auto cfh3 = reinterpret_cast<ColumnFamilyHandleImpl*>(handles_[3]);
+    auto cfh3 = static_cast_with_check<ColumnFamilyHandleImpl>(handles_[3]);
     ASSERT_EQ(4U, cfh3->GetID());
     Close();
     Destroy();
@@ -658,7 +657,7 @@ TEST_P(FlushEmptyCFTestWithParam, FlushEmptyCFTest) {
   Flush(0);
   ASSERT_OK(Put(1, "bar", "v3"));  // seqID 4
   ASSERT_OK(Put(1, "foo", "v4"));  // seqID 5
-  db_->FlushWAL(false);
+  ASSERT_OK(db_->FlushWAL(/*sync=*/false));
 
   // Preserve file system state up to here to simulate a crash condition.
   fault_env->SetFilesystemActive(false);
@@ -721,7 +720,7 @@ TEST_P(FlushEmptyCFTestWithParam, FlushEmptyCFTest2) {
   // Write to log file D
   ASSERT_OK(Put(1, "bar", "v4"));  // seqID 7
   ASSERT_OK(Put(1, "bar", "v5"));  // seqID 8
-  db_->FlushWAL(false);
+  ASSERT_OK(db_->FlushWAL(/*sync=*/false));
   // Preserve file system state up to here to simulate a crash condition.
   fault_env->SetFilesystemActive(false);
   std::vector<std::string> names;
@@ -850,13 +849,15 @@ TEST_P(ColumnFamilyTest, WriteBatchFailure) {
   Open();
   CreateColumnFamiliesAndReopen({"one", "two"});
   WriteBatch batch;
-  batch.Put(handles_[0], Slice("existing"), Slice("column-family"));
-  batch.Put(handles_[1], Slice("non-existing"), Slice("column-family"));
+  ASSERT_OK(batch.Put(handles_[0], Slice("existing"), Slice("column-family")));
+  ASSERT_OK(
+      batch.Put(handles_[1], Slice("non-existing"), Slice("column-family")));
   ASSERT_OK(db_->Write(WriteOptions(), &batch));
   DropColumnFamilies({1});
   WriteOptions woptions_ignore_missing_cf;
   woptions_ignore_missing_cf.ignore_missing_column_families = true;
-  batch.Put(handles_[0], Slice("still here"), Slice("column-family"));
+  ASSERT_OK(
+      batch.Put(handles_[0], Slice("still here"), Slice("column-family")));
   ASSERT_OK(db_->Write(woptions_ignore_missing_cf, &batch));
   ASSERT_EQ("column-family", Get(0, "still here"));
   Status s = db_->Write(WriteOptions(), &batch);
@@ -895,10 +896,10 @@ TEST_P(ColumnFamilyTest, IgnoreRecoveredLog) {
   ASSERT_OK(env_->CreateDirIfMissing(dbname_));
   ASSERT_OK(env_->CreateDirIfMissing(backup_logs));
   std::vector<std::string> old_files;
-  env_->GetChildren(backup_logs, &old_files);
+  ASSERT_OK(env_->GetChildren(backup_logs, &old_files));
   for (auto& file : old_files) {
     if (file != "." && file != "..") {
-      env_->DeleteFile(backup_logs + "/" + file);
+      ASSERT_OK(env_->DeleteFile(backup_logs + "/" + file));
     }
   }
 
@@ -926,7 +927,7 @@ TEST_P(ColumnFamilyTest, IgnoreRecoveredLog) {
 
   // copy the logs to backup
   std::vector<std::string> logs;
-  env_->GetChildren(db_options_.wal_dir, &logs);
+  ASSERT_OK(env_->GetChildren(db_options_.wal_dir, &logs));
   for (auto& log : logs) {
     if (log != ".." && log != ".") {
       CopyFile(db_options_.wal_dir + "/" + log, backup_logs + "/" + log);
@@ -991,6 +992,7 @@ TEST_P(ColumnFamilyTest, FlushTest) {
     ASSERT_OK(Put(0, "foofoo", "bar"));
 
     for (auto* it : iterators) {
+      ASSERT_OK(it->status());
       delete it;
     }
   }
@@ -1088,8 +1090,8 @@ TEST_P(ColumnFamilyTest, CrashAfterFlush) {
   CreateColumnFamilies({"one"});
 
   WriteBatch batch;
-  batch.Put(handles_[0], Slice("foo"), Slice("bar"));
-  batch.Put(handles_[1], Slice("foo"), Slice("bar"));
+  ASSERT_OK(batch.Put(handles_[0], Slice("foo"), Slice("bar")));
+  ASSERT_OK(batch.Put(handles_[1], Slice("foo"), Slice("bar")));
   ASSERT_OK(db_->Write(WriteOptions(), &batch));
   Flush(0);
   fault_env->SetFilesystemActive(false);
@@ -2075,6 +2077,7 @@ std::string IterStatus(Iterator* iter) {
   if (iter->Valid()) {
     result = iter->key().ToString() + "->" + iter->value().ToString();
   } else {
+    EXPECT_OK(iter->status());
     result = "(invalid)";
   }
   return result;
@@ -2329,7 +2332,7 @@ TEST_P(ColumnFamilyTest, ReadDroppedColumnFamily) {
         ASSERT_OK(db_->DropColumnFamily(handles_[2]));
       } else {
         // delete CF two
-        db_->DestroyColumnFamilyHandle(handles_[2]);
+        ASSERT_OK(db_->DestroyColumnFamilyHandle(handles_[2]));
         handles_[2] = nullptr;
       }
       // Make sure iterator created can still be used.
@@ -2385,7 +2388,6 @@ TEST_P(ColumnFamilyTest, LiveIteratorWithDroppedColumnFamily) {
   // 1MB should create ~10 files for each CF
   int kKeysNum = 10000;
   PutRandomData(1, kKeysNum, 100);
-
   {
     std::unique_ptr<Iterator> iterator(
         db_->NewIterator(ReadOptions(), handles_[1]));
@@ -2432,6 +2434,9 @@ TEST_P(ColumnFamilyTest, FlushAndDropRaceCondition) {
 
   env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask, &sleeping_task,
                  Env::Priority::HIGH);
+  // Make sure the task is sleeping. Otherwise, it might start to execute
+  // after sleeping_task.WaitUntilDone() and cause TSAN warning.
+  sleeping_task.WaitUntilSleeping();
 
   // 1MB should create ~10 files for each CF
   int kKeysNum = 10000;
@@ -2446,6 +2451,9 @@ TEST_P(ColumnFamilyTest, FlushAndDropRaceCondition) {
   // now we sleep again. this is just so we're certain that flush job finished
   env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask, &sleeping_task,
                  Env::Priority::HIGH);
+  // Make sure the task is sleeping. Otherwise, it might start to execute
+  // after sleeping_task.WaitUntilDone() and cause TSAN warning.
+  sleeping_task.WaitUntilSleeping();
   sleeping_task.WakeUp();
   sleeping_task.WaitUntilDone();
 
@@ -2995,6 +3003,9 @@ TEST_P(ColumnFamilyTest, FlushCloseWALFiles) {
   test::SleepingBackgroundTask sleeping_task;
   env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask, &sleeping_task,
                  Env::Priority::HIGH);
+  // Make sure the task is sleeping. Otherwise, it might start to execute
+  // after sleeping_task.WaitUntilDone() and cause TSAN warning.
+  sleeping_task.WaitUntilSleeping();
 
   WriteOptions wo;
   wo.sync = true;
@@ -3027,6 +3038,7 @@ TEST_P(ColumnFamilyTest, IteratorCloseWALFile1) {
   ASSERT_OK(Put(1, "fodor", "mirko"));
   // Create an iterator holding the current super version.
   Iterator* it = db_->NewIterator(ReadOptions(), handles_[1]);
+  ASSERT_OK(it->status());
   // A flush will make `it` hold the last reference of its super version.
   Flush(1);
 
@@ -3040,6 +3052,9 @@ TEST_P(ColumnFamilyTest, IteratorCloseWALFile1) {
   test::SleepingBackgroundTask sleeping_task;
   env_->Schedule(&test::SleepingBackgroundTask::DoSleepTask, &sleeping_task,
                  Env::Priority::HIGH);
+  // Make sure the task is sleeping. Otherwise, it might start to execute
+  // after sleeping_task.WaitUntilDone() and cause TSAN warning.
+  sleeping_task.WaitUntilSleeping();
 
   WriteOptions wo;
   wo.sync = true;
@@ -3076,6 +3091,7 @@ TEST_P(ColumnFamilyTest, IteratorCloseWALFile2) {
   ReadOptions ro;
   ro.background_purge_on_iterator_cleanup = true;
   Iterator* it = db_->NewIterator(ro, handles_[1]);
+  ASSERT_OK(it->status());
   // A flush will make `it` hold the last reference of its super version.
   Flush(1);
 
@@ -3170,6 +3186,8 @@ TEST_P(ColumnFamilyTest, ForwardIteratorCloseWALFile) {
   // Deleting the iterator will clear its super version, triggering
   // closing all files
   it->Seek("");
+  ASSERT_OK(it->status());
+
   ASSERT_EQ(2, env.num_open_wal_file_.load());
   ASSERT_EQ(0, env.delete_count_.load());
 
@@ -3200,8 +3218,8 @@ TEST_P(ColumnFamilyTest, LogSyncConflictFlush) {
   Open();
   CreateColumnFamiliesAndReopen({"one", "two"});
 
-  Put(0, "", "");
-  Put(1, "foo", "bar");
+  ASSERT_OK(Put(0, "", ""));
+  ASSERT_OK(Put(1, "foo", "bar"));
 
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency(
       {{"DBImpl::SyncWAL:BeforeMarkLogsSynced:1",
@@ -3211,11 +3229,11 @@ TEST_P(ColumnFamilyTest, LogSyncConflictFlush) {
 
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
 
-  ROCKSDB_NAMESPACE::port::Thread thread([&] { db_->SyncWAL(); });
+  ROCKSDB_NAMESPACE::port::Thread thread([&] { ASSERT_OK(db_->SyncWAL()); });
 
   TEST_SYNC_POINT("ColumnFamilyTest::LogSyncConflictFlush:1");
   Flush(1);
-  Put(1, "foo", "bar");
+  ASSERT_OK(Put(1, "foo", "bar"));
   Flush(1);
 
   TEST_SYNC_POINT("ColumnFamilyTest::LogSyncConflictFlush:2");
@@ -3298,7 +3316,7 @@ TEST_P(ColumnFamilyTest, DISABLED_LogTruncationTest) {
   Close();
 
   // cleanup
-  env_->DeleteDir(backup_logs);
+  ASSERT_OK(env_->DeleteDir(backup_logs));
 }
 
 TEST_P(ColumnFamilyTest, DefaultCfPathsTest) {
@@ -3354,15 +3372,17 @@ TEST_P(ColumnFamilyTest, MultipleCFPathsTest) {
 
   // Re-open and verify the keys.
   Reopen({ColumnFamilyOptions(), cf_opt1, cf_opt2});
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
   for (int cf = 1; cf != 3; ++cf) {
     ReadOptions read_options;
     read_options.readahead_size = 0;
     auto it = dbi->NewIterator(read_options, handles_[cf]);
     for (it->SeekToFirst(); it->Valid(); it->Next()) {
+      ASSERT_OK(it->status());
       Slice key(it->key());
       ASSERT_NE(keys_[cf].end(), keys_[cf].find(key.ToString()));
     }
+    ASSERT_OK(it->status());
     delete it;
 
     for (const auto& key : keys_[cf]) {

db/compact_files_test.cc

@@ -16,6 +16,7 @@
 #include "rocksdb/env.h"
 #include "test_util/sync_point.h"
 #include "test_util/testharness.h"
+#include "util/cast_util.h"
 #include "util/string_util.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -148,7 +149,7 @@ TEST_F(CompactFilesTest, ObsoleteFiles) {
 
   auto l0_files = collector->GetFlushedFiles();
   ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files, 1));
-  reinterpret_cast<DBImpl*>(db)->TEST_WaitForCompact();
+  static_cast_with_check<DBImpl>(db)->TEST_WaitForCompact();
 
   // verify all compaction input files are deleted
   for (auto fname : l0_files) {
@@ -183,13 +184,13 @@ TEST_F(CompactFilesTest, NotCutOutputOnLevel0) {
   for (int i = 0; i < 500; ++i) {
     db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
   }
-  reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
+  static_cast_with_check<DBImpl>(db)->TEST_WaitForFlushMemTable();
   auto l0_files_1 = collector->GetFlushedFiles();
   collector->ClearFlushedFiles();
   for (int i = 0; i < 500; ++i) {
     db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
   }
-  reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
+  static_cast_with_check<DBImpl>(db)->TEST_WaitForFlushMemTable();
   auto l0_files_2 = collector->GetFlushedFiles();
   ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files_1, 0));
   ASSERT_OK(db->CompactFiles(CompactionOptions(), l0_files_2, 0));
@@ -383,7 +384,7 @@ TEST_F(CompactFilesTest, GetCompactionJobInfo) {
   for (int i = 0; i < 500; ++i) {
     db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
   }
-  reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
+  static_cast_with_check<DBImpl>(db)->TEST_WaitForFlushMemTable();
   auto l0_files_1 = collector->GetFlushedFiles();
   CompactionOptions co;
   co.compression = CompressionType::kLZ4Compression;

db/compacted_db_impl.cc

@@ -5,9 +5,11 @@
 
 #ifndef ROCKSDB_LITE
 #include "db/compacted_db_impl.h"
+
 #include "db/db_impl/db_impl.h"
 #include "db/version_set.h"
 #include "table/get_context.h"
+#include "util/cast_util.h"
 
 namespace ROCKSDB_NAMESPACE {
 
@@ -39,8 +41,11 @@ Status CompactedDBImpl::Get(const ReadOptions& options, ColumnFamilyHandle*,
                          GetContext::kNotFound, key, value, nullptr, nullptr,
                          nullptr, true, nullptr, nullptr);
   LookupKey lkey(key, kMaxSequenceNumber);
-  files_.files[FindFile(key)].fd.table_reader->Get(options, lkey.internal_key(),
-                                                   &get_context, nullptr);
+  Status s = files_.files[FindFile(key)].fd.table_reader->Get(
+      options, lkey.internal_key(), &get_context, nullptr);
+  if (!s.ok() && !s.IsNotFound()) {
+    return s;
+  }
   if (get_context.State() == GetContext::kFound) {
     return Status::OK();
   }
@@ -72,10 +77,14 @@ std::vector<Status> CompactedDBImpl::MultiGet(const ReadOptions& options,
                              GetContext::kNotFound, keys[idx], &pinnable_val,
                              nullptr, nullptr, nullptr, true, nullptr, nullptr);
       LookupKey lkey(keys[idx], kMaxSequenceNumber);
-      r->Get(options, lkey.internal_key(), &get_context, nullptr);
-      value.assign(pinnable_val.data(), pinnable_val.size());
-      if (get_context.State() == GetContext::kFound) {
-        statuses[idx] = Status::OK();
+      Status s = r->Get(options, lkey.internal_key(), &get_context, nullptr);
+      if (!s.ok() && !s.IsNotFound()) {
+        statuses[idx] = s;
+      } else {
+        value.assign(pinnable_val.data(), pinnable_val.size());
+        if (get_context.State() == GetContext::kFound) {
+          statuses[idx] = Status::OK();
+        }
       }
     }
     ++idx;
@@ -90,8 +99,8 @@ Status CompactedDBImpl::Init(const Options& options) {
                             ColumnFamilyOptions(options));
   Status s = Recover({cf}, true /* read only */, false, true);
   if (s.ok()) {
-    cfd_ = reinterpret_cast<ColumnFamilyHandleImpl*>(
-              DefaultColumnFamily())->cfd();
+    cfd_ = static_cast_with_check<ColumnFamilyHandleImpl>(DefaultColumnFamily())
+               ->cfd();
     cfd_->InstallSuperVersion(&sv_context, &mutex_);
   }
   mutex_.Unlock();
@@ -147,7 +156,7 @@ Status CompactedDBImpl::Open(const Options& options,
   std::unique_ptr<CompactedDBImpl> db(new CompactedDBImpl(db_options, dbname));
   Status s = db->Init(options);
   if (s.ok()) {
-    db->StartTimedTasks();
+    db->StartPeriodicWorkScheduler();
     ROCKS_LOG_INFO(db->immutable_db_options_.info_log,
                    "Opened the db as fully compacted mode");
     LogFlush(db->immutable_db_options_.info_log);

db/compaction/compaction.cc

@@ -7,12 +7,14 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
+#include "db/compaction/compaction.h"
+
 #include <cinttypes>
 #include <vector>
 
 #include "db/column_family.h"
-#include "db/compaction/compaction.h"
 #include "rocksdb/compaction_filter.h"
+#include "rocksdb/sst_partitioner.h"
 #include "test_util/sync_point.h"
 #include "util/string_util.h"
 
@@ -205,6 +207,7 @@ bool Compaction::IsFullCompaction(
 Compaction::Compaction(VersionStorageInfo* vstorage,
                        const ImmutableCFOptions& _immutable_cf_options,
                        const MutableCFOptions& _mutable_cf_options,
+                       const MutableDBOptions& _mutable_db_options,
                        std::vector<CompactionInputFiles> _inputs,
                        int _output_level, uint64_t _target_file_size,
                        uint64_t _max_compaction_bytes, uint32_t _output_path_id,
@@ -243,7 +246,7 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
     compaction_reason_ = CompactionReason::kManualCompaction;
   }
   if (max_subcompactions_ == 0) {
-    max_subcompactions_ = immutable_cf_options_.max_subcompactions;
+    max_subcompactions_ = _mutable_db_options.max_subcompactions;
   }
   if (!bottommost_level_) {
     // Currently we only enable dictionary compression during compaction to the
@@ -328,6 +331,8 @@ bool Compaction::IsTrivialMove() const {
 
   // assert inputs_.size() == 1
 
+  std::unique_ptr<SstPartitioner> partitioner = CreateSstPartitioner();
+
   for (const auto& file : inputs_.front().files) {
     std::vector<FileMetaData*> file_grand_parents;
     if (output_level_ + 1 >= number_levels_) {
@@ -340,6 +345,13 @@ bool Compaction::IsTrivialMove() const {
     if (compaction_size > max_compaction_bytes_) {
       return false;
     }
+
+    if (partitioner.get() != nullptr) {
+      if (!partitioner->CanDoTrivialMove(file->smallest.user_key(),
+                                         file->largest.user_key())) {
+        return false;
+      }
+    }
   }
 
   return true;
@@ -525,6 +537,21 @@ std::unique_ptr<CompactionFilter> Compaction::CreateCompactionFilter() const {
       context);
 }
 
+std::unique_ptr<SstPartitioner> Compaction::CreateSstPartitioner() const {
+  if (!immutable_cf_options_.sst_partitioner_factory) {
+    return nullptr;
+  }
+
+  SstPartitioner::Context context;
+  context.is_full_compaction = is_full_compaction_;
+  context.is_manual_compaction = is_manual_compaction_;
+  context.output_level = output_level_;
+  context.smallest_user_key = smallest_user_key_;
+  context.largest_user_key = largest_user_key_;
+  return immutable_cf_options_.sst_partitioner_factory->CreatePartitioner(
+      context);
+}
+
 bool Compaction::IsOutputLevelEmpty() const {
   return inputs_.back().level != output_level_ || inputs_.back().empty();
 }

db/compaction/compaction.h

@@ -11,6 +11,7 @@
 #include "db/version_set.h"
 #include "memory/arena.h"
 #include "options/cf_options.h"
+#include "rocksdb/sst_partitioner.h"
 #include "util/autovector.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -71,6 +72,7 @@ class Compaction {
   Compaction(VersionStorageInfo* input_version,
              const ImmutableCFOptions& immutable_cf_options,
              const MutableCFOptions& mutable_cf_options,
+             const MutableDBOptions& mutable_db_options,
              std::vector<CompactionInputFiles> inputs, int output_level,
              uint64_t target_file_size, uint64_t max_compaction_bytes,
              uint32_t output_path_id, CompressionType compression,
@@ -255,6 +257,9 @@ class Compaction {
   // Create a CompactionFilter from compaction_filter_factory
   std::unique_ptr<CompactionFilter> CreateCompactionFilter() const;
 
+  // Create a SstPartitioner from sst_partitioner_factory
+  std::unique_ptr<SstPartitioner> CreateSstPartitioner() const;
+
   // Is the input level corresponding to output_level_ empty?
   bool IsOutputLevelEmpty() const;
 

db/compaction/compaction_iterator.cc

@@ -3,9 +3,11 @@
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
 
+#include "db/compaction/compaction_iterator.h"
+
 #include <cinttypes>
 
-#include "db/compaction/compaction_iterator.h"
+#include "db/blob/blob_file_builder.h"
 #include "db/snapshot_checker.h"
 #include "port/likely.h"
 #include "rocksdb/listener.h"
@@ -36,16 +38,18 @@ CompactionIterator::CompactionIterator(
     SequenceNumber earliest_write_conflict_snapshot,
     const SnapshotChecker* snapshot_checker, Env* env,
     bool report_detailed_time, bool expect_valid_internal_key,
-    CompactionRangeDelAggregator* range_del_agg, const Compaction* compaction,
-    const CompactionFilter* compaction_filter,
+    CompactionRangeDelAggregator* range_del_agg,
+    BlobFileBuilder* blob_file_builder, bool allow_data_in_errors,
+    const Compaction* compaction, const CompactionFilter* compaction_filter,
     const std::atomic<bool>* shutting_down,
     const SequenceNumber preserve_deletes_seqnum,
-    const std::atomic<bool>* manual_compaction_paused,
+    const std::atomic<int>* manual_compaction_paused,
     const std::shared_ptr<Logger> info_log)
     : CompactionIterator(
           input, cmp, merge_helper, last_sequence, snapshots,
           earliest_write_conflict_snapshot, snapshot_checker, env,
           report_detailed_time, expect_valid_internal_key, range_del_agg,
+          blob_file_builder, allow_data_in_errors,
           std::unique_ptr<CompactionProxy>(
               compaction ? new CompactionProxy(compaction) : nullptr),
           compaction_filter, shutting_down, preserve_deletes_seqnum,
@@ -58,11 +62,12 @@ CompactionIterator::CompactionIterator(
     const SnapshotChecker* snapshot_checker, Env* env,
     bool report_detailed_time, bool expect_valid_internal_key,
     CompactionRangeDelAggregator* range_del_agg,
+    BlobFileBuilder* blob_file_builder, bool allow_data_in_errors,
     std::unique_ptr<CompactionProxy> compaction,
     const CompactionFilter* compaction_filter,
     const std::atomic<bool>* shutting_down,
     const SequenceNumber preserve_deletes_seqnum,
-    const std::atomic<bool>* manual_compaction_paused,
+    const std::atomic<int>* manual_compaction_paused,
     const std::shared_ptr<Logger> info_log)
     : input_(input),
       cmp_(cmp),
@@ -74,6 +79,7 @@ CompactionIterator::CompactionIterator(
       report_detailed_time_(report_detailed_time),
       expect_valid_internal_key_(expect_valid_internal_key),
       range_del_agg_(range_del_agg),
+      blob_file_builder_(blob_file_builder),
       compaction_(std::move(compaction)),
       compaction_filter_(compaction_filter),
       shutting_down_(shutting_down),
@@ -83,11 +89,14 @@ CompactionIterator::CompactionIterator(
       current_user_key_snapshot_(0),
       merge_out_iter_(merge_helper_),
       current_key_committed_(false),
-      info_log_(info_log) {
+      info_log_(info_log),
+      allow_data_in_errors_(allow_data_in_errors) {
   assert(compaction_filter_ == nullptr || compaction_ != nullptr);
   assert(snapshots_ != nullptr);
-  bottommost_level_ =
-      compaction_ == nullptr ? false : compaction_->bottommost_level();
+  bottommost_level_ = compaction_ == nullptr
+                          ? false
+                          : compaction_->bottommost_level() &&
+                                !compaction_->allow_ingest_behind();
   if (compaction_ != nullptr) {
     level_ptrs_ = std::vector<size_t>(compaction_->number_levels(), 0);
   }
@@ -142,12 +151,11 @@ void CompactionIterator::Next() {
     if (merge_out_iter_.Valid()) {
       key_ = merge_out_iter_.key();
       value_ = merge_out_iter_.value();
-      bool valid_key __attribute__((__unused__));
-      valid_key =  ParseInternalKey(key_, &ikey_);
+      Status s = ParseInternalKey(key_, &ikey_);
       // MergeUntil stops when it encounters a corrupt key and does not
       // include them in the result, so we expect the keys here to be valid.
-      assert(valid_key);
-      if (!valid_key) {
+      assert(s.ok());
+      if (!s.ok()) {
         ROCKS_LOG_FATAL(info_log_, "Invalid key (%s) in compaction",
                         key_.ToString(true).c_str());
       }
@@ -182,7 +190,7 @@ void CompactionIterator::Next() {
   PrepareOutput();
 }
 
-void CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
+bool CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
                                               Slice* skip_until) {
   if (compaction_filter_ != nullptr &&
       (ikey_.type == kTypeValue || ikey_.type == kTypeBlobIndex)) {
@@ -225,14 +233,31 @@ void CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
       value_.clear();
       iter_stats_.num_record_drop_user++;
     } else if (filter == CompactionFilter::Decision::kChangeValue) {
+      if (ikey_.type == kTypeBlobIndex) {
+        // value transfer from blob file to inlined data
+        ikey_.type = kTypeValue;
+        current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
+      }
       value_ = compaction_filter_value_;
     } else if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil) {
       *need_skip = true;
       compaction_filter_skip_until_.ConvertFromUserKey(kMaxSequenceNumber,
                                                        kValueTypeForSeek);
       *skip_until = compaction_filter_skip_until_.Encode();
+    } else if (filter == CompactionFilter::Decision::kChangeBlobIndex) {
+      if (ikey_.type == kTypeValue) {
+        // value transfer from inlined data to blob file
+        ikey_.type = kTypeBlobIndex;
+        current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
+      }
+      value_ = compaction_filter_value_;
+    } else if (filter == CompactionFilter::Decision::kIOError) {
+      status_ =
+          Status::IOError("Failed to access blob during compaction filter");
+      return false;
     }
   }
+  return true;
 }
 
 void CompactionIterator::NextFromInput() {
@@ -245,20 +270,28 @@ void CompactionIterator::NextFromInput() {
     value_ = input_->value();
     iter_stats_.num_input_records++;
 
-    if (!ParseInternalKey(key_, &ikey_)) {
+    Status pikStatus = ParseInternalKey(key_, &ikey_);
+    if (!pikStatus.ok()) {
+      iter_stats_.num_input_corrupt_records++;
+
       // If `expect_valid_internal_key_` is false, return the corrupted key
       // and let the caller decide what to do with it.
       // TODO(noetzli): We should have a more elegant solution for this.
       if (expect_valid_internal_key_) {
-        assert(!"Corrupted internal key not expected.");
-        status_ = Status::Corruption("Corrupted internal key not expected.");
-        break;
+        std::string msg("Corrupted internal key not expected.");
+        if (allow_data_in_errors_) {
+          msg.append(" Corrupt key: " + ikey_.user_key.ToString(/*hex=*/true) +
+                     ". ");
+          msg.append("key type: " + std::to_string(ikey_.type) + ".");
+          msg.append("seq: " + std::to_string(ikey_.sequence) + ".");
+        }
+        status_ = Status::Corruption(msg.c_str());
+        return;
       }
       key_ = current_key_.SetInternalKey(key_);
       has_current_user_key_ = false;
       current_user_key_sequence_ = kMaxSequenceNumber;
       current_user_key_snapshot_ = 0;
-      iter_stats_.num_input_corrupt_records++;
       valid_ = true;
       break;
     }
@@ -295,8 +328,9 @@ void CompactionIterator::NextFromInput() {
 
       // Apply the compaction filter to the first committed version of the user
       // key.
-      if (current_key_committed_) {
-        InvokeFilterIfNeeded(&need_skip, &skip_until);
+      if (current_key_committed_ &&
+          !InvokeFilterIfNeeded(&need_skip, &skip_until)) {
+        break;
       }
     } else {
       // Update the current key to reflect the new sequence number/type without
@@ -316,8 +350,9 @@ void CompactionIterator::NextFromInput() {
         current_key_committed_ = KeyCommitted(ikey_.sequence);
         // Apply the compaction filter to the first committed version of the
         // user key.
-        if (current_key_committed_) {
-          InvokeFilterIfNeeded(&need_skip, &skip_until);
+        if (current_key_committed_ &&
+            !InvokeFilterIfNeeded(&need_skip, &skip_until)) {
+          break;
         }
       }
     }
@@ -331,8 +366,7 @@ void CompactionIterator::NextFromInput() {
     // If there are no snapshots, then this kv affect visibility at tip.
     // Otherwise, search though all existing snapshots to find the earliest
     // snapshot that is affected by this kv.
-    SequenceNumber last_sequence __attribute__((__unused__));
-    last_sequence = current_user_key_sequence_;
+    SequenceNumber last_sequence = current_user_key_sequence_;
     current_user_key_sequence_ = ikey_.sequence;
     SequenceNumber last_snapshot = current_user_key_snapshot_;
     SequenceNumber prev_snapshot = 0;  // 0 means no previous snapshot
@@ -404,7 +438,8 @@ void CompactionIterator::NextFromInput() {
 
       // Check whether the next key exists, is not corrupt, and is the same key
       // as the single delete.
-      if (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
+      if (input_->Valid() &&
+          ParseInternalKey(input_->key(), &next_ikey) == Status::OK() &&
           cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
         // Check whether the next key belongs to the same snapshot as the
         // SingleDelete.
@@ -544,11 +579,15 @@ void CompactionIterator::NextFromInput() {
       // Handle the case where we have a delete key at the bottom most level
       // We can skip outputting the key iff there are no subsequent puts for this
       // key
+      assert(!compaction_ || compaction_->KeyNotExistsBeyondOutputLevel(
+                                 ikey_.user_key, &level_ptrs_));
       ParsedInternalKey next_ikey;
       input_->Next();
       // Skip over all versions of this key that happen to occur in the same snapshot
       // range as the delete
-      while (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
+      while (!IsPausingManualCompaction() && !IsShuttingDown() &&
+             input_->Valid() &&
+             (ParseInternalKey(input_->key(), &next_ikey) == Status::OK()) &&
              cmp_->Equal(ikey_.user_key, next_ikey.user_key) &&
              (prev_snapshot == 0 ||
               DEFINITELY_NOT_IN_SNAPSHOT(next_ikey.sequence, prev_snapshot))) {
@@ -556,7 +595,8 @@ void CompactionIterator::NextFromInput() {
       }
       // If you find you still need to output a row with this key, we need to output the
       // delete too
-      if (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
+      if (input_->Valid() &&
+          (ParseInternalKey(input_->key(), &next_ikey) == Status::OK()) &&
           cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
         valid_ = true;
         at_next_ = true;
@@ -585,12 +625,11 @@ void CompactionIterator::NextFromInput() {
         //       These will be correctly set below.
         key_ = merge_out_iter_.key();
         value_ = merge_out_iter_.value();
-        bool valid_key __attribute__((__unused__));
-        valid_key = ParseInternalKey(key_, &ikey_);
+        pikStatus = ParseInternalKey(key_, &ikey_);
         // MergeUntil stops when it encounters a corrupt key and does not
         // include them in the result, so we expect the keys here to valid.
-        assert(valid_key);
-        if (!valid_key) {
+        assert(pikStatus.ok());
+        if (!pikStatus.ok()) {
           ROCKS_LOG_FATAL(info_log_, "Invalid key (%s) in compaction",
                           key_.ToString(true).c_str());
         }
@@ -640,20 +679,37 @@ void CompactionIterator::NextFromInput() {
 
 void CompactionIterator::PrepareOutput() {
   if (valid_) {
-    if (compaction_filter_ && ikey_.type == kTypeBlobIndex) {
-      const auto blob_decision = compaction_filter_->PrepareBlobOutput(
-          user_key(), value_, &compaction_filter_value_);
+    if (ikey_.type == kTypeValue) {
+      if (blob_file_builder_) {
+        blob_index_.clear();
+        const Status s =
+            blob_file_builder_->Add(user_key(), value_, &blob_index_);
 
-      if (blob_decision == CompactionFilter::BlobDecision::kCorruption) {
-        status_ = Status::Corruption(
-            "Corrupted blob reference encountered during GC");
-        valid_ = false;
-      } else if (blob_decision == CompactionFilter::BlobDecision::kIOError) {
-        status_ = Status::IOError("Could not relocate blob during GC");
-        valid_ = false;
-      } else if (blob_decision ==
-                 CompactionFilter::BlobDecision::kChangeValue) {
-        value_ = compaction_filter_value_;
+        if (!s.ok()) {
+          status_ = s;
+          valid_ = false;
+        } else if (!blob_index_.empty()) {
+          value_ = blob_index_;
+          ikey_.type = kTypeBlobIndex;
+          current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
+        }
+      }
+    } else if (ikey_.type == kTypeBlobIndex) {
+      if (compaction_filter_) {
+        const auto blob_decision = compaction_filter_->PrepareBlobOutput(
+            user_key(), value_, &compaction_filter_value_);
+
+        if (blob_decision == CompactionFilter::BlobDecision::kCorruption) {
+          status_ = Status::Corruption(
+              "Corrupted blob reference encountered during GC");
+          valid_ = false;
+        } else if (blob_decision == CompactionFilter::BlobDecision::kIOError) {
+          status_ = Status::IOError("Could not relocate blob during GC");
+          valid_ = false;
+        } else if (blob_decision ==
+                   CompactionFilter::BlobDecision::kChangeValue) {
+          value_ = compaction_filter_value_;
+        }
       }
     }
 

db/compaction/compaction_iterator.h

@@ -21,6 +21,8 @@
 
 namespace ROCKSDB_NAMESPACE {
 
+class BlobFileBuilder;
+
 class CompactionIterator {
  public:
   // A wrapper around Compaction. Has a much smaller interface, only what
@@ -59,34 +61,38 @@ class CompactionIterator {
     const Compaction* compaction_;
   };
 
-  CompactionIterator(
-      InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
-      SequenceNumber last_sequence, std::vector<SequenceNumber>* snapshots,
-      SequenceNumber earliest_write_conflict_snapshot,
-      const SnapshotChecker* snapshot_checker, Env* env,
-      bool report_detailed_time, bool expect_valid_internal_key,
-      CompactionRangeDelAggregator* range_del_agg,
-      const Compaction* compaction = nullptr,
-      const CompactionFilter* compaction_filter = nullptr,
-      const std::atomic<bool>* shutting_down = nullptr,
-      const SequenceNumber preserve_deletes_seqnum = 0,
-      const std::atomic<bool>* manual_compaction_paused = nullptr,
-      const std::shared_ptr<Logger> info_log = nullptr);
+  CompactionIterator(InternalIterator* input, const Comparator* cmp,
+                     MergeHelper* merge_helper, SequenceNumber last_sequence,
+                     std::vector<SequenceNumber>* snapshots,
+                     SequenceNumber earliest_write_conflict_snapshot,
+                     const SnapshotChecker* snapshot_checker, Env* env,
+                     bool report_detailed_time, bool expect_valid_internal_key,
+                     CompactionRangeDelAggregator* range_del_agg,
+                     BlobFileBuilder* blob_file_builder,
+                     bool allow_data_in_errors,
+                     const Compaction* compaction = nullptr,
+                     const CompactionFilter* compaction_filter = nullptr,
+                     const std::atomic<bool>* shutting_down = nullptr,
+                     const SequenceNumber preserve_deletes_seqnum = 0,
+                     const std::atomic<int>* manual_compaction_paused = nullptr,
+                     const std::shared_ptr<Logger> info_log = nullptr);
 
   // Constructor with custom CompactionProxy, used for tests.
-  CompactionIterator(
-      InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
-      SequenceNumber last_sequence, std::vector<SequenceNumber>* snapshots,
-      SequenceNumber earliest_write_conflict_snapshot,
-      const SnapshotChecker* snapshot_checker, Env* env,
-      bool report_detailed_time, bool expect_valid_internal_key,
-      CompactionRangeDelAggregator* range_del_agg,
-      std::unique_ptr<CompactionProxy> compaction,
-      const CompactionFilter* compaction_filter = nullptr,
-      const std::atomic<bool>* shutting_down = nullptr,
-      const SequenceNumber preserve_deletes_seqnum = 0,
-      const std::atomic<bool>* manual_compaction_paused = nullptr,
-      const std::shared_ptr<Logger> info_log = nullptr);
+  CompactionIterator(InternalIterator* input, const Comparator* cmp,
+                     MergeHelper* merge_helper, SequenceNumber last_sequence,
+                     std::vector<SequenceNumber>* snapshots,
+                     SequenceNumber earliest_write_conflict_snapshot,
+                     const SnapshotChecker* snapshot_checker, Env* env,
+                     bool report_detailed_time, bool expect_valid_internal_key,
+                     CompactionRangeDelAggregator* range_del_agg,
+                     BlobFileBuilder* blob_file_builder,
+                     bool allow_data_in_errors,
+                     std::unique_ptr<CompactionProxy> compaction,
+                     const CompactionFilter* compaction_filter = nullptr,
+                     const std::atomic<bool>* shutting_down = nullptr,
+                     const SequenceNumber preserve_deletes_seqnum = 0,
+                     const std::atomic<int>* manual_compaction_paused = nullptr,
+                     const std::shared_ptr<Logger> info_log = nullptr);
 
   ~CompactionIterator();
 
@@ -121,7 +127,8 @@ class CompactionIterator {
   void PrepareOutput();
 
   // Invoke compaction filter if needed.
-  void InvokeFilterIfNeeded(bool* need_skip, Slice* skip_until);
+  // Return true on success, false on failures (e.g.: kIOError).
+  bool InvokeFilterIfNeeded(bool* need_skip, Slice* skip_until);
 
   // Given a sequence number, return the sequence number of the
   // earliest snapshot that this sequence number is visible in.
@@ -162,10 +169,11 @@ class CompactionIterator {
   bool report_detailed_time_;
   bool expect_valid_internal_key_;
   CompactionRangeDelAggregator* range_del_agg_;
+  BlobFileBuilder* blob_file_builder_;
   std::unique_ptr<CompactionProxy> compaction_;
   const CompactionFilter* compaction_filter_;
   const std::atomic<bool>* shutting_down_;
-  const std::atomic<bool>* manual_compaction_paused_;
+  const std::atomic<int>* manual_compaction_paused_;
   const SequenceNumber preserve_deletes_seqnum_;
   bool bottommost_level_;
   bool valid_ = false;
@@ -210,6 +218,7 @@ class CompactionIterator {
   // PinnedIteratorsManager used to pin input_ Iterator blocks while reading
   // merge operands and then releasing them after consuming them.
   PinnedIteratorsManager pinned_iters_mgr_;
+  std::string blob_index_;
   std::string compaction_filter_value_;
   InternalKey compaction_filter_skip_until_;
   // "level_ptrs" holds indices that remember which file of an associated
@@ -226,6 +235,8 @@ class CompactionIterator {
   bool current_key_committed_;
   std::shared_ptr<Logger> info_log_;
 
+  bool allow_data_in_errors_;
+
   bool IsShuttingDown() {
     // This is a best-effort facility, so memory_order_relaxed is sufficient.
     return shutting_down_ && shutting_down_->load(std::memory_order_relaxed);
@@ -234,7 +245,7 @@ class CompactionIterator {
   bool IsPausingManualCompaction() {
     // This is a best-effort facility, so memory_order_relaxed is sufficient.
     return manual_compaction_paused_ &&
-           manual_compaction_paused_->load(std::memory_order_relaxed);
+           manual_compaction_paused_->load(std::memory_order_relaxed) > 0;
   }
 };
 }  // namespace ROCKSDB_NAMESPACE

db/compaction/compaction_iterator_test.cc

@@ -169,13 +169,15 @@ class FakeCompaction : public CompactionIterator::CompactionProxy {
   Slice GetLargestUserKey() const override {
     return "\xff\xff\xff\xff\xff\xff\xff\xff\xff";
   }
-  bool allow_ingest_behind() const override { return false; }
+  bool allow_ingest_behind() const override { return is_allow_ingest_behind; }
 
   bool preserve_deletes() const override { return false; }
 
   bool key_not_exists_beyond_output_level = false;
 
   bool is_bottommost_level = false;
+
+  bool is_allow_ingest_behind = false;
 };
 
 // A simplifed snapshot checker which assumes each snapshot has a global
@@ -237,6 +239,7 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
     if (filter || bottommost_level) {
       compaction_proxy_ = new FakeCompaction();
       compaction_proxy_->is_bottommost_level = bottommost_level;
+      compaction_proxy_->is_allow_ingest_behind = AllowIngestBehind();
       compaction.reset(compaction_proxy_);
     }
     bool use_snapshot_checker = UseSnapshotChecker() || GetParam();
@@ -255,7 +258,9 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
         iter_.get(), cmp_, merge_helper_.get(), last_sequence, &snapshots_,
         earliest_write_conflict_snapshot, snapshot_checker_.get(),
         Env::Default(), false /* report_detailed_time */, false,
-        range_del_agg_.get(), std::move(compaction), filter, &shutting_down_));
+        range_del_agg_.get(), nullptr /* blob_file_builder */,
+        false /*allow_data_in_errors*/, std::move(compaction), filter,
+        &shutting_down_));
   }
 
   void AddSnapshot(SequenceNumber snapshot,
@@ -266,6 +271,8 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
 
   virtual bool UseSnapshotChecker() const { return false; }
 
+  virtual bool AllowIngestBehind() const { return false; }
+
   void RunTest(
       const std::vector<std::string>& input_keys,
       const std::vector<std::string>& input_values,
@@ -288,6 +295,7 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
       ASSERT_EQ(expected_values[i], c_iter_->value().ToString()) << info;
       c_iter_->Next();
     }
+    ASSERT_OK(c_iter_->status());
     ASSERT_FALSE(c_iter_->Valid());
   }
 
@@ -312,6 +320,7 @@ TEST_P(CompactionIteratorTest, EmptyResult) {
                  test::KeyStr("a", 3, kTypeValue)},
                 {"", "val"}, {}, {}, 5);
   c_iter_->SeekToFirst();
+  ASSERT_OK(c_iter_->status());
   ASSERT_FALSE(c_iter_->Valid());
 }
 
@@ -333,6 +342,7 @@ TEST_P(CompactionIteratorTest, CorruptionAfterSingleDeletion) {
   ASSERT_TRUE(c_iter_->Valid());
   ASSERT_EQ(test::KeyStr("b", 10, kTypeValue), c_iter_->key().ToString());
   c_iter_->Next();
+  ASSERT_OK(c_iter_->status());
   ASSERT_FALSE(c_iter_->Valid());
 }
 
@@ -349,6 +359,7 @@ TEST_P(CompactionIteratorTest, SimpleRangeDeletion) {
   ASSERT_TRUE(c_iter_->Valid());
   ASSERT_EQ(test::KeyStr("night", 3, kTypeValue), c_iter_->key().ToString());
   c_iter_->Next();
+  ASSERT_OK(c_iter_->status());
   ASSERT_FALSE(c_iter_->Valid());
 }
 
@@ -370,6 +381,7 @@ TEST_P(CompactionIteratorTest, RangeDeletionWithSnapshots) {
   ASSERT_TRUE(c_iter_->Valid());
   ASSERT_EQ(test::KeyStr("night", 40, kTypeValue), c_iter_->key().ToString());
   c_iter_->Next();
+  ASSERT_OK(c_iter_->status());
   ASSERT_FALSE(c_iter_->Valid());
 }
 
@@ -463,6 +475,7 @@ TEST_P(CompactionIteratorTest, CompactionFilterSkipUntil) {
   ASSERT_EQ(test::KeyStr("h", 91, kTypeValue), c_iter_->key().ToString());
   ASSERT_EQ("hv91", c_iter_->value().ToString());
   c_iter_->Next();
+  ASSERT_OK(c_iter_->status());
   ASSERT_FALSE(c_iter_->Valid());
 
   // Check that the compaction iterator did the correct sequence of calls on
@@ -656,6 +669,7 @@ TEST_P(CompactionIteratorTest, SingleMergeOperand) {
   ASSERT_TRUE(c_iter_->Valid());
   ASSERT_EQ("bv1bv2", c_iter_->value().ToString());
   c_iter_->Next();
+  ASSERT_OK(c_iter_->status());
   ASSERT_EQ("cv1cv2", c_iter_->value().ToString());
 }
 
@@ -697,6 +711,18 @@ TEST_P(CompactionIteratorTest, RemoveSingleDeletionAtBottomLevel) {
           nullptr /*compaction_filter*/, true /*bottommost_level*/);
 }
 
+TEST_P(CompactionIteratorTest, ConvertToPutAtBottom) {
+  std::shared_ptr<MergeOperator> merge_op =
+      MergeOperators::CreateStringAppendOperator();
+  RunTest({test::KeyStr("a", 4, kTypeMerge), test::KeyStr("a", 3, kTypeMerge),
+           test::KeyStr("a", 2, kTypeMerge), test::KeyStr("b", 1, kTypeValue)},
+          {"a4", "a3", "a2", "b1"},
+          {test::KeyStr("a", 0, kTypeValue), test::KeyStr("b", 0, kTypeValue)},
+          {"a2,a3,a4", "b1"}, kMaxSequenceNumber /*last_committed_seq*/,
+          merge_op.get(), nullptr /*compaction_filter*/,
+          true /*bottomost_level*/);
+}
+
 INSTANTIATE_TEST_CASE_P(CompactionIteratorTestInstance, CompactionIteratorTest,
                         testing::Values(true, false));
 
@@ -968,6 +994,45 @@ TEST_F(CompactionIteratorWithSnapshotCheckerTest, CompactionFilter_FullMerge) {
       compaction_filter.get());
 }
 
+// Tests how CompactionIterator work together with AllowIngestBehind.
+class CompactionIteratorWithAllowIngestBehindTest
+    : public CompactionIteratorTest {
+ public:
+  bool AllowIngestBehind() const override { return true; }
+};
+
+// When allow_ingest_behind is set, compaction iterator is not targeting
+// the bottommost level since there is no guarantee there won't be further
+// data ingested under the compaction output in future.
+TEST_P(CompactionIteratorWithAllowIngestBehindTest, NoConvertToPutAtBottom) {
+  std::shared_ptr<MergeOperator> merge_op =
+      MergeOperators::CreateStringAppendOperator();
+  RunTest({test::KeyStr("a", 4, kTypeMerge), test::KeyStr("a", 3, kTypeMerge),
+           test::KeyStr("a", 2, kTypeMerge), test::KeyStr("b", 1, kTypeValue)},
+          {"a4", "a3", "a2", "b1"},
+          {test::KeyStr("a", 4, kTypeMerge), test::KeyStr("b", 1, kTypeValue)},
+          {"a2,a3,a4", "b1"}, kMaxSequenceNumber /*last_committed_seq*/,
+          merge_op.get(), nullptr /*compaction_filter*/,
+          true /*bottomost_level*/);
+}
+
+TEST_P(CompactionIteratorWithAllowIngestBehindTest,
+       MergeToPutIfEncounteredPutAtBottom) {
+  std::shared_ptr<MergeOperator> merge_op =
+      MergeOperators::CreateStringAppendOperator();
+  RunTest({test::KeyStr("a", 4, kTypeMerge), test::KeyStr("a", 3, kTypeMerge),
+           test::KeyStr("a", 2, kTypeValue), test::KeyStr("b", 1, kTypeValue)},
+          {"a4", "a3", "a2", "b1"},
+          {test::KeyStr("a", 4, kTypeValue), test::KeyStr("b", 1, kTypeValue)},
+          {"a2,a3,a4", "b1"}, kMaxSequenceNumber /*last_committed_seq*/,
+          merge_op.get(), nullptr /*compaction_filter*/,
+          true /*bottomost_level*/);
+}
+
+INSTANTIATE_TEST_CASE_P(CompactionIteratorWithAllowIngestBehindTestInstance,
+                        CompactionIteratorWithAllowIngestBehindTest,
+                        testing::Values(true, false));
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/compaction/compaction_job.cc

@@ -7,6 +7,8 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
+#include "db/compaction/compaction_job.h"
+
 #include <algorithm>
 #include <cinttypes>
 #include <functional>
@@ -19,7 +21,6 @@
 #include <vector>
 
 #include "db/builder.h"
-#include "db/compaction/compaction_job.h"
 #include "db/db_impl/db_impl.h"
 #include "db/db_iter.h"
 #include "db/dbformat.h"
@@ -31,6 +32,7 @@
 #include "db/memtable_list.h"
 #include "db/merge_context.h"
 #include "db/merge_helper.h"
+#include "db/output_validator.h"
 #include "db/range_del_aggregator.h"
 #include "db/version_set.h"
 #include "file/filename.h"
@@ -45,6 +47,7 @@
 #include "port/port.h"
 #include "rocksdb/db.h"
 #include "rocksdb/env.h"
+#include "rocksdb/sst_partitioner.h"
 #include "rocksdb/statistics.h"
 #include "rocksdb/status.h"
 #include "rocksdb/table.h"
@@ -54,6 +57,7 @@
 #include "table/table_builder.h"
 #include "test_util/sync_point.h"
 #include "util/coding.h"
+#include "util/hash.h"
 #include "util/mutexlock.h"
 #include "util/random.h"
 #include "util/stop_watch.h"
@@ -116,9 +120,18 @@ struct CompactionJob::SubcompactionState {
   // The return status of this subcompaction
   Status status;
 
+  // The return IO Status of this subcompaction
+  IOStatus io_status;
+
   // Files produced by this subcompaction
   struct Output {
+    Output(FileMetaData&& _meta, const InternalKeyComparator& _icmp,
+           bool _enable_order_check, bool _enable_hash)
+        : meta(std::move(_meta)),
+          validator(_icmp, _enable_order_check, _enable_hash),
+          finished(false) {}
     FileMetaData meta;
+    OutputValidator validator;
     bool finished;
     std::shared_ptr<const TableProperties> table_properties;
   };
@@ -127,9 +140,10 @@ struct CompactionJob::SubcompactionState {
   std::vector<Output> outputs;
   std::unique_ptr<WritableFileWriter> outfile;
   std::unique_ptr<TableBuilder> builder;
+
   Output* current_output() {
     if (outputs.empty()) {
-      // This subcompaction's outptut could be empty if compaction was aborted
+      // This subcompaction's output could be empty if compaction was aborted
       // before this subcompaction had a chance to generate any output files.
       // When subcompactions are executed sequentially this is more likely and
       // will be particulalry likely for the later subcompactions to be empty.
@@ -140,13 +154,13 @@ struct CompactionJob::SubcompactionState {
     }
   }
 
-  uint64_t current_output_file_size;
+  uint64_t current_output_file_size = 0;
 
   // State during the subcompaction
-  uint64_t total_bytes;
-  uint64_t num_output_records;
+  uint64_t total_bytes = 0;
+  uint64_t num_output_records = 0;
   CompactionJobStats compaction_job_stats;
-  uint64_t approx_size;
+  uint64_t approx_size = 0;
   // An index that used to speed up ShouldStopBefore().
   size_t grandparent_index = 0;
   // The number of bytes overlapping between the current output and
@@ -155,49 +169,25 @@ struct CompactionJob::SubcompactionState {
   // A flag determine whether the key has been seen in ShouldStopBefore()
   bool seen_key = false;
 
-  SubcompactionState(Compaction* c, Slice* _start, Slice* _end,
-                     uint64_t size = 0)
-      : compaction(c),
-        start(_start),
-        end(_end),
-        outfile(nullptr),
-        builder(nullptr),
-        current_output_file_size(0),
-        total_bytes(0),
-        num_output_records(0),
-        approx_size(size),
-        grandparent_index(0),
-        overlapped_bytes(0),
-        seen_key(false) {
+  SubcompactionState(Compaction* c, Slice* _start, Slice* _end, uint64_t size)
+      : compaction(c), start(_start), end(_end), approx_size(size) {
     assert(compaction != nullptr);
   }
 
-  SubcompactionState(SubcompactionState&& o) { *this = std::move(o); }
-
-  SubcompactionState& operator=(SubcompactionState&& o) {
-    compaction = std::move(o.compaction);
-    start = std::move(o.start);
-    end = std::move(o.end);
-    status = std::move(o.status);
-    outputs = std::move(o.outputs);
-    outfile = std::move(o.outfile);
-    builder = std::move(o.builder);
-    current_output_file_size = std::move(o.current_output_file_size);
-    total_bytes = std::move(o.total_bytes);
-    num_output_records = std::move(o.num_output_records);
-    compaction_job_stats = std::move(o.compaction_job_stats);
-    approx_size = std::move(o.approx_size);
-    grandparent_index = std::move(o.grandparent_index);
-    overlapped_bytes = std::move(o.overlapped_bytes);
-    seen_key = std::move(o.seen_key);
-    return *this;
+  // Adds the key and value to the builder
+  // If paranoid is true, adds the key-value to the paranoid hash
+  Status AddToBuilder(const Slice& key, const Slice& value) {
+    auto curr = current_output();
+    assert(builder != nullptr);
+    assert(curr != nullptr);
+    Status s = curr->validator.Add(key, value);
+    if (!s.ok()) {
+      return s;
+    }
+    builder->Add(key, value);
+    return Status::OK();
   }
 
-  // Because member std::unique_ptrs do not have these.
-  SubcompactionState(const SubcompactionState&) = delete;
-
-  SubcompactionState& operator=(const SubcompactionState&) = delete;
-
   // Returns true iff we should stop building the current output
   // before processing "internal_key".
   bool ShouldStopBefore(const Slice& internal_key, uint64_t curr_file_size) {
@@ -286,10 +276,8 @@ void CompactionJob::AggregateStatistics() {
     compact_->total_bytes += sc.total_bytes;
     compact_->num_output_records += sc.num_output_records;
   }
-  if (compaction_job_stats_) {
-    for (SubcompactionState& sc : compact_->sub_compact_states) {
-      compaction_job_stats_->Add(sc.compaction_job_stats);
-    }
+  for (SubcompactionState& sc : compact_->sub_compact_states) {
+    compaction_job_stats_->Add(sc.compaction_job_stats);
   }
 }
 
@@ -305,8 +293,9 @@ CompactionJob::CompactionJob(
     const SnapshotChecker* snapshot_checker, std::shared_ptr<Cache> table_cache,
     EventLogger* event_logger, bool paranoid_file_checks, bool measure_io_stats,
     const std::string& dbname, CompactionJobStats* compaction_job_stats,
-    Env::Priority thread_pri, const std::atomic<bool>* manual_compaction_paused,
-    const std::string& db_id, const std::string& db_session_id)
+    Env::Priority thread_pri, const std::shared_ptr<IOTracer>& io_tracer,
+    const std::atomic<int>* manual_compaction_paused, const std::string& db_id,
+    const std::string& db_session_id)
     : job_id_(job_id),
       compact_(new CompactionState(compaction)),
       compaction_job_stats_(compaction_job_stats),
@@ -317,7 +306,8 @@ CompactionJob::CompactionJob(
       db_options_(db_options),
       file_options_(file_options),
       env_(db_options.env),
-      fs_(db_options.fs.get()),
+      io_tracer_(io_tracer),
+      fs_(db_options.fs, io_tracer),
       file_options_for_read_(
           fs_->OptimizeForCompactionTableRead(file_options, db_options_)),
       versions_(versions),
@@ -340,6 +330,7 @@ CompactionJob::CompactionJob(
       measure_io_stats_(measure_io_stats),
       write_hint_(Env::WLTH_NOT_SET),
       thread_pri_(thread_pri) {
+  assert(compaction_job_stats_ != nullptr);
   assert(log_buffer_ != nullptr);
   const auto* cfd = compact_->compaction->column_family_data();
   ThreadStatusUtil::SetColumnFamily(cfd, cfd->ioptions()->env,
@@ -391,10 +382,9 @@ void CompactionJob::ReportStartedCompaction(Compaction* compaction) {
   // to ensure GetThreadList() can always show them all together.
   ThreadStatusUtil::SetThreadOperation(ThreadStatus::OP_COMPACTION);
 
-  if (compaction_job_stats_) {
-    compaction_job_stats_->is_manual_compaction =
-        compaction->is_manual_compaction();
-  }
+  compaction_job_stats_->is_manual_compaction =
+      compaction->is_manual_compaction();
+  compaction_job_stats_->is_full_compaction = compaction->is_full_compaction();
 }
 
 void CompactionJob::Prepare() {
@@ -426,7 +416,11 @@ void CompactionJob::Prepare() {
     RecordInHistogram(stats_, NUM_SUBCOMPACTIONS_SCHEDULED,
                       compact_->sub_compact_states.size());
   } else {
-    compact_->sub_compact_states.emplace_back(c, nullptr, nullptr);
+    constexpr Slice* start = nullptr;
+    constexpr Slice* end = nullptr;
+    constexpr uint64_t size = 0;
+
+    compact_->sub_compact_states.emplace_back(c, start, end, size);
   }
 }
 
@@ -609,38 +603,42 @@ Status CompactionJob::Run() {
 
   // Check if any thread encountered an error during execution
   Status status;
+  IOStatus io_s;
   for (const auto& state : compact_->sub_compact_states) {
     if (!state.status.ok()) {
       status = state.status;
+      io_s = state.io_status;
       break;
     }
   }
-
-  IOStatus io_s;
+  if (io_status_.ok()) {
+    io_status_ = io_s;
+  }
   if (status.ok() && output_directory_) {
     io_s = output_directory_->Fsync(IOOptions(), nullptr);
   }
-  if (!io_s.ok()) {
+  if (io_status_.ok()) {
     io_status_ = io_s;
+  }
+  if (status.ok()) {
     status = io_s;
   }
-
   if (status.ok()) {
     thread_pool.clear();
-    std::vector<const FileMetaData*> files_meta;
+    std::vector<const CompactionJob::SubcompactionState::Output*> files_output;
     for (const auto& state : compact_->sub_compact_states) {
       for (const auto& output : state.outputs) {
-        files_meta.emplace_back(&output.meta);
+        files_output.emplace_back(&output);
       }
     }
     ColumnFamilyData* cfd = compact_->compaction->column_family_data();
     auto prefix_extractor =
         compact_->compaction->mutable_cf_options()->prefix_extractor.get();
-    std::atomic<size_t> next_file_meta_idx(0);
+    std::atomic<size_t> next_file_idx(0);
     auto verify_table = [&](Status& output_status) {
       while (true) {
-        size_t file_idx = next_file_meta_idx.fetch_add(1);
-        if (file_idx >= files_meta.size()) {
+        size_t file_idx = next_file_idx.fetch_add(1);
+        if (file_idx >= files_output.size()) {
           break;
         }
         // Verify that the table is usable
@@ -649,9 +647,11 @@ Status CompactionJob::Run() {
         // No matter whether use_direct_io_for_flush_and_compaction is true,
         // we will regard this verification as user reads since the goal is
         // to cache it here for further user reads
+        ReadOptions read_options;
         InternalIterator* iter = cfd->table_cache()->NewIterator(
-            ReadOptions(), file_options_, cfd->internal_comparator(),
-            *files_meta[file_idx], /*range_del_agg=*/nullptr, prefix_extractor,
+            read_options, file_options_, cfd->internal_comparator(),
+            files_output[file_idx]->meta, /*range_del_agg=*/nullptr,
+            prefix_extractor,
             /*table_reader_ptr=*/nullptr,
             cfd->internal_stats()->GetFileReadHist(
                 compact_->compaction->output_level()),
@@ -665,8 +665,22 @@ Status CompactionJob::Run() {
         auto s = iter->status();
 
         if (s.ok() && paranoid_file_checks_) {
-          for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {}
-          s = iter->status();
+          OutputValidator validator(cfd->internal_comparator(),
+                                    /*_enable_order_check=*/true,
+                                    /*_enable_hash=*/true);
+          for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+            s = validator.Add(iter->key(), iter->value());
+            if (!s.ok()) {
+              break;
+            }
+          }
+          if (s.ok()) {
+            s = iter->status();
+          }
+          if (s.ok() &&
+              !validator.CompareValidator(files_output[file_idx]->validator)) {
+            s = Status::Corruption("Paranoid checksums do not match");
+          }
         }
 
         delete iter;
@@ -724,7 +738,6 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
   cfd->internal_stats()->AddCompactionStats(
       compact_->compaction->output_level(), thread_pri_, compaction_stats_);
 
-  versions_->SetIOStatusOK();
   if (status.ok()) {
     status = InstallCompactionResults(mutable_cf_options);
   }
@@ -789,14 +802,12 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
          << compact_->sub_compact_states.size() << "output_compression"
          << CompressionTypeToString(compact_->compaction->output_compression());
 
-  if (compaction_job_stats_ != nullptr) {
-    stream << "num_single_delete_mismatches"
-           << compaction_job_stats_->num_single_del_mismatch;
-    stream << "num_single_delete_fallthrough"
-           << compaction_job_stats_->num_single_del_fallthru;
-  }
+  stream << "num_single_delete_mismatches"
+         << compaction_job_stats_->num_single_del_mismatch;
+  stream << "num_single_delete_fallthrough"
+         << compaction_job_stats_->num_single_del_fallthru;
 
-  if (measure_io_stats_ && compaction_job_stats_ != nullptr) {
+  if (measure_io_stats_) {
     stream << "file_write_nanos" << compaction_job_stats_->file_write_nanos;
     stream << "file_range_sync_nanos"
            << compaction_job_stats_->file_range_sync_nanos;
@@ -842,11 +853,20 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
 
   CompactionRangeDelAggregator range_del_agg(&cfd->internal_comparator(),
                                              existing_snapshots_);
+  ReadOptions read_options;
+  read_options.verify_checksums = true;
+  read_options.fill_cache = false;
+  // Compaction iterators shouldn't be confined to a single prefix.
+  // Compactions use Seek() for
+  // (a) concurrent compactions,
+  // (b) CompactionFilter::Decision::kRemoveAndSkipUntil.
+  read_options.total_order_seek = true;
 
   // Although the v2 aggregator is what the level iterator(s) know about,
   // the AddTombstones calls will be propagated down to the v1 aggregator.
-  std::unique_ptr<InternalIterator> input(versions_->MakeInputIterator(
-      sub_compact->compaction, &range_del_agg, file_options_for_read_));
+  std::unique_ptr<InternalIterator> input(
+      versions_->MakeInputIterator(read_options, sub_compact->compaction,
+                                   &range_del_agg, file_options_for_read_));
 
   AutoThreadOperationStageUpdater stage_updater(
       ThreadStatus::STAGE_COMPACTION_PROCESS_KV);
@@ -883,7 +903,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   TEST_SYNC_POINT_CALLBACK(
       "CompactionJob::Run():PausingManualCompaction:1",
       reinterpret_cast<void*>(
-          const_cast<std::atomic<bool>*>(manual_compaction_paused_)));
+          const_cast<std::atomic<int>*>(manual_compaction_paused_)));
 
   Slice* start = sub_compact->start;
   Slice* end = sub_compact->end;
@@ -899,9 +919,11 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   sub_compact->c_iter.reset(new CompactionIterator(
       input.get(), cfd->user_comparator(), &merge, versions_->LastSequence(),
       &existing_snapshots_, earliest_write_conflict_snapshot_,
-      snapshot_checker_, env_, ShouldReportDetailedTime(env_, stats_), false,
-      &range_del_agg, sub_compact->compaction, compaction_filter,
-      shutting_down_, preserve_deletes_seqnum_, manual_compaction_paused_,
+      snapshot_checker_, env_, ShouldReportDetailedTime(env_, stats_),
+      /*expect_valid_internal_key=*/true, &range_del_agg,
+      /* blob_file_builder */ nullptr, db_options_.allow_data_in_errors,
+      sub_compact->compaction, compaction_filter, shutting_down_,
+      preserve_deletes_seqnum_, manual_compaction_paused_,
       db_options_.info_log));
   auto c_iter = sub_compact->c_iter.get();
   c_iter->SeekToFirst();
@@ -914,6 +936,12 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   }
   const auto& c_iter_stats = c_iter->iter_stats();
 
+  std::unique_ptr<SstPartitioner> partitioner =
+      sub_compact->compaction->output_level() == 0
+          ? nullptr
+          : sub_compact->compaction->CreateSstPartitioner();
+  std::string last_key_for_partitioner;
+
   while (status.ok() && !cfd->IsDropped() && c_iter->Valid()) {
     // Invariant: c_iter.status() is guaranteed to be OK if c_iter->Valid()
     // returns true.
@@ -940,9 +968,11 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
         break;
       }
     }
-    assert(sub_compact->builder != nullptr);
-    assert(sub_compact->current_output() != nullptr);
-    sub_compact->builder->Add(key, value);
+    status = sub_compact->AddToBuilder(key, value);
+    if (!status.ok()) {
+      break;
+    }
+
     sub_compact->current_output_file_size =
         sub_compact->builder->EstimatedFileSize();
     const ParsedInternalKey& ikey = c_iter->ikey();
@@ -959,33 +989,39 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
     // going to be 1.2MB and max_output_file_size = 1MB, prefer to have 0.6MB
     // and 0.6MB instead of 1MB and 0.2MB)
     bool output_file_ended = false;
-    Status input_status;
     if (sub_compact->compaction->output_level() != 0 &&
         sub_compact->current_output_file_size >=
             sub_compact->compaction->max_output_file_size()) {
       // (1) this key terminates the file. For historical reasons, the iterator
       // status before advancing will be given to FinishCompactionOutputFile().
-      input_status = input->status();
       output_file_ended = true;
     }
     TEST_SYNC_POINT_CALLBACK(
         "CompactionJob::Run():PausingManualCompaction:2",
         reinterpret_cast<void*>(
-            const_cast<std::atomic<bool>*>(manual_compaction_paused_)));
+            const_cast<std::atomic<int>*>(manual_compaction_paused_)));
+    if (partitioner.get()) {
+      last_key_for_partitioner.assign(c_iter->user_key().data_,
+                                      c_iter->user_key().size_);
+    }
     c_iter->Next();
     if (c_iter->status().IsManualCompactionPaused()) {
       break;
     }
-    if (!output_file_ended && c_iter->Valid() &&
-        sub_compact->compaction->output_level() != 0 &&
-        sub_compact->ShouldStopBefore(c_iter->key(),
-                                      sub_compact->current_output_file_size) &&
-        sub_compact->builder != nullptr) {
-      // (2) this key belongs to the next file. For historical reasons, the
-      // iterator status after advancing will be given to
-      // FinishCompactionOutputFile().
-      input_status = input->status();
-      output_file_ended = true;
+    if (!output_file_ended && c_iter->Valid()) {
+      if (((partitioner.get() &&
+            partitioner->ShouldPartition(PartitionerRequest(
+                last_key_for_partitioner, c_iter->user_key(),
+                sub_compact->current_output_file_size)) == kRequired) ||
+           (sub_compact->compaction->output_level() != 0 &&
+            sub_compact->ShouldStopBefore(
+                c_iter->key(), sub_compact->current_output_file_size))) &&
+          sub_compact->builder != nullptr) {
+        // (2) this key belongs to the next file. For historical reasons, the
+        // iterator status after advancing will be given to
+        // FinishCompactionOutputFile().
+        output_file_ended = true;
+      }
     }
     if (output_file_ended) {
       const Slice* next_key = nullptr;
@@ -993,9 +1029,9 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
         next_key = &c_iter->key();
       }
       CompactionIterationStats range_del_out_stats;
-      status =
-          FinishCompactionOutputFile(input_status, sub_compact, &range_del_agg,
-                                     &range_del_out_stats, next_key);
+      status = FinishCompactionOutputFile(input->status(), sub_compact,
+                                          &range_del_agg, &range_del_out_stats,
+                                          next_key);
       RecordDroppedKeys(range_del_out_stats,
                         &sub_compact->compaction_job_stats);
     }
@@ -1029,7 +1065,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   }
   if ((status.ok() || status.IsColumnFamilyDropped()) &&
       (manual_compaction_paused_ &&
-       manual_compaction_paused_->load(std::memory_order_relaxed))) {
+       manual_compaction_paused_->load(std::memory_order_relaxed) > 0)) {
     status = Status::Incomplete(Status::SubCode::kManualCompactionPaused);
   }
   if (status.ok()) {
@@ -1051,7 +1087,7 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
     CompactionIterationStats range_del_out_stats;
     Status s = FinishCompactionOutputFile(status, sub_compact, &range_del_agg,
                                           &range_del_out_stats);
-    if (status.ok()) {
+    if (!s.ok() && status.ok()) {
       status = s;
     }
     RecordDroppedKeys(range_del_out_stats, &sub_compact->compaction_job_stats);
@@ -1077,6 +1113,16 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
       SetPerfLevel(prev_perf_level);
     }
   }
+#ifdef ROCKSDB_ASSERT_STATUS_CHECKED
+  if (!status.ok()) {
+    if (sub_compact->c_iter) {
+      sub_compact->c_iter->status().PermitUncheckedError();
+    }
+    if (input) {
+      input->status().PermitUncheckedError();
+    }
+  }
+#endif  // ROCKSDB_ASSERT_STATUS_CHECKED
 
   sub_compact->c_iter.reset();
   input.reset();
@@ -1137,6 +1183,8 @@ Status CompactionJob::FinishCompactionOutputFile(
 
   ColumnFamilyData* cfd = sub_compact->compaction->column_family_data();
   const Comparator* ucmp = cfd->user_comparator();
+  std::string file_checksum = kUnknownFileChecksum;
+  std::string file_checksum_func_name = kUnknownFileChecksumFuncName;
 
   // Check for iterator errors
   Status s = input_status;
@@ -1210,6 +1258,7 @@ Status CompactionJob::FinishCompactionOutputFile(
     } else {
       it->SeekToFirst();
     }
+    TEST_SYNC_POINT("CompactionJob::FinishCompactionOutputFile1");
     for (; it->Valid(); it->Next()) {
       auto tombstone = it->Tombstone();
       if (upper_bound != nullptr) {
@@ -1237,6 +1286,7 @@ Status CompactionJob::FinishCompactionOutputFile(
       auto kv = tombstone.Serialize();
       assert(lower_bound == nullptr ||
              ucmp->Compare(*lower_bound, kv.second) < 0);
+      // Range tombstone is not supported by output validator yet.
       sub_compact->builder->Add(kv.first.Encode(), kv.second);
       InternalKey smallest_candidate = std::move(kv.first);
       if (lower_bound != nullptr &&
@@ -1310,9 +1360,9 @@ Status CompactionJob::FinishCompactionOutputFile(
   } else {
     sub_compact->builder->Abandon();
   }
-  if (!sub_compact->builder->io_status().ok()) {
-    io_status_ = sub_compact->builder->io_status();
-    s = io_status_;
+  IOStatus io_s = sub_compact->builder->io_status();
+  if (s.ok()) {
+    s = io_s;
   }
   const uint64_t current_bytes = sub_compact->builder->FileSize();
   if (s.ok()) {
@@ -1322,24 +1372,30 @@ Status CompactionJob::FinishCompactionOutputFile(
   sub_compact->total_bytes += current_bytes;
 
   // Finish and check for file errors
-  IOStatus io_s;
   if (s.ok()) {
     StopWatch sw(env_, stats_, COMPACTION_OUTFILE_SYNC_MICROS);
     io_s = sub_compact->outfile->Sync(db_options_.use_fsync);
   }
-  if (io_s.ok()) {
+  if (s.ok() && io_s.ok()) {
     io_s = sub_compact->outfile->Close();
   }
-  if (io_s.ok()) {
+  if (s.ok() && io_s.ok()) {
     // Add the checksum information to file metadata.
     meta->file_checksum = sub_compact->outfile->GetFileChecksum();
     meta->file_checksum_func_name =
         sub_compact->outfile->GetFileChecksumFuncName();
+    file_checksum = meta->file_checksum;
+    file_checksum_func_name = meta->file_checksum_func_name;
   }
-  if (!io_s.ok()) {
-    io_status_ = io_s;
+  if (s.ok()) {
     s = io_s;
   }
+  if (sub_compact->io_status.ok()) {
+    sub_compact->io_status = io_s;
+    // Since this error is really a copy of the
+    // "normal" status, it does not also need to be checked
+    sub_compact->io_status.PermitUncheckedError();
+  }
   sub_compact->outfile.reset();
 
   TableProperties tp;
@@ -1389,14 +1445,18 @@ Status CompactionJob::FinishCompactionOutputFile(
   EventHelpers::LogAndNotifyTableFileCreationFinished(
       event_logger_, cfd->ioptions()->listeners, dbname_, cfd->GetName(), fname,
       job_id_, output_fd, oldest_blob_file_number, tp,
-      TableFileCreationReason::kCompaction, s);
+      TableFileCreationReason::kCompaction, s, file_checksum,
+      file_checksum_func_name);
 
 #ifndef ROCKSDB_LITE
   // Report new file to SstFileManagerImpl
   auto sfm =
       static_cast<SstFileManagerImpl*>(db_options_.sst_file_manager.get());
   if (sfm && meta != nullptr && meta->fd.GetPathId() == 0) {
-    sfm->OnAddFile(fname);
+    Status add_s = sfm->OnAddFile(fname);
+    if (!add_s.ok() && s.ok()) {
+      s = add_s;
+    }
     if (sfm->IsMaxAllowedSpaceReached()) {
       // TODO(ajkr): should we return OK() if max space was reached by the final
       // compaction output file (similarly to how flush works when full)?
@@ -1405,7 +1465,9 @@ Status CompactionJob::FinishCompactionOutputFile(
           "CompactionJob::FinishCompactionOutputFile:"
           "MaxAllowedSpaceReached");
       InstrumentedMutexLock l(db_mutex_);
-      db_error_handler_->SetBGError(s, BackgroundErrorReason::kCompaction);
+      // Should handle return error?
+      db_error_handler_->SetBGError(s, BackgroundErrorReason::kCompaction)
+          .PermitUncheckedError();
     }
   }
 #endif
@@ -1456,10 +1518,22 @@ Status CompactionJob::InstallCompactionResults(
 
 void CompactionJob::RecordCompactionIOStats() {
   RecordTick(stats_, COMPACT_READ_BYTES, IOSTATS(bytes_read));
+  RecordTick(stats_, COMPACT_WRITE_BYTES, IOSTATS(bytes_written));
+  CompactionReason compaction_reason =
+      compact_->compaction->compaction_reason();
+  if (compaction_reason == CompactionReason::kFilesMarkedForCompaction) {
+    RecordTick(stats_, COMPACT_READ_BYTES_MARKED, IOSTATS(bytes_read));
+    RecordTick(stats_, COMPACT_WRITE_BYTES_MARKED, IOSTATS(bytes_written));
+  } else if (compaction_reason == CompactionReason::kPeriodicCompaction) {
+    RecordTick(stats_, COMPACT_READ_BYTES_PERIODIC, IOSTATS(bytes_read));
+    RecordTick(stats_, COMPACT_WRITE_BYTES_PERIODIC, IOSTATS(bytes_written));
+  } else if (compaction_reason == CompactionReason::kTtl) {
+    RecordTick(stats_, COMPACT_READ_BYTES_TTL, IOSTATS(bytes_read));
+    RecordTick(stats_, COMPACT_WRITE_BYTES_TTL, IOSTATS(bytes_written));
+  }
   ThreadStatusUtil::IncreaseThreadOperationProperty(
       ThreadStatus::COMPACTION_BYTES_READ, IOSTATS(bytes_read));
   IOSTATS_RESET(bytes_read);
-  RecordTick(stats_, COMPACT_WRITE_BYTES, IOSTATS(bytes_written));
   ThreadStatusUtil::IncreaseThreadOperationProperty(
       ThreadStatus::COMPACTION_BYTES_WRITTEN, IOSTATS(bytes_written));
   IOSTATS_RESET(bytes_written);
@@ -1488,7 +1562,16 @@ Status CompactionJob::OpenCompactionOutputFile(
   TEST_SYNC_POINT_CALLBACK("CompactionJob::OpenCompactionOutputFile",
                            &syncpoint_arg);
 #endif
-  Status s = NewWritableFile(fs_, fname, &writable_file, file_options_);
+  Status s;
+  IOStatus io_s =
+      NewWritableFile(fs_.get(), fname, &writable_file, file_options_);
+  s = io_s;
+  if (sub_compact->io_status.ok()) {
+    sub_compact->io_status = io_s;
+    // Since this error is really a copy of the io_s that is checked below as s,
+    // it does not also need to be checked.
+    sub_compact->io_status.PermitUncheckedError();
+  }
   if (!s.ok()) {
     ROCKS_LOG_ERROR(
         db_options_.info_log,
@@ -1500,7 +1583,8 @@ Status CompactionJob::OpenCompactionOutputFile(
     EventHelpers::LogAndNotifyTableFileCreationFinished(
         event_logger_, cfd->ioptions()->listeners, dbname_, cfd->GetName(),
         fname, job_id_, FileDescriptor(), kInvalidBlobFileNumber,
-        TableProperties(), TableFileCreationReason::kCompaction, s);
+        TableProperties(), TableFileCreationReason::kCompaction, s,
+        kUnknownFileChecksum, kUnknownFileChecksumFuncName);
     return s;
   }
 
@@ -1522,13 +1606,17 @@ Status CompactionJob::OpenCompactionOutputFile(
 
   // Initialize a SubcompactionState::Output and add it to sub_compact->outputs
   {
-    SubcompactionState::Output out;
-    out.meta.fd = FileDescriptor(file_number,
-                                 sub_compact->compaction->output_path_id(), 0);
-    out.meta.oldest_ancester_time = oldest_ancester_time;
-    out.meta.file_creation_time = current_time;
-    out.finished = false;
-    sub_compact->outputs.push_back(out);
+    FileMetaData meta;
+    meta.fd = FileDescriptor(file_number,
+                             sub_compact->compaction->output_path_id(), 0);
+    meta.oldest_ancester_time = oldest_ancester_time;
+    meta.file_creation_time = current_time;
+    sub_compact->outputs.emplace_back(
+        std::move(meta), cfd->internal_comparator(),
+        /*enable_order_check=*/
+        sub_compact->compaction->mutable_cf_options()
+            ->check_flush_compaction_key_order,
+        /*enable_hash=*/paranoid_file_checks_);
   }
 
   writable_file->SetIOPriority(Env::IOPriority::IO_LOW);
@@ -1537,10 +1625,10 @@ Status CompactionJob::OpenCompactionOutputFile(
       sub_compact->compaction->OutputFilePreallocationSize()));
   const auto& listeners =
       sub_compact->compaction->immutable_cf_options()->listeners;
-  sub_compact->outfile.reset(
-      new WritableFileWriter(std::move(writable_file), fname, file_options_,
-                             env_, db_options_.statistics.get(), listeners,
-                             db_options_.file_checksum_gen_factory.get()));
+  sub_compact->outfile.reset(new WritableFileWriter(
+      std::move(writable_file), fname, file_options_, env_, io_tracer_,
+      db_options_.statistics.get(), listeners,
+      db_options_.file_checksum_gen_factory.get()));
 
   // If the Column family flag is to only optimize filters for hits,
   // we can skip creating filters if this is the bottommost_level where
@@ -1581,6 +1669,9 @@ void CompactionJob::CleanupCompaction() {
         TableCache::Evict(table_cache_.get(), out.meta.fd.GetNumber());
       }
     }
+    // TODO: sub_compact.io_status is not checked like status. Not sure if thats
+    // intentional. So ignoring the io_status as of now.
+    sub_compact.io_status.PermitUncheckedError();
   }
   delete compact_;
   compact_ = nullptr;
@@ -1657,32 +1748,29 @@ void CompactionJob::UpdateCompactionInputStatsHelper(int* num_files,
 void CompactionJob::UpdateCompactionJobStats(
     const InternalStats::CompactionStats& stats) const {
 #ifndef ROCKSDB_LITE
-  if (compaction_job_stats_) {
-    compaction_job_stats_->elapsed_micros = stats.micros;
+  compaction_job_stats_->elapsed_micros = stats.micros;
 
-    // input information
-    compaction_job_stats_->total_input_bytes =
-        stats.bytes_read_non_output_levels + stats.bytes_read_output_level;
-    compaction_job_stats_->num_input_records = stats.num_input_records;
-    compaction_job_stats_->num_input_files =
-        stats.num_input_files_in_non_output_levels +
-        stats.num_input_files_in_output_level;
-    compaction_job_stats_->num_input_files_at_output_level =
-        stats.num_input_files_in_output_level;
+  // input information
+  compaction_job_stats_->total_input_bytes =
+      stats.bytes_read_non_output_levels + stats.bytes_read_output_level;
+  compaction_job_stats_->num_input_records = stats.num_input_records;
+  compaction_job_stats_->num_input_files =
+      stats.num_input_files_in_non_output_levels +
+      stats.num_input_files_in_output_level;
+  compaction_job_stats_->num_input_files_at_output_level =
+      stats.num_input_files_in_output_level;
 
-    // output information
-    compaction_job_stats_->total_output_bytes = stats.bytes_written;
-    compaction_job_stats_->num_output_records = compact_->num_output_records;
-    compaction_job_stats_->num_output_files = stats.num_output_files;
+  // output information
+  compaction_job_stats_->total_output_bytes = stats.bytes_written;
+  compaction_job_stats_->num_output_records = compact_->num_output_records;
+  compaction_job_stats_->num_output_files = stats.num_output_files;
 
-    if (compact_->NumOutputFiles() > 0U) {
-      CopyPrefix(compact_->SmallestUserKey(),
-                 CompactionJobStats::kMaxPrefixLength,
-                 &compaction_job_stats_->smallest_output_key_prefix);
-      CopyPrefix(compact_->LargestUserKey(),
-                 CompactionJobStats::kMaxPrefixLength,
-                 &compaction_job_stats_->largest_output_key_prefix);
-    }
+  if (compact_->NumOutputFiles() > 0U) {
+    CopyPrefix(compact_->SmallestUserKey(),
+               CompactionJobStats::kMaxPrefixLength,
+               &compaction_job_stats_->smallest_output_key_prefix);
+    CopyPrefix(compact_->LargestUserKey(), CompactionJobStats::kMaxPrefixLength,
+               &compaction_job_stats_->largest_output_key_prefix);
   }
 #else
   (void)stats;

db/compaction/compaction_job.h

@@ -62,25 +62,23 @@ class VersionSet;
 // if needed.
 class CompactionJob {
  public:
-  CompactionJob(int job_id, Compaction* compaction,
-                const ImmutableDBOptions& db_options,
-                const FileOptions& file_options, VersionSet* versions,
-                const std::atomic<bool>* shutting_down,
-                const SequenceNumber preserve_deletes_seqnum,
-                LogBuffer* log_buffer, FSDirectory* db_directory,
-                FSDirectory* output_directory, Statistics* stats,
-                InstrumentedMutex* db_mutex, ErrorHandler* db_error_handler,
-                std::vector<SequenceNumber> existing_snapshots,
-                SequenceNumber earliest_write_conflict_snapshot,
-                const SnapshotChecker* snapshot_checker,
-                std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
-                bool paranoid_file_checks, bool measure_io_stats,
-                const std::string& dbname,
-                CompactionJobStats* compaction_job_stats,
-                Env::Priority thread_pri,
-                const std::atomic<bool>* manual_compaction_paused = nullptr,
-                const std::string& db_id = "",
-                const std::string& db_session_id = "");
+  CompactionJob(
+      int job_id, Compaction* compaction, const ImmutableDBOptions& db_options,
+      const FileOptions& file_options, VersionSet* versions,
+      const std::atomic<bool>* shutting_down,
+      const SequenceNumber preserve_deletes_seqnum, LogBuffer* log_buffer,
+      FSDirectory* db_directory, FSDirectory* output_directory,
+      Statistics* stats, InstrumentedMutex* db_mutex,
+      ErrorHandler* db_error_handler,
+      std::vector<SequenceNumber> existing_snapshots,
+      SequenceNumber earliest_write_conflict_snapshot,
+      const SnapshotChecker* snapshot_checker,
+      std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
+      bool paranoid_file_checks, bool measure_io_stats,
+      const std::string& dbname, CompactionJobStats* compaction_job_stats,
+      Env::Priority thread_pri, const std::shared_ptr<IOTracer>& io_tracer,
+      const std::atomic<int>* manual_compaction_paused = nullptr,
+      const std::string& db_id = "", const std::string& db_session_id = "");
 
   ~CompactionJob();
 
@@ -160,12 +158,13 @@ class CompactionJob {
   const FileOptions file_options_;
 
   Env* env_;
-  FileSystem* fs_;
+  std::shared_ptr<IOTracer> io_tracer_;
+  FileSystemPtr fs_;
   // env_option optimized for compaction table reads
   FileOptions file_options_for_read_;
   VersionSet* versions_;
   const std::atomic<bool>* shutting_down_;
-  const std::atomic<bool>* manual_compaction_paused_;
+  const std::atomic<int>* manual_compaction_paused_;
   const SequenceNumber preserve_deletes_seqnum_;
   LogBuffer* log_buffer_;
   FSDirectory* db_directory_;

db/compaction/compaction_job_stats_test.cc

@@ -52,6 +52,7 @@
 #include "test_util/sync_point.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"
+#include "util/cast_util.h"
 #include "util/compression.h"
 #include "util/hash.h"
 #include "util/mutexlock.h"
@@ -126,9 +127,7 @@ class CompactionJobStatsTest : public testing::Test,
   static void SetUpTestCase() {}
   static void TearDownTestCase() {}
 
-  DBImpl* dbfull() {
-    return reinterpret_cast<DBImpl*>(db_);
-  }
+  DBImpl* dbfull() { return static_cast_with_check<DBImpl>(db_); }
 
   void CreateColumnFamilies(const std::vector<std::string>& cfs,
                             const Options& options) {
@@ -460,6 +459,7 @@ class CompactionJobStatsChecker : public EventListener {
     ASSERT_EQ(current_stats.num_output_files,
         stats.num_output_files);
 
+    ASSERT_EQ(current_stats.is_full_compaction, stats.is_full_compaction);
     ASSERT_EQ(current_stats.is_manual_compaction,
         stats.is_manual_compaction);
 
@@ -572,7 +572,7 @@ CompactionJobStats NewManualCompactionJobStats(
     uint64_t num_input_records, size_t key_size, size_t value_size,
     size_t num_output_files, uint64_t num_output_records,
     double compression_ratio, uint64_t num_records_replaced,
-    bool is_manual = true) {
+    bool is_full = false, bool is_manual = true) {
   CompactionJobStats stats;
   stats.Reset();
 
@@ -596,6 +596,7 @@ CompactionJobStats NewManualCompactionJobStats(
   stats.total_input_raw_value_bytes =
       num_input_records * value_size;
 
+  stats.is_full_compaction = is_full;
   stats.is_manual_compaction = is_manual;
 
   stats.num_records_replaced = num_records_replaced;
@@ -797,7 +798,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
     }
 
     ASSERT_OK(Flush(1));
-    reinterpret_cast<DBImpl*>(db_)->TEST_WaitForCompact();
+    ASSERT_OK(static_cast_with_check<DBImpl>(db_)->TEST_WaitForCompact());
 
     stats_checker->set_verify_next_comp_io_stats(true);
     std::atomic<bool> first_prepare_write(true);
@@ -895,7 +896,7 @@ TEST_P(CompactionJobStatsTest, DeletionStatsTest) {
   CompactRangeOptions cr_options;
   cr_options.change_level = true;
   cr_options.target_level = 2;
-  db_->CompactRange(cr_options, handles_[1], nullptr, nullptr);
+  ASSERT_OK(db_->CompactRange(cr_options, handles_[1], nullptr, nullptr));
   ASSERT_GT(NumTableFilesAtLevel(2, 1), 0);
 
   // Stage 2: Generate files including keys from the entire key range
@@ -982,26 +983,21 @@ TEST_P(CompactionJobStatsTest, UniversalCompactionTest) {
     if (num_input_units == 0) {
       continue;
     }
+    // A full compaction only happens when the number of flushes equals to
+    // the number of compaction input runs.
+    bool is_full = num_flushes == num_input_units;
     // The following statement determines the expected smallest key
-    // based on whether it is a full compaction.  A full compaction only
-    // happens when the number of flushes equals to the number of compaction
-    // input runs.
-    uint64_t smallest_key =
-        (num_flushes == num_input_units) ?
-            key_base : key_base * (num_flushes - 1);
+    // based on whether it is a full compaction.
+    uint64_t smallest_key = is_full ? key_base : key_base * (num_flushes - 1);
 
-    stats_checker->AddExpectedStats(
-        NewManualCompactionJobStats(
-            Key(smallest_key, 10),
-            Key(smallest_key + key_base * num_input_units - key_interval, 10),
-            num_input_units,
-            num_input_units > 2 ? num_input_units / 2 : 0,
-            num_keys_per_table * num_input_units,
-            kKeySize, kValueSize,
-            num_input_units,
-            num_keys_per_table * num_input_units,
-            1.0, 0, false));
-    dbfull()->TEST_WaitForCompact();
+    stats_checker->AddExpectedStats(NewManualCompactionJobStats(
+        Key(smallest_key, 10),
+        Key(smallest_key + key_base * num_input_units - key_interval, 10),
+        num_input_units, num_input_units > 2 ? num_input_units / 2 : 0,
+        num_keys_per_table * num_input_units, kKeySize, kValueSize,
+        num_input_units, num_keys_per_table * num_input_units, 1.0, 0, is_full,
+        false));
+    ASSERT_OK(dbfull()->TEST_WaitForCompact());
   }
   ASSERT_EQ(stats_checker->NumberOfUnverifiedStats(), 3U);
 
@@ -1012,7 +1008,7 @@ TEST_P(CompactionJobStatsTest, UniversalCompactionTest) {
         &rnd, start_key, start_key + key_base - 1,
         kKeySize, kValueSize, key_interval,
         compression_ratio, 1);
-    reinterpret_cast<DBImpl*>(db_)->TEST_WaitForCompact();
+    ASSERT_OK(static_cast_with_check<DBImpl>(db_)->TEST_WaitForCompact());
   }
   ASSERT_EQ(stats_checker->NumberOfUnverifiedStats(), 0U);
 }

db/compaction/compaction_job_test.cc

@@ -76,12 +76,13 @@ class CompactionJobTest : public testing::Test {
         dbname_(test::PerThreadDBPath("compaction_job_test")),
         db_options_(),
         mutable_cf_options_(cf_options_),
+        mutable_db_options_(),
         table_cache_(NewLRUCache(50000, 16)),
         write_buffer_manager_(db_options_.db_write_buffer_size),
-        versions_(new VersionSet(dbname_, &db_options_, env_options_,
-                                 table_cache_.get(), &write_buffer_manager_,
-                                 &write_controller_,
-                                 /*block_cache_tracer=*/nullptr)),
+        versions_(new VersionSet(
+            dbname_, &db_options_, env_options_, table_cache_.get(),
+            &write_buffer_manager_, &write_controller_,
+            /*block_cache_tracer=*/nullptr, /*io_tracer=*/nullptr)),
         shutting_down_(false),
         preserve_deletes_seqnum_(0),
         mock_table_factory_(new mock::MockTableFactory()),
@@ -129,7 +130,7 @@ class CompactionJobTest : public testing::Test {
     return blob_index;
   }
 
-  void AddMockFile(const stl_wrappers::KVMap& contents, int level = 0) {
+  void AddMockFile(const mock::KVVector& contents, int level = 0) {
     assert(contents.size() > 0);
 
     bool first_key = true;
@@ -143,7 +144,7 @@ class CompactionJobTest : public testing::Test {
       std::string skey;
       std::string value;
       std::tie(skey, value) = kv;
-      bool parsed = ParseInternalKey(skey, &key);
+      const Status pikStatus = ParseInternalKey(skey, &key);
 
       smallest_seqno = std::min(smallest_seqno, key.sequence);
       largest_seqno = std::max(largest_seqno, key.sequence);
@@ -161,7 +162,7 @@ class CompactionJobTest : public testing::Test {
 
       first_key = false;
 
-      if (parsed && key.type == kTypeBlobIndex) {
+      if (pikStatus.ok() && key.type == kTypeBlobIndex) {
         BlobIndex blob_index;
         const Status s = blob_index.DecodeFrom(value);
         if (!s.ok()) {
@@ -191,8 +192,9 @@ class CompactionJobTest : public testing::Test {
                  kUnknownFileChecksum, kUnknownFileChecksumFuncName);
 
     mutex_.Lock();
-    versions_->LogAndApply(versions_->GetColumnFamilySet()->GetDefault(),
-                           mutable_cf_options_, &edit, &mutex_);
+    EXPECT_OK(
+        versions_->LogAndApply(versions_->GetColumnFamilySet()->GetDefault(),
+                               mutable_cf_options_, &edit, &mutex_));
     mutex_.Unlock();
   }
 
@@ -203,8 +205,8 @@ class CompactionJobTest : public testing::Test {
   }
 
   // returns expected result after compaction
-  stl_wrappers::KVMap CreateTwoFiles(bool gen_corrupted_keys) {
-    auto expected_results = mock::MakeMockFile();
+  mock::KVVector CreateTwoFiles(bool gen_corrupted_keys) {
+    stl_wrappers::KVMap expected_results;
     const int kKeysPerFile = 10000;
     const int kCorruptKeysPerFile = 200;
     const int kMatchingKeys = kKeysPerFile / 2;
@@ -230,28 +232,34 @@ class CompactionJobTest : public testing::Test {
           test::CorruptKeyType(&internal_key);
           test::CorruptKeyType(&bottommost_internal_key);
         }
-        contents.insert({ internal_key.Encode().ToString(), value });
+        contents.push_back({internal_key.Encode().ToString(), value});
         if (i == 1 || k < kMatchingKeys || corrupt_id(k - kMatchingKeys)) {
           expected_results.insert(
-              { bottommost_internal_key.Encode().ToString(), value });
+              {bottommost_internal_key.Encode().ToString(), value});
         }
       }
+      mock::SortKVVector(&contents);
 
       AddMockFile(contents);
     }
 
     SetLastSequence(sequence_number);
 
-    return expected_results;
+    mock::KVVector expected_results_kvvector;
+    for (auto& kv : expected_results) {
+      expected_results_kvvector.push_back({kv.first, kv.second});
+    }
+
+    return expected_results_kvvector;
   }
 
   void NewDB() {
     DestroyDB(dbname_, Options());
     EXPECT_OK(env_->CreateDirIfMissing(dbname_));
-    versions_.reset(new VersionSet(dbname_, &db_options_, env_options_,
-                                   table_cache_.get(), &write_buffer_manager_,
-                                   &write_controller_,
-                                   /*block_cache_tracer=*/nullptr));
+    versions_.reset(
+        new VersionSet(dbname_, &db_options_, env_options_, table_cache_.get(),
+                       &write_buffer_manager_, &write_controller_,
+                       /*block_cache_tracer=*/nullptr, /*io_tracer=*/nullptr));
     compaction_job_stats_.Reset();
     SetIdentityFile(env_, dbname_);
 
@@ -283,6 +291,8 @@ class CompactionJobTest : public testing::Test {
     // Make "CURRENT" file that points to the new manifest file.
     s = SetCurrentFile(fs_.get(), dbname_, 1, nullptr);
 
+    ASSERT_OK(s);
+
     std::vector<ColumnFamilyDescriptor> column_families;
     cf_options_.table_factory = mock_table_factory_;
     cf_options_.merge_operator = merge_op_;
@@ -295,7 +305,7 @@ class CompactionJobTest : public testing::Test {
 
   void RunCompaction(
       const std::vector<std::vector<FileMetaData*>>& input_files,
-      const stl_wrappers::KVMap& expected_results,
+      const mock::KVVector& expected_results,
       const std::vector<SequenceNumber>& snapshots = {},
       SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber,
       int output_level = 1, bool verify = true,
@@ -316,9 +326,10 @@ class CompactionJobTest : public testing::Test {
 
     Compaction compaction(
         cfd->current()->storage_info(), *cfd->ioptions(),
-        *cfd->GetLatestMutableCFOptions(), compaction_input_files, output_level,
-        1024 * 1024, 10 * 1024 * 1024, 0, kNoCompression,
-        cfd->GetLatestMutableCFOptions()->compression_opts, 0, {}, true);
+        *cfd->GetLatestMutableCFOptions(), mutable_db_options_,
+        compaction_input_files, output_level, 1024 * 1024, 10 * 1024 * 1024, 0,
+        kNoCompression, cfd->GetLatestMutableCFOptions()->compression_opts, 0,
+        {}, true);
     compaction.SetInputVersion(cfd->current());
 
     LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, db_options_.info_log.get());
@@ -332,7 +343,7 @@ class CompactionJobTest : public testing::Test {
         nullptr, nullptr, &mutex_, &error_handler_, snapshots,
         earliest_write_conflict_snapshot, snapshot_checker, table_cache_,
         &event_logger, false, false, dbname_, &compaction_job_stats_,
-        Env::Priority::USER);
+        Env::Priority::USER, nullptr /* IOTracer */);
     VerifyInitializationOfCompactionJobStats(compaction_job_stats_);
 
     compaction_job.Prepare();
@@ -340,8 +351,10 @@ class CompactionJobTest : public testing::Test {
     Status s;
     s = compaction_job.Run();
     ASSERT_OK(s);
+    ASSERT_OK(compaction_job.io_status());
     mutex_.Lock();
     ASSERT_OK(compaction_job.Install(*cfd->GetLatestMutableCFOptions()));
+    ASSERT_OK(compaction_job.io_status());
     mutex_.Unlock();
 
     if (verify) {
@@ -370,6 +383,7 @@ class CompactionJobTest : public testing::Test {
   ImmutableDBOptions db_options_;
   ColumnFamilyOptions cf_options_;
   MutableCFOptions mutable_cf_options_;
+  MutableDBOptions mutable_db_options_;
   std::shared_ptr<Cache> table_cache_;
   WriteController write_controller_;
   WriteBufferManager write_buffer_manager_;
@@ -395,7 +409,7 @@ TEST_F(CompactionJobTest, Simple) {
   RunCompaction({ files }, expected_results);
 }
 
-TEST_F(CompactionJobTest, SimpleCorrupted) {
+TEST_F(CompactionJobTest, DISABLED_SimpleCorrupted) {
   NewDB();
 
   auto expected_results = CreateTwoFiles(true);
@@ -636,7 +650,7 @@ TEST_F(CompactionJobTest, FilterAllMergeOperands) {
   SetLastSequence(11U);
   auto files = cfd_->current()->storage_info()->LevelFiles(0);
 
-  stl_wrappers::KVMap empty_map;
+  mock::KVVector empty_map;
   RunCompaction({files}, empty_map);
 }
 
@@ -989,7 +1003,7 @@ TEST_F(CompactionJobTest, MultiSingleDelete) {
 // single deletion and the (single) deletion gets removed while the corrupt key
 // gets written out. TODO(noetzli): We probably want a better way to treat
 // corrupt keys.
-TEST_F(CompactionJobTest, CorruptionAfterDeletion) {
+TEST_F(CompactionJobTest, DISABLED_CorruptionAfterDeletion) {
   NewDB();
 
   auto file1 =

db/compaction/compaction_picker.cc

@@ -332,7 +332,7 @@ Compaction* CompactionPicker::CompactFiles(
     const CompactionOptions& compact_options,
     const std::vector<CompactionInputFiles>& input_files, int output_level,
     VersionStorageInfo* vstorage, const MutableCFOptions& mutable_cf_options,
-    uint32_t output_path_id) {
+    const MutableDBOptions& mutable_db_options, uint32_t output_path_id) {
   assert(input_files.size());
   // This compaction output should not overlap with a running compaction as
   // `SanitizeCompactionInputFiles` should've checked earlier and db mutex
@@ -356,8 +356,8 @@ Compaction* CompactionPicker::CompactFiles(
     compression_type = compact_options.compression;
   }
   auto c = new Compaction(
-      vstorage, ioptions_, mutable_cf_options, input_files, output_level,
-      compact_options.output_file_size_limit,
+      vstorage, ioptions_, mutable_cf_options, mutable_db_options, input_files,
+      output_level, compact_options.output_file_size_limit,
       mutable_cf_options.max_compaction_bytes, output_path_id, compression_type,
       GetCompressionOptions(mutable_cf_options, vstorage, output_level),
       compact_options.max_subcompactions,
@@ -563,7 +563,8 @@ void CompactionPicker::GetGrandparents(
 
 Compaction* CompactionPicker::CompactRange(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, int input_level, int output_level,
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    int input_level, int output_level,
     const CompactRangeOptions& compact_range_options, const InternalKey* begin,
     const InternalKey* end, InternalKey** compaction_end, bool* manual_conflict,
     uint64_t max_file_num_to_ignore) {
@@ -626,8 +627,8 @@ Compaction* CompactionPicker::CompactRange(
     }
 
     Compaction* c = new Compaction(
-        vstorage, ioptions_, mutable_cf_options, std::move(inputs),
-        output_level,
+        vstorage, ioptions_, mutable_cf_options, mutable_db_options,
+        std::move(inputs), output_level,
         MaxFileSizeForLevel(mutable_cf_options, output_level,
                             ioptions_.compaction_style),
         /* max_compaction_bytes */ LLONG_MAX,
@@ -638,6 +639,7 @@ Compaction* CompactionPicker::CompactRange(
         compact_range_options.max_subcompactions, /* grandparents */ {},
         /* is manual */ true);
     RegisterCompaction(c);
+    vstorage->ComputeCompactionScore(ioptions_, mutable_cf_options);
     return c;
   }
 
@@ -778,8 +780,8 @@ Compaction* CompactionPicker::CompactRange(
   std::vector<FileMetaData*> grandparents;
   GetGrandparents(vstorage, inputs, output_level_inputs, &grandparents);
   Compaction* compaction = new Compaction(
-      vstorage, ioptions_, mutable_cf_options, std::move(compaction_inputs),
-      output_level,
+      vstorage, ioptions_, mutable_cf_options, mutable_db_options,
+      std::move(compaction_inputs), output_level,
       MaxFileSizeForLevel(mutable_cf_options, output_level,
                           ioptions_.compaction_style, vstorage->base_level(),
                           ioptions_.level_compaction_dynamic_level_bytes),

db/compaction/compaction_picker.h

@@ -56,7 +56,8 @@ class CompactionPicker {
   // describes the compaction.  Caller should delete the result.
   virtual Compaction* PickCompaction(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, LogBuffer* log_buffer,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      LogBuffer* log_buffer,
       SequenceNumber earliest_memtable_seqno = kMaxSequenceNumber) = 0;
 
   // Return a compaction object for compacting the range [begin,end] in
@@ -72,7 +73,8 @@ class CompactionPicker {
   // *compaction_end should point to valid InternalKey!
   virtual Compaction* CompactRange(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, int input_level, int output_level,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      int input_level, int output_level,
       const CompactRangeOptions& compact_range_options,
       const InternalKey* begin, const InternalKey* end,
       InternalKey** compaction_end, bool* manual_conflict,
@@ -113,6 +115,7 @@ class CompactionPicker {
                            const std::vector<CompactionInputFiles>& input_files,
                            int output_level, VersionStorageInfo* vstorage,
                            const MutableCFOptions& mutable_cf_options,
+                           const MutableDBOptions& mutable_db_options,
                            uint32_t output_path_id);
 
   // Converts a set of compaction input file numbers into
@@ -250,6 +253,7 @@ class NullCompactionPicker : public CompactionPicker {
   Compaction* PickCompaction(
       const std::string& /*cf_name*/,
       const MutableCFOptions& /*mutable_cf_options*/,
+      const MutableDBOptions& /*mutable_db_options*/,
       VersionStorageInfo* /*vstorage*/, LogBuffer* /* log_buffer */,
       SequenceNumber /* earliest_memtable_seqno */) override {
     return nullptr;
@@ -258,6 +262,7 @@ class NullCompactionPicker : public CompactionPicker {
   // Always return "nullptr"
   Compaction* CompactRange(const std::string& /*cf_name*/,
                            const MutableCFOptions& /*mutable_cf_options*/,
+                           const MutableDBOptions& /*mutable_db_options*/,
                            VersionStorageInfo* /*vstorage*/,
                            int /*input_level*/, int /*output_level*/,
                            const CompactRangeOptions& /*compact_range_options*/,

db/compaction/compaction_picker_fifo.cc

@@ -36,7 +36,8 @@ bool FIFOCompactionPicker::NeedsCompaction(
 
 Compaction* FIFOCompactionPicker::PickTTLCompaction(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    LogBuffer* log_buffer) {
   assert(mutable_cf_options.ttl > 0);
 
   const int kLevel0 = 0;
@@ -108,8 +109,9 @@ Compaction* FIFOCompactionPicker::PickTTLCompaction(
   }
 
   Compaction* c = new Compaction(
-      vstorage, ioptions_, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
-      kNoCompression, mutable_cf_options.compression_opts,
+      vstorage, ioptions_, mutable_cf_options, mutable_db_options,
+      std::move(inputs), 0, 0, 0, 0, kNoCompression,
+      mutable_cf_options.compression_opts,
       /* max_subcompactions */ 0, {}, /* is manual */ false,
       vstorage->CompactionScore(0),
       /* is deletion compaction */ true, CompactionReason::kFIFOTtl);
@@ -118,7 +120,8 @@ Compaction* FIFOCompactionPicker::PickTTLCompaction(
 
 Compaction* FIFOCompactionPicker::PickSizeCompaction(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    LogBuffer* log_buffer) {
   const int kLevel0 = 0;
   const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
   uint64_t total_size = GetTotalFilesSize(level_files);
@@ -147,8 +150,8 @@ Compaction* FIFOCompactionPicker::PickSizeCompaction(
               max_compact_bytes_per_del_file,
               mutable_cf_options.max_compaction_bytes, &comp_inputs)) {
         Compaction* c = new Compaction(
-            vstorage, ioptions_, mutable_cf_options, {comp_inputs}, 0,
-            16 * 1024 * 1024 /* output file size limit */,
+            vstorage, ioptions_, mutable_cf_options, mutable_db_options,
+            {comp_inputs}, 0, 16 * 1024 * 1024 /* output file size limit */,
             0 /* max compaction bytes, not applicable */,
             0 /* output path ID */, mutable_cf_options.compression,
             mutable_cf_options.compression_opts, 0 /* max_subcompactions */, {},
@@ -198,8 +201,9 @@ Compaction* FIFOCompactionPicker::PickSizeCompaction(
   }
 
   Compaction* c = new Compaction(
-      vstorage, ioptions_, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
-      kNoCompression, mutable_cf_options.compression_opts,
+      vstorage, ioptions_, mutable_cf_options, mutable_db_options,
+      std::move(inputs), 0, 0, 0, 0, kNoCompression,
+      mutable_cf_options.compression_opts,
       /* max_subcompactions */ 0, {}, /* is manual */ false,
       vstorage->CompactionScore(0),
       /* is deletion compaction */ true, CompactionReason::kFIFOMaxSize);
@@ -208,16 +212,18 @@ Compaction* FIFOCompactionPicker::PickSizeCompaction(
 
 Compaction* FIFOCompactionPicker::PickCompaction(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer,
-    SequenceNumber /*earliest_memtable_seqno*/) {
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    LogBuffer* log_buffer, SequenceNumber /*earliest_memtable_seqno*/) {
   assert(vstorage->num_levels() == 1);
 
   Compaction* c = nullptr;
   if (mutable_cf_options.ttl > 0) {
-    c = PickTTLCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
+    c = PickTTLCompaction(cf_name, mutable_cf_options, mutable_db_options,
+                          vstorage, log_buffer);
   }
   if (c == nullptr) {
-    c = PickSizeCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
+    c = PickSizeCompaction(cf_name, mutable_cf_options, mutable_db_options,
+                           vstorage, log_buffer);
   }
   RegisterCompaction(c);
   return c;
@@ -225,7 +231,8 @@ Compaction* FIFOCompactionPicker::PickCompaction(
 
 Compaction* FIFOCompactionPicker::CompactRange(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, int input_level, int output_level,
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    int input_level, int output_level,
     const CompactRangeOptions& /*compact_range_options*/,
     const InternalKey* /*begin*/, const InternalKey* /*end*/,
     InternalKey** compaction_end, bool* /*manual_conflict*/,
@@ -238,8 +245,8 @@ Compaction* FIFOCompactionPicker::CompactRange(
   assert(output_level == 0);
   *compaction_end = nullptr;
   LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, ioptions_.info_log);
-  Compaction* c =
-      PickCompaction(cf_name, mutable_cf_options, vstorage, &log_buffer);
+  Compaction* c = PickCompaction(cf_name, mutable_cf_options,
+                                 mutable_db_options, vstorage, &log_buffer);
   log_buffer.FlushBufferToLog();
   return c;
 }

db/compaction/compaction_picker_fifo.h

@@ -21,12 +21,14 @@ class FIFOCompactionPicker : public CompactionPicker {
 
   virtual Compaction* PickCompaction(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* version, LogBuffer* log_buffer,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* version,
+      LogBuffer* log_buffer,
       SequenceNumber earliest_memtable_seqno = kMaxSequenceNumber) override;
 
   virtual Compaction* CompactRange(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, int input_level, int output_level,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      int input_level, int output_level,
       const CompactRangeOptions& compact_range_options,
       const InternalKey* begin, const InternalKey* end,
       InternalKey** compaction_end, bool* manual_conflict,
@@ -41,11 +43,13 @@ class FIFOCompactionPicker : public CompactionPicker {
  private:
   Compaction* PickTTLCompaction(const std::string& cf_name,
                                 const MutableCFOptions& mutable_cf_options,
+                                const MutableDBOptions& mutable_db_options,
                                 VersionStorageInfo* version,
                                 LogBuffer* log_buffer);
 
   Compaction* PickSizeCompaction(const std::string& cf_name,
                                  const MutableCFOptions& mutable_cf_options,
+                                 const MutableDBOptions& mutable_db_options,
                                  VersionStorageInfo* version,
                                  LogBuffer* log_buffer);
 };

db/compaction/compaction_picker_level.cc

@@ -49,14 +49,16 @@ class LevelCompactionBuilder {
                          CompactionPicker* compaction_picker,
                          LogBuffer* log_buffer,
                          const MutableCFOptions& mutable_cf_options,
-                         const ImmutableCFOptions& ioptions)
+                         const ImmutableCFOptions& ioptions,
+                         const MutableDBOptions& mutable_db_options)
       : cf_name_(cf_name),
         vstorage_(vstorage),
         earliest_mem_seqno_(earliest_mem_seqno),
         compaction_picker_(compaction_picker),
         log_buffer_(log_buffer),
         mutable_cf_options_(mutable_cf_options),
-        ioptions_(ioptions) {}
+        ioptions_(ioptions),
+        mutable_db_options_(mutable_db_options) {}
 
   // Pick and return a compaction.
   Compaction* PickCompaction();
@@ -120,6 +122,7 @@ class LevelCompactionBuilder {
 
   const MutableCFOptions& mutable_cf_options_;
   const ImmutableCFOptions& ioptions_;
+  const MutableDBOptions& mutable_db_options_;
   // Pick a path ID to place a newly generated file, with its level
   static uint32_t GetPathId(const ImmutableCFOptions& ioptions,
                             const MutableCFOptions& mutable_cf_options,
@@ -323,8 +326,8 @@ Compaction* LevelCompactionBuilder::PickCompaction() {
 
 Compaction* LevelCompactionBuilder::GetCompaction() {
   auto c = new Compaction(
-      vstorage_, ioptions_, mutable_cf_options_, std::move(compaction_inputs_),
-      output_level_,
+      vstorage_, ioptions_, mutable_cf_options_, mutable_db_options_,
+      std::move(compaction_inputs_), output_level_,
       MaxFileSizeForLevel(mutable_cf_options_, output_level_,
                           ioptions_.compaction_style, vstorage_->base_level(),
                           ioptions_.level_compaction_dynamic_level_bytes),
@@ -497,10 +500,11 @@ bool LevelCompactionBuilder::PickIntraL0Compaction() {
 
 Compaction* LevelCompactionPicker::PickCompaction(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer,
-    SequenceNumber earliest_mem_seqno) {
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    LogBuffer* log_buffer, SequenceNumber earliest_mem_seqno) {
   LevelCompactionBuilder builder(cf_name, vstorage, earliest_mem_seqno, this,
-                                 log_buffer, mutable_cf_options, ioptions_);
+                                 log_buffer, mutable_cf_options, ioptions_,
+                                 mutable_db_options);
   return builder.PickCompaction();
 }
 }  // namespace ROCKSDB_NAMESPACE

db/compaction/compaction_picker_level.h

@@ -22,7 +22,8 @@ class LevelCompactionPicker : public CompactionPicker {
       : CompactionPicker(ioptions, icmp) {}
   virtual Compaction* PickCompaction(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, LogBuffer* log_buffer,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      LogBuffer* log_buffer,
       SequenceNumber earliest_memtable_seqno = kMaxSequenceNumber) override;
 
   virtual bool NeedsCompaction(

db/compaction/compaction_picker_test.cc

@@ -33,6 +33,7 @@ class CompactionPickerTest : public testing::Test {
   Options options_;
   ImmutableCFOptions ioptions_;
   MutableCFOptions mutable_cf_options_;
+  MutableDBOptions mutable_db_options_;
   LevelCompactionPicker level_compaction_picker;
   std::string cf_name_;
   CountingLogger logger_;
@@ -52,6 +53,7 @@ class CompactionPickerTest : public testing::Test {
         icmp_(ucmp_),
         ioptions_(options_),
         mutable_cf_options_(options_),
+        mutable_db_options_(),
         level_compaction_picker(ioptions_, &icmp_),
         cf_name_("dummy"),
         log_buffer_(InfoLogLevel::INFO_LEVEL, &logger_),
@@ -181,7 +183,8 @@ TEST_F(CompactionPickerTest, Empty) {
   NewVersionStorage(6, kCompactionStyleLevel);
   UpdateVersionStorageInfo();
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() == nullptr);
 }
 
@@ -192,7 +195,8 @@ TEST_F(CompactionPickerTest, Single) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() == nullptr);
 }
 
@@ -205,7 +209,8 @@ TEST_F(CompactionPickerTest, Level0Trigger) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_files(0));
   ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
@@ -218,7 +223,8 @@ TEST_F(CompactionPickerTest, Level1Trigger) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   ASSERT_EQ(66U, compaction->input(0, 0)->fd.GetNumber());
@@ -236,7 +242,8 @@ TEST_F(CompactionPickerTest, Level1Trigger2) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   ASSERT_EQ(2U, compaction->num_input_files(1));
@@ -267,7 +274,8 @@ TEST_F(CompactionPickerTest, LevelMaxScore) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   ASSERT_EQ(7U, compaction->input(0, 0)->fd.GetNumber());
@@ -314,7 +322,8 @@ TEST_F(CompactionPickerTest, Level0TriggerDynamic) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_files(0));
   ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
@@ -338,7 +347,8 @@ TEST_F(CompactionPickerTest, Level0TriggerDynamic2) {
   ASSERT_EQ(vstorage_->base_level(), num_levels - 2);
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_files(0));
   ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
@@ -363,7 +373,8 @@ TEST_F(CompactionPickerTest, Level0TriggerDynamic3) {
   ASSERT_EQ(vstorage_->base_level(), num_levels - 3);
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_files(0));
   ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
@@ -392,7 +403,8 @@ TEST_F(CompactionPickerTest, Level0TriggerDynamic4) {
   ASSERT_EQ(vstorage_->base_level(), num_levels - 3);
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_files(0));
   ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
@@ -424,7 +436,8 @@ TEST_F(CompactionPickerTest, LevelTriggerDynamic4) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   ASSERT_EQ(5U, compaction->input(0, 0)->fd.GetNumber());
@@ -481,7 +494,8 @@ TEST_F(CompactionPickerTest, CompactionUniversalIngestBehindReservedLevel) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   // output level should be the one above the bottom-most
   ASSERT_EQ(1, compaction->output_level());
@@ -515,7 +529,8 @@ TEST_F(CompactionPickerTest, CannotTrivialMoveUniversal) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_TRUE(!compaction->is_trivial_move());
 }
@@ -541,7 +556,8 @@ TEST_F(CompactionPickerTest, AllowsTrivialMoveUniversal) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_TRUE(compaction->is_trivial_move());
 }
@@ -569,7 +585,8 @@ TEST_F(CompactionPickerTest, UniversalPeriodicCompaction1) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_TRUE(compaction);
   ASSERT_EQ(4, compaction->output_level());
@@ -599,7 +616,8 @@ TEST_F(CompactionPickerTest, UniversalPeriodicCompaction2) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_FALSE(compaction);
 }
@@ -625,7 +643,8 @@ TEST_F(CompactionPickerTest, UniversalPeriodicCompaction3) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_FALSE(compaction);
 }
@@ -655,7 +674,8 @@ TEST_F(CompactionPickerTest, UniversalPeriodicCompaction4) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
   ASSERT_TRUE(!compaction ||
               compaction->start_level() != compaction->output_level());
 }
@@ -675,7 +695,8 @@ TEST_F(CompactionPickerTest, UniversalPeriodicCompaction5) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
   ASSERT_TRUE(compaction);
   ASSERT_EQ(0, compaction->start_level());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -699,7 +720,8 @@ TEST_F(CompactionPickerTest, UniversalPeriodicCompaction6) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
   ASSERT_TRUE(compaction);
   ASSERT_EQ(4, compaction->start_level());
   ASSERT_EQ(2U, compaction->num_input_files(0));
@@ -759,7 +781,8 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping1) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   // Pick file 8 because it overlaps with 0 files on level 3.
@@ -791,7 +814,8 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping2) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   // Picking file 7 because overlapping ratio is the biggest.
@@ -818,7 +842,8 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping3) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   // Picking file 8 because overlapping ratio is the biggest.
@@ -847,7 +872,8 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping4) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   // Picking file 8 because overlapping ratio is the biggest.
@@ -874,7 +900,8 @@ TEST_F(CompactionPickerTest, ParentIndexResetBug) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
 }
 
 // This test checks ExpandWhileOverlapping() by having overlapping user keys
@@ -891,7 +918,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-              cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_levels());
   ASSERT_EQ(2U, compaction->num_input_files(0));
@@ -910,7 +938,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys2) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-              cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(2U, compaction->num_input_files(0));
@@ -937,7 +966,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys3) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-              cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(5U, compaction->num_input_files(0));
@@ -967,7 +997,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys4) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -990,7 +1021,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys5) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() == nullptr);
 }
 
@@ -1011,7 +1043,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys6) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -1031,7 +1064,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys7) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_GE(1U, compaction->num_input_files(0));
@@ -1059,7 +1093,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys8) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(3U, compaction->num_input_files(0));
@@ -1091,7 +1126,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys9) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(5U, compaction->num_input_files(0));
@@ -1131,7 +1167,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys10) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -1169,7 +1206,8 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys11) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -1206,7 +1244,8 @@ TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri1) {
   ASSERT_EQ(0, vstorage_->CompactionScoreLevel(0));
   ASSERT_EQ(1, vstorage_->CompactionScoreLevel(1));
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() == nullptr);
 }
 
@@ -1236,7 +1275,8 @@ TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri2) {
   ASSERT_EQ(0, vstorage_->CompactionScoreLevel(0));
   ASSERT_EQ(1, vstorage_->CompactionScoreLevel(1));
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
 }
 
@@ -1269,7 +1309,8 @@ TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri3) {
   ASSERT_EQ(1, vstorage_->CompactionScoreLevel(0));
   ASSERT_EQ(0, vstorage_->CompactionScoreLevel(1));
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
 }
 
@@ -1563,7 +1604,8 @@ TEST_F(CompactionPickerTest, MaxCompactionBytesHit) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -1587,7 +1629,8 @@ TEST_F(CompactionPickerTest, MaxCompactionBytesNotHit) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(2U, compaction->num_input_levels());
   ASSERT_EQ(3U, compaction->num_input_files(0));
@@ -1616,11 +1659,38 @@ TEST_F(CompactionPickerTest, IsTrivialMoveOn) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-    cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_TRUE(compaction->IsTrivialMove());
 }
 
+TEST_F(CompactionPickerTest, IsTrivialMoveOffSstPartitioned) {
+  mutable_cf_options_.max_bytes_for_level_base = 10000u;
+  mutable_cf_options_.max_compaction_bytes = 10001u;
+  ioptions_.level_compaction_dynamic_level_bytes = false;
+  ioptions_.sst_partitioner_factory = NewSstPartitionerFixedPrefixFactory(1);
+  NewVersionStorage(6, kCompactionStyleLevel);
+  // A compaction should be triggered and pick file 2
+  Add(1, 1U, "100", "150", 3000U);
+  Add(1, 2U, "151", "200", 3001U);
+  Add(1, 3U, "201", "250", 3000U);
+  Add(1, 4U, "251", "300", 3000U);
+
+  Add(3, 5U, "120", "130", 7000U);
+  Add(3, 6U, "170", "180", 7000U);
+  Add(3, 7U, "220", "230", 7000U);
+  Add(3, 8U, "270", "280", 7000U);
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  // No trivial move, because partitioning is applied
+  ASSERT_TRUE(!compaction->IsTrivialMove());
+}
+
 TEST_F(CompactionPickerTest, IsTrivialMoveOff) {
   mutable_cf_options_.max_bytes_for_level_base = 1000000u;
   mutable_cf_options_.max_compaction_bytes = 10000u;
@@ -1637,7 +1707,8 @@ TEST_F(CompactionPickerTest, IsTrivialMoveOff) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-    cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_FALSE(compaction->IsTrivialMove());
 }
@@ -1662,7 +1733,8 @@ TEST_F(CompactionPickerTest, CacheNextCompactionIndex) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -1671,7 +1743,8 @@ TEST_F(CompactionPickerTest, CacheNextCompactionIndex) {
   ASSERT_EQ(2, vstorage_->NextCompactionIndex(1 /* level */));
 
   compaction.reset(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_levels());
   ASSERT_EQ(1U, compaction->num_input_files(0));
@@ -1680,7 +1753,8 @@ TEST_F(CompactionPickerTest, CacheNextCompactionIndex) {
   ASSERT_EQ(3, vstorage_->NextCompactionIndex(1 /* level */));
 
   compaction.reset(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() == nullptr);
   ASSERT_EQ(4, vstorage_->NextCompactionIndex(1 /* level */));
 }
@@ -1705,7 +1779,8 @@ TEST_F(CompactionPickerTest, IntraL0MaxCompactionBytesNotHit) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_levels());
   ASSERT_EQ(5U, compaction->num_input_files(0));
@@ -1735,7 +1810,8 @@ TEST_F(CompactionPickerTest, IntraL0MaxCompactionBytesHit) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_levels());
   ASSERT_EQ(4U, compaction->num_input_files(0));
@@ -1767,7 +1843,8 @@ TEST_F(CompactionPickerTest, IntraL0ForEarliestSeqno) {
   UpdateVersionStorageInfo();
 
   std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_, 107));
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_, 107));
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_levels());
   ASSERT_EQ(4U, compaction->num_input_files(0));
@@ -1799,7 +1876,8 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionFullOverlap) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_TRUE(compaction);
   // Validate that its a compaction to reduce sorted runs
@@ -1816,7 +1894,8 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionFullOverlap) {
 
   std::unique_ptr<Compaction> compaction2(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
   ASSERT_FALSE(compaction2);
 }
 
@@ -1840,7 +1919,8 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionFullOverlap2) {
 
   std::unique_ptr<Compaction> compaction(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
 
   ASSERT_TRUE(compaction);
   // Validate that its a delete triggered compaction
@@ -1858,7 +1938,8 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionFullOverlap2) {
 
   std::unique_ptr<Compaction> compaction2(
       universal_compaction_picker.PickCompaction(
-          cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
   ASSERT_FALSE(compaction2);
 }
 
@@ -1898,7 +1979,8 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionStartOutputOverlap) {
 
     std::unique_ptr<Compaction> compaction(
         universal_compaction_picker.PickCompaction(
-            cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+            cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+            &log_buffer_));
 
     ASSERT_TRUE(compaction);
     // Validate that its a delete triggered compaction
@@ -1928,11 +2010,177 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionStartOutputOverlap) {
     random_index = 0;
     std::unique_ptr<Compaction> compaction2(
         universal_compaction_picker.PickCompaction(
-            cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+            cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+            &log_buffer_));
     ASSERT_FALSE(compaction2);
     DeleteVersionStorage();
   }
 }
+
+TEST_F(CompactionPickerTest, UniversalMarkedL0NoOverlap) {
+  const uint64_t kFileSize = 100000;
+
+  ioptions_.compaction_style = kCompactionStyleUniversal;
+  UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+
+  // This test covers the case where a delete triggered compaction is
+  // scheduled and should result in a full compaction
+  NewVersionStorage(1, kCompactionStyleUniversal);
+
+  // Mark file number 4 for compaction
+  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 260, 300, 0, true);
+  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 201, 250);
+  Add(0, 3U, "301", "350", 4 * kFileSize, 0, 101, 150);
+  Add(0, 6U, "501", "750", 8 * kFileSize, 0, 50, 100);
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(
+      universal_compaction_picker.PickCompaction(
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
+
+  ASSERT_TRUE(compaction);
+  // Validate that its a delete triggered compaction
+  ASSERT_EQ(CompactionReason::kFilesMarkedForCompaction,
+            compaction->compaction_reason());
+  ASSERT_EQ(0, compaction->output_level());
+  ASSERT_EQ(0, compaction->start_level());
+  ASSERT_EQ(4U, compaction->num_input_files(0));
+  ASSERT_TRUE(file_map_[4].first->being_compacted);
+  ASSERT_TRUE(file_map_[5].first->being_compacted);
+  ASSERT_TRUE(file_map_[3].first->being_compacted);
+  ASSERT_TRUE(file_map_[6].first->being_compacted);
+}
+
+TEST_F(CompactionPickerTest, UniversalMarkedL0WithOverlap) {
+  const uint64_t kFileSize = 100000;
+
+  ioptions_.compaction_style = kCompactionStyleUniversal;
+  UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+
+  // This test covers the case where a file is being compacted, and a
+  // delete triggered compaction is then scheduled. The latter should stop
+  // at the first file being compacted
+  NewVersionStorage(1, kCompactionStyleUniversal);
+
+  // Mark file number 4 for compaction
+  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 260, 300, 0, true);
+  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 201, 250);
+  Add(0, 3U, "301", "350", 4 * kFileSize, 0, 101, 150);
+  Add(0, 6U, "501", "750", 8 * kFileSize, 0, 50, 100);
+  UpdateVersionStorageInfo();
+  file_map_[3].first->being_compacted = true;
+
+  std::unique_ptr<Compaction> compaction(
+      universal_compaction_picker.PickCompaction(
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
+
+  ASSERT_TRUE(compaction);
+  // Validate that its a delete triggered compaction
+  ASSERT_EQ(CompactionReason::kFilesMarkedForCompaction,
+            compaction->compaction_reason());
+  ASSERT_EQ(0, compaction->output_level());
+  ASSERT_EQ(0, compaction->start_level());
+  ASSERT_EQ(2U, compaction->num_input_files(0));
+  ASSERT_TRUE(file_map_[4].first->being_compacted);
+  ASSERT_TRUE(file_map_[5].first->being_compacted);
+}
+
+TEST_F(CompactionPickerTest, UniversalMarkedL0Overlap2) {
+  const uint64_t kFileSize = 100000;
+
+  ioptions_.compaction_style = kCompactionStyleUniversal;
+  UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+
+  // This test covers the case where a delete triggered compaction is
+  // scheduled first, followed by a "regular" compaction. The latter
+  // should fail
+  NewVersionStorage(1, kCompactionStyleUniversal);
+
+  // Mark file number 4 for compaction
+  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 260, 300);
+  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 201, 250, 0, true);
+  Add(0, 3U, "301", "350", 4 * kFileSize, 0, 101, 150);
+  Add(0, 6U, "501", "750", 8 * kFileSize, 0, 50, 100);
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(
+      universal_compaction_picker.PickCompaction(
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
+
+  ASSERT_TRUE(compaction);
+  // Validate that its a delete triggered compaction
+  ASSERT_EQ(CompactionReason::kFilesMarkedForCompaction,
+            compaction->compaction_reason());
+  ASSERT_EQ(0, compaction->output_level());
+  ASSERT_EQ(0, compaction->start_level());
+  ASSERT_EQ(3U, compaction->num_input_files(0));
+  ASSERT_TRUE(file_map_[5].first->being_compacted);
+  ASSERT_TRUE(file_map_[3].first->being_compacted);
+  ASSERT_TRUE(file_map_[6].first->being_compacted);
+
+  AddVersionStorage();
+  Add(0, 1U, "150", "200", kFileSize, 0, 500, 550);
+  Add(0, 2U, "201", "250", kFileSize, 0, 401, 450);
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction2(
+      universal_compaction_picker.PickCompaction(
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          &log_buffer_));
+  ASSERT_TRUE(compaction2);
+  ASSERT_EQ(3U, compaction->num_input_files(0));
+  ASSERT_TRUE(file_map_[1].first->being_compacted);
+  ASSERT_TRUE(file_map_[2].first->being_compacted);
+  ASSERT_TRUE(file_map_[4].first->being_compacted);
+}
+
+TEST_F(CompactionPickerTest, UniversalMarkedManualCompaction) {
+  const uint64_t kFileSize = 100000;
+  const int kNumLevels = 7;
+
+  // This test makes sure the `files_marked_for_compaction_` is updated after
+  // creating manual compaction.
+  ioptions_.compaction_style = kCompactionStyleUniversal;
+  UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+
+  NewVersionStorage(kNumLevels, kCompactionStyleUniversal);
+
+  // Add 3 files marked for compaction
+  Add(0, 3U, "301", "350", 4 * kFileSize, 0, 101, 150, 0, true);
+  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 260, 300, 0, true);
+  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 201, 250, 0, true);
+  UpdateVersionStorageInfo();
+
+  // All 3 files are marked for compaction
+  ASSERT_EQ(3U, vstorage_->FilesMarkedForCompaction().size());
+
+  bool manual_conflict = false;
+  InternalKey* manual_end = NULL;
+  std::unique_ptr<Compaction> compaction(
+      universal_compaction_picker.CompactRange(
+          cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+          ColumnFamilyData::kCompactAllLevels, 6, CompactRangeOptions(), NULL,
+          NULL, &manual_end, &manual_conflict, port::kMaxUint64));
+
+  ASSERT_TRUE(compaction);
+
+  ASSERT_EQ(CompactionReason::kManualCompaction,
+            compaction->compaction_reason());
+  ASSERT_EQ(kNumLevels - 1, compaction->output_level());
+  ASSERT_EQ(0, compaction->start_level());
+  ASSERT_EQ(3U, compaction->num_input_files(0));
+  ASSERT_TRUE(file_map_[3].first->being_compacted);
+  ASSERT_TRUE(file_map_[4].first->being_compacted);
+  ASSERT_TRUE(file_map_[5].first->being_compacted);
+
+  // After creating the manual compaction, all files should be cleared from
+  // `FilesMarkedForCompaction`. So they won't be picked by others.
+  ASSERT_EQ(0U, vstorage_->FilesMarkedForCompaction().size());
+}
+
 #endif  // ROCKSDB_LITE
 
 }  // namespace ROCKSDB_NAMESPACE

db/compaction/compaction_picker_universal.cc

@@ -31,17 +31,16 @@ namespace {
 // PickCompaction().
 class UniversalCompactionBuilder {
  public:
-  UniversalCompactionBuilder(const ImmutableCFOptions& ioptions,
-                             const InternalKeyComparator* icmp,
-                             const std::string& cf_name,
-                             const MutableCFOptions& mutable_cf_options,
-                             VersionStorageInfo* vstorage,
-                             UniversalCompactionPicker* picker,
-                             LogBuffer* log_buffer)
+  UniversalCompactionBuilder(
+      const ImmutableCFOptions& ioptions, const InternalKeyComparator* icmp,
+      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      UniversalCompactionPicker* picker, LogBuffer* log_buffer)
       : ioptions_(ioptions),
         icmp_(icmp),
         cf_name_(cf_name),
         mutable_cf_options_(mutable_cf_options),
+        mutable_db_options_(mutable_db_options),
         vstorage_(vstorage),
         picker_(picker),
         log_buffer_(log_buffer) {}
@@ -115,6 +114,7 @@ class UniversalCompactionBuilder {
   std::vector<SortedRun> sorted_runs_;
   const std::string& cf_name_;
   const MutableCFOptions& mutable_cf_options_;
+  const MutableDBOptions& mutable_db_options_;
   VersionStorageInfo* vstorage_;
   UniversalCompactionPicker* picker_;
   LogBuffer* log_buffer_;
@@ -277,11 +277,11 @@ bool UniversalCompactionPicker::NeedsCompaction(
 
 Compaction* UniversalCompactionPicker::PickCompaction(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer,
-    SequenceNumber /* earliest_memtable_seqno */) {
+    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+    LogBuffer* log_buffer, SequenceNumber /* earliest_memtable_seqno */) {
   UniversalCompactionBuilder builder(ioptions_, icmp_, cf_name,
-                                     mutable_cf_options, vstorage, this,
-                                     log_buffer);
+                                     mutable_cf_options, mutable_db_options,
+                                     vstorage, this, log_buffer);
   return builder.PickCompaction();
 }
 
@@ -461,7 +461,6 @@ Compaction* UniversalCompactionBuilder::PickCompaction() {
 
 // validate that all the chosen files of L0 are non overlapping in time
 #ifndef NDEBUG
-  SequenceNumber prev_smallest_seqno = 0U;
   bool is_first = true;
 
   size_t level_index = 0U;
@@ -471,7 +470,6 @@ Compaction* UniversalCompactionBuilder::PickCompaction() {
       if (is_first) {
         is_first = false;
       }
-      prev_smallest_seqno = f->fd.smallest_seqno;
     }
     level_index = 1U;
   }
@@ -483,16 +481,7 @@ Compaction* UniversalCompactionBuilder::PickCompaction() {
                               &largest_seqno);
       if (is_first) {
         is_first = false;
-      } else if (prev_smallest_seqno > 0) {
-        // A level is considered as the bottommost level if there are
-        // no files in higher levels or if files in higher levels do
-        // not overlap with the files being compacted. Sequence numbers
-        // of files in bottommost level can be set to 0 to help
-        // compression. As a result, the following assert may not hold
-        // if the prev_smallest_seqno is 0.
-        assert(prev_smallest_seqno > largest_seqno);
       }
-      prev_smallest_seqno = smallest_seqno;
     }
   }
 #endif
@@ -730,8 +719,8 @@ Compaction* UniversalCompactionBuilder::PickCompactionToReduceSortedRuns(
     compaction_reason = CompactionReason::kUniversalSortedRunNum;
   }
   return new Compaction(
-      vstorage_, ioptions_, mutable_cf_options_, std::move(inputs),
-      output_level,
+      vstorage_, ioptions_, mutable_cf_options_, mutable_db_options_,
+      std::move(inputs), output_level,
       MaxFileSizeForLevel(mutable_cf_options_, output_level,
                           kCompactionStyleUniversal),
       LLONG_MAX, path_id,
@@ -841,34 +830,48 @@ Compaction* UniversalCompactionBuilder::PickDeleteTriggeredCompaction() {
   std::vector<CompactionInputFiles> inputs;
 
   if (vstorage_->num_levels() == 1) {
-#if defined(ENABLE_SINGLE_LEVEL_DTC)
     // This is single level universal. Since we're basically trying to reclaim
     // space by processing files marked for compaction due to high tombstone
     // density, let's do the same thing as compaction to reduce size amp which
     // has the same goals.
-    bool compact = false;
+    int start_index = -1;
 
     start_level_inputs.level = 0;
     start_level_inputs.files.clear();
     output_level = 0;
-    for (FileMetaData* f : vstorage_->LevelFiles(0)) {
+    // Find the first file marked for compaction. Ignore the last file
+    for (size_t loop = 0; loop + 1 < sorted_runs_.size(); loop++) {
+      SortedRun* sr = &sorted_runs_[loop];
+      if (sr->being_compacted) {
+        continue;
+      }
+      FileMetaData* f = vstorage_->LevelFiles(0)[loop];
       if (f->marked_for_compaction) {
-        compact = true;
-      }
-      if (compact) {
         start_level_inputs.files.push_back(f);
+        start_index =
+            static_cast<int>(loop);  // Consider this as the first candidate.
+        break;
       }
     }
+    if (start_index < 0) {
+      // Either no file marked, or they're already being compacted
+      return nullptr;
+    }
+
+    for (size_t loop = start_index + 1; loop < sorted_runs_.size(); loop++) {
+      SortedRun* sr = &sorted_runs_[loop];
+      if (sr->being_compacted) {
+        break;
+      }
+
+      FileMetaData* f = vstorage_->LevelFiles(0)[loop];
+      start_level_inputs.files.push_back(f);
+    }
     if (start_level_inputs.size() <= 1) {
       // If only the last file in L0 is marked for compaction, ignore it
       return nullptr;
     }
     inputs.push_back(start_level_inputs);
-#else
-    // Disable due to a known race condition.
-    // TODO: Reenable once the race condition is fixed
-    return nullptr;
-#endif  // ENABLE_SINGLE_LEVEL_DTC
   } else {
     int start_level;
 
@@ -944,8 +947,8 @@ Compaction* UniversalCompactionBuilder::PickDeleteTriggeredCompaction() {
   uint32_t path_id =
       GetPathId(ioptions_, mutable_cf_options_, estimated_total_size);
   return new Compaction(
-      vstorage_, ioptions_, mutable_cf_options_, std::move(inputs),
-      output_level,
+      vstorage_, ioptions_, mutable_cf_options_, mutable_db_options_,
+      std::move(inputs), output_level,
       MaxFileSizeForLevel(mutable_cf_options_, output_level,
                           kCompactionStyleUniversal),
       /* max_grandparent_overlap_bytes */ LLONG_MAX, path_id,
@@ -993,6 +996,9 @@ Compaction* UniversalCompactionBuilder::PickCompactionToOldest(
       comp_reason_print_string = "size amp";
     } else {
       assert(false);
+      comp_reason_print_string = "unknown: ";
+      comp_reason_print_string.append(
+          std::to_string(static_cast<int>(compaction_reason)));
     }
 
     char file_num_buf[256];
@@ -1014,14 +1020,14 @@ Compaction* UniversalCompactionBuilder::PickCompactionToOldest(
   // compaction_options_universal.compression_size_percent,
   // because we always compact all the files, so always compress.
   return new Compaction(
-      vstorage_, ioptions_, mutable_cf_options_, std::move(inputs),
-      output_level,
+      vstorage_, ioptions_, mutable_cf_options_, mutable_db_options_,
+      std::move(inputs), output_level,
       MaxFileSizeForLevel(mutable_cf_options_, output_level,
                           kCompactionStyleUniversal),
       LLONG_MAX, path_id,
-      GetCompressionType(ioptions_, vstorage_, mutable_cf_options_, start_level,
-                         1, true /* enable_compression */),
-      GetCompressionOptions(mutable_cf_options_, vstorage_, start_level,
+      GetCompressionType(ioptions_, vstorage_, mutable_cf_options_,
+                         output_level, 1, true /* enable_compression */),
+      GetCompressionOptions(mutable_cf_options_, vstorage_, output_level,
                             true /* enable_compression */),
       /* max_subcompactions */ 0, /* grandparents */ {}, /* is manual */ false,
       score_, false /* deletion_compaction */, compaction_reason);

db/compaction/compaction_picker_universal.h

@@ -20,7 +20,8 @@ class UniversalCompactionPicker : public CompactionPicker {
       : CompactionPicker(ioptions, icmp) {}
   virtual Compaction* PickCompaction(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, LogBuffer* log_buffer,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      LogBuffer* log_buffer,
       SequenceNumber earliest_memtable_seqno = kMaxSequenceNumber) override;
   virtual int MaxOutputLevel() const override { return NumberLevels() - 1; }
 

db/compaction/sst_partitioner.cc

@@ -0,0 +1,44 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+
+#include "rocksdb/sst_partitioner.h"
+
+#include <algorithm>
+
+namespace ROCKSDB_NAMESPACE {
+
+PartitionerResult SstPartitionerFixedPrefix::ShouldPartition(
+    const PartitionerRequest& request) {
+  Slice last_key_fixed(*request.prev_user_key);
+  if (last_key_fixed.size() > len_) {
+    last_key_fixed.size_ = len_;
+  }
+  Slice current_key_fixed(*request.current_user_key);
+  if (current_key_fixed.size() > len_) {
+    current_key_fixed.size_ = len_;
+  }
+  return last_key_fixed.compare(current_key_fixed) != 0 ? kRequired
+                                                        : kNotRequired;
+}
+
+bool SstPartitionerFixedPrefix::CanDoTrivialMove(
+    const Slice& smallest_user_key, const Slice& largest_user_key) {
+  return ShouldPartition(PartitionerRequest(smallest_user_key, largest_user_key,
+                                            0)) == kNotRequired;
+}
+
+std::unique_ptr<SstPartitioner>
+SstPartitionerFixedPrefixFactory::CreatePartitioner(
+    const SstPartitioner::Context& /* context */) const {
+  return std::unique_ptr<SstPartitioner>(new SstPartitionerFixedPrefix(len_));
+}
+
+std::shared_ptr<SstPartitionerFactory> NewSstPartitionerFixedPrefixFactory(
+    size_t prefix_len) {
+  return std::make_shared<SstPartitionerFixedPrefixFactory>(prefix_len);
+}
+
+}  // namespace ROCKSDB_NAMESPACE

db/comparator_db_test.cc

@@ -13,6 +13,7 @@
 #include "test_util/testutil.h"
 #include "util/hash.h"
 #include "util/kv_map.h"
+#include "util/random.h"
 #include "util/string_util.h"
 #include "utilities/merge_operators.h"
 
@@ -342,12 +343,12 @@ TEST_P(ComparatorDBTest, SimpleSuffixReverseComparator) {
     std::vector<std::string> source_prefixes;
     // Randomly generate 5 prefixes
     for (int i = 0; i < 5; i++) {
-      source_prefixes.push_back(test::RandomHumanReadableString(&rnd, 8));
+      source_prefixes.push_back(rnd.HumanReadableString(8));
     }
     for (int j = 0; j < 20; j++) {
       int prefix_index = rnd.Uniform(static_cast<int>(source_prefixes.size()));
       std::string key = source_prefixes[prefix_index] +
-                        test::RandomHumanReadableString(&rnd, rnd.Uniform(8));
+                        rnd.HumanReadableString(rnd.Uniform(8));
       source_strings.push_back(key);
     }
 

db/corruption_test.cc

@@ -9,13 +9,13 @@
 
 #ifndef ROCKSDB_LITE
 
-#include "rocksdb/db.h"
-
 #include <errno.h>
 #include <fcntl.h>
 #include <sys/stat.h>
 #include <sys/types.h>
+
 #include <cinttypes>
+
 #include "db/db_impl/db_impl.h"
 #include "db/db_test_util.h"
 #include "db/log_format.h"
@@ -24,13 +24,16 @@
 #include "file/filename.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/convenience.h"
+#include "rocksdb/db.h"
 #include "rocksdb/env.h"
 #include "rocksdb/table.h"
 #include "rocksdb/write_batch.h"
 #include "table/block_based/block_based_table_builder.h"
 #include "table/meta_blocks.h"
+#include "table/mock_table.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"
+#include "util/random.h"
 #include "util/string_util.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -53,7 +56,8 @@ class CorruptionTest : public testing::Test {
     options_.wal_recovery_mode = WALRecoveryMode::kTolerateCorruptedTailRecords;
     options_.env = &env_;
     dbname_ = test::PerThreadDBPath("corruption_test");
-    DestroyDB(dbname_, options_);
+    Status s = DestroyDB(dbname_, options_);
+    EXPECT_OK(s);
 
     db_ = nullptr;
     options_.create_if_missing = true;
@@ -66,6 +70,7 @@ class CorruptionTest : public testing::Test {
 
   ~CorruptionTest() override {
     delete db_;
+    db_ = nullptr;
     DestroyDB(dbname_, Options());
   }
 
@@ -101,22 +106,24 @@ class CorruptionTest : public testing::Test {
     ASSERT_OK(::ROCKSDB_NAMESPACE::RepairDB(dbname_, options_));
   }
 
-  void Build(int n, int flush_every = 0) {
+  void Build(int n, int start, int flush_every) {
     std::string key_space, value_space;
     WriteBatch batch;
     for (int i = 0; i < n; i++) {
       if (flush_every != 0 && i != 0 && i % flush_every == 0) {
-        DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-        dbi->TEST_FlushMemTable();
+        DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+        ASSERT_OK(dbi->TEST_FlushMemTable());
       }
       //if ((i % 100) == 0) fprintf(stderr, "@ %d of %d\n", i, n);
-      Slice key = Key(i, &key_space);
+      Slice key = Key(i + start, &key_space);
       batch.Clear();
-      batch.Put(key, Value(i, &value_space));
+      ASSERT_OK(batch.Put(key, Value(i + start, &value_space)));
       ASSERT_OK(db_->Write(WriteOptions(), &batch));
     }
   }
 
+  void Build(int n, int flush_every = 0) { Build(n, 0, flush_every); }
+
   void Check(int min_expected, int max_expected) {
     uint64_t next_expected = 0;
     uint64_t missed = 0;
@@ -131,6 +138,7 @@ class CorruptionTest : public testing::Test {
     // occurred.
     Iterator* iter = db_->NewIterator(ReadOptions(false, true));
     for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+      ASSERT_OK(iter->status());
       uint64_t key;
       Slice in(iter->key());
       if (!ConsumeDecimalNumber(&in, &key) ||
@@ -147,6 +155,7 @@ class CorruptionTest : public testing::Test {
         correct++;
       }
     }
+    iter->status().PermitUncheckedError();
     delete iter;
 
     fprintf(stderr,
@@ -219,11 +228,11 @@ class CorruptionTest : public testing::Test {
       // preserves the implementation that was in place when all of the
       // magic values in this file were picked.
       *storage = std::string(kValueSize, ' ');
-      return Slice(*storage);
     } else {
       Random r(k);
-      return test::RandomString(&r, kValueSize, storage);
+      *storage = r.RandomString(kValueSize);
     }
+    return Slice(*storage);
   }
 };
 
@@ -266,7 +275,7 @@ TEST_F(CorruptionTest, NewFileErrorDuringWrite) {
   bool failed = false;
   for (int i = 0; i < num; i++) {
     WriteBatch batch;
-    batch.Put("a", Value(100, &value_storage));
+    ASSERT_OK(batch.Put("a", Value(100, &value_storage)));
     s = db_->Write(WriteOptions(), &batch);
     if (!s.ok()) {
       failed = true;
@@ -281,10 +290,10 @@ TEST_F(CorruptionTest, NewFileErrorDuringWrite) {
 
 TEST_F(CorruptionTest, TableFile) {
   Build(100);
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
-  dbi->TEST_CompactRange(1, nullptr, nullptr);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr));
+  ASSERT_OK(dbi->TEST_CompactRange(1, nullptr, nullptr));
 
   Corrupt(kTableFile, 100, 1);
   Check(99, 99);
@@ -304,10 +313,10 @@ TEST_F(CorruptionTest, VerifyChecksumReadahead) {
   Reopen(&options);
 
   Build(10000);
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
-  dbi->TEST_CompactRange(1, nullptr, nullptr);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr));
+  ASSERT_OK(dbi->TEST_CompactRange(1, nullptr, nullptr));
 
   senv.count_random_reads_ = true;
   senv.random_read_counter_.Reset();
@@ -351,14 +360,14 @@ TEST_F(CorruptionTest, TableFileIndexData) {
   Reopen(&options);
   // build 2 tables, flush at 5000
   Build(10000, 5000);
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-  dbi->TEST_FlushMemTable();
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
 
   // corrupt an index block of an entire file
   Corrupt(kTableFile, -2000, 500);
   options.paranoid_checks = false;
   Reopen(&options);
-  dbi = reinterpret_cast<DBImpl*>(db_);
+  dbi = static_cast_with_check<DBImpl>(db_);
   // one full file may be readable, since only one was corrupted
   // the other file should be fully non-readable, since index was corrupted
   Check(0, 5000);
@@ -398,9 +407,9 @@ TEST_F(CorruptionTest, SequenceNumberRecovery) {
 
 TEST_F(CorruptionTest, CorruptedDescriptor) {
   ASSERT_OK(db_->Put(WriteOptions(), "foo", "hello"));
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr));
 
   Corrupt(kDescriptorFile, 0, 1000);
   Status s = TryReopen();
@@ -417,10 +426,10 @@ TEST_F(CorruptionTest, CompactionInputError) {
   Options options;
   Reopen(&options);
   Build(10);
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
-  dbi->TEST_CompactRange(1, nullptr, nullptr);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr));
+  ASSERT_OK(dbi->TEST_CompactRange(1, nullptr, nullptr));
   ASSERT_EQ(1, Property("rocksdb.num-files-at-level2"));
 
   Corrupt(kTableFile, 100, 1);
@@ -439,25 +448,25 @@ TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
   options.write_buffer_size = 131072;
   options.max_write_buffer_number = 2;
   Reopen(&options);
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
 
   // Fill levels >= 1
   for (int level = 1; level < dbi->NumberLevels(); level++) {
-    dbi->Put(WriteOptions(), "", "begin");
-    dbi->Put(WriteOptions(), "~", "end");
-    dbi->TEST_FlushMemTable();
+    ASSERT_OK(dbi->Put(WriteOptions(), "", "begin"));
+    ASSERT_OK(dbi->Put(WriteOptions(), "~", "end"));
+    ASSERT_OK(dbi->TEST_FlushMemTable());
     for (int comp_level = 0; comp_level < dbi->NumberLevels() - level;
          ++comp_level) {
-      dbi->TEST_CompactRange(comp_level, nullptr, nullptr);
+      ASSERT_OK(dbi->TEST_CompactRange(comp_level, nullptr, nullptr));
     }
   }
 
   Reopen(&options);
 
-  dbi = reinterpret_cast<DBImpl*>(db_);
+  dbi = static_cast_with_check<DBImpl>(db_);
   Build(10);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_WaitForCompact();
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_WaitForCompact());
   ASSERT_EQ(1, Property("rocksdb.num-files-at-level0"));
 
   CorruptTableFileAtLevel(0, 100, 1);
@@ -481,8 +490,8 @@ TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
 
 TEST_F(CorruptionTest, UnrelatedKeys) {
   Build(10);
-  DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
-  dbi->TEST_FlushMemTable();
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
   Corrupt(kTableFile, 100, 1);
   ASSERT_NOK(dbi->VerifyChecksum());
 
@@ -491,7 +500,7 @@ TEST_F(CorruptionTest, UnrelatedKeys) {
   std::string v;
   ASSERT_OK(db_->Get(ReadOptions(), Key(1000, &tmp1), &v));
   ASSERT_EQ(Value(1000, &tmp2).ToString(), v);
-  dbi->TEST_FlushMemTable();
+  ASSERT_OK(dbi->TEST_FlushMemTable());
   ASSERT_OK(db_->Get(ReadOptions(), Key(1000, &tmp1), &v));
   ASSERT_EQ(Value(1000, &tmp2).ToString(), v);
 }
@@ -532,7 +541,7 @@ TEST_F(CorruptionTest, FileSystemStateCorrupted) {
     Reopen(&options);
     Build(10);
     ASSERT_OK(db_->Flush(FlushOptions()));
-    DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
+    DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
     std::vector<LiveFileMetaData> metadata;
     dbi->GetLiveFilesMetaData(&metadata);
     ASSERT_GT(metadata.size(), size_t(0));
@@ -544,20 +553,277 @@ TEST_F(CorruptionTest, FileSystemStateCorrupted) {
     if (iter == 0) {  // corrupt file size
       std::unique_ptr<WritableFile> file;
       env_.NewWritableFile(filename, &file, EnvOptions());
-      file->Append(Slice("corrupted sst"));
+      ASSERT_OK(file->Append(Slice("corrupted sst")));
       file.reset();
       Status x = TryReopen(&options);
       ASSERT_TRUE(x.IsCorruption());
     } else {  // delete the file
-      env_.DeleteFile(filename);
+      ASSERT_OK(env_.DeleteFile(filename));
       Status x = TryReopen(&options);
       ASSERT_TRUE(x.IsPathNotFound());
     }
 
-    DestroyDB(dbname_, options_);
+    ASSERT_OK(DestroyDB(dbname_, options_));
   }
 }
 
+static const auto& corruption_modes = {
+    mock::MockTableFactory::kCorruptNone, mock::MockTableFactory::kCorruptKey,
+    mock::MockTableFactory::kCorruptValue,
+    mock::MockTableFactory::kCorruptReorderKey};
+
+TEST_F(CorruptionTest, ParanoidFileChecksOnFlush) {
+  Options options;
+  options.check_flush_compaction_key_order = false;
+  options.paranoid_file_checks = true;
+  options.create_if_missing = true;
+  Status s;
+  for (const auto& mode : corruption_modes) {
+    delete db_;
+    db_ = nullptr;
+    s = DestroyDB(dbname_, options);
+    ASSERT_OK(s);
+    std::shared_ptr<mock::MockTableFactory> mock =
+        std::make_shared<mock::MockTableFactory>();
+    options.table_factory = mock;
+    mock->SetCorruptionMode(mode);
+    ASSERT_OK(DB::Open(options, dbname_, &db_));
+    assert(db_ != nullptr);
+    Build(10);
+    s = db_->Flush(FlushOptions());
+    if (mode == mock::MockTableFactory::kCorruptNone) {
+      ASSERT_OK(s);
+    } else {
+      ASSERT_NOK(s);
+    }
+  }
+}
+
+TEST_F(CorruptionTest, ParanoidFileChecksOnCompact) {
+  Options options;
+  options.paranoid_file_checks = true;
+  options.create_if_missing = true;
+  options.check_flush_compaction_key_order = false;
+  Status s;
+  for (const auto& mode : corruption_modes) {
+    delete db_;
+    db_ = nullptr;
+    s = DestroyDB(dbname_, options);
+    std::shared_ptr<mock::MockTableFactory> mock =
+        std::make_shared<mock::MockTableFactory>();
+    options.table_factory = mock;
+    ASSERT_OK(DB::Open(options, dbname_, &db_));
+    assert(db_ != nullptr);
+    Build(100, 2);
+    // ASSERT_OK(db_->Flush(FlushOptions()));
+    DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+    ASSERT_OK(dbi->TEST_FlushMemTable());
+    mock->SetCorruptionMode(mode);
+    s = dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true);
+    if (mode == mock::MockTableFactory::kCorruptNone) {
+      ASSERT_OK(s);
+    } else {
+      ASSERT_NOK(s);
+    }
+  }
+}
+
+TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRangeFirst) {
+  Options options;
+  options.check_flush_compaction_key_order = false;
+  options.paranoid_file_checks = true;
+  options.create_if_missing = true;
+  for (bool do_flush : {true, false}) {
+    delete db_;
+    db_ = nullptr;
+    ASSERT_OK(DestroyDB(dbname_, options));
+    ASSERT_OK(DB::Open(options, dbname_, &db_));
+    std::string start, end;
+    assert(db_ != nullptr);
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(3, &start), Key(7, &end)));
+    auto snap = db_->GetSnapshot();
+    ASSERT_NE(snap, nullptr);
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(8, &start), Key(9, &end)));
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(2, &start), Key(5, &end)));
+    Build(10);
+    if (do_flush) {
+      ASSERT_OK(db_->Flush(FlushOptions()));
+    } else {
+      DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+      ASSERT_OK(dbi->TEST_FlushMemTable());
+      ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true));
+    }
+    db_->ReleaseSnapshot(snap);
+  }
+}
+
+TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRange) {
+  Options options;
+  options.check_flush_compaction_key_order = false;
+  options.paranoid_file_checks = true;
+  options.create_if_missing = true;
+  for (bool do_flush : {true, false}) {
+    delete db_;
+    db_ = nullptr;
+    ASSERT_OK(DestroyDB(dbname_, options));
+    ASSERT_OK(DB::Open(options, dbname_, &db_));
+    assert(db_ != nullptr);
+    Build(10, 0, 0);
+    std::string start, end;
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(5, &start), Key(15, &end)));
+    auto snap = db_->GetSnapshot();
+    ASSERT_NE(snap, nullptr);
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(8, &start), Key(9, &end)));
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(12, &start), Key(17, &end)));
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(2, &start), Key(4, &end)));
+    Build(10, 10, 0);
+    if (do_flush) {
+      ASSERT_OK(db_->Flush(FlushOptions()));
+    } else {
+      DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+      ASSERT_OK(dbi->TEST_FlushMemTable());
+      ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true));
+    }
+    db_->ReleaseSnapshot(snap);
+  }
+}
+
+TEST_F(CorruptionTest, ParanoidFileChecksWithDeleteRangeLast) {
+  Options options;
+  options.check_flush_compaction_key_order = false;
+  options.paranoid_file_checks = true;
+  options.create_if_missing = true;
+  for (bool do_flush : {true, false}) {
+    delete db_;
+    db_ = nullptr;
+    ASSERT_OK(DestroyDB(dbname_, options));
+    ASSERT_OK(DB::Open(options, dbname_, &db_));
+    assert(db_ != nullptr);
+    std::string start, end;
+    Build(10);
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(3, &start), Key(7, &end)));
+    auto snap = db_->GetSnapshot();
+    ASSERT_NE(snap, nullptr);
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(6, &start), Key(8, &end)));
+    ASSERT_OK(db_->DeleteRange(WriteOptions(), db_->DefaultColumnFamily(),
+                               Key(2, &start), Key(5, &end)));
+    if (do_flush) {
+      ASSERT_OK(db_->Flush(FlushOptions()));
+    } else {
+      DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+      ASSERT_OK(dbi->TEST_FlushMemTable());
+      ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true));
+    }
+    db_->ReleaseSnapshot(snap);
+  }
+}
+
+TEST_F(CorruptionTest, LogCorruptionErrorsInCompactionIterator) {
+  Options options;
+  options.create_if_missing = true;
+  options.allow_data_in_errors = true;
+  auto mode = mock::MockTableFactory::kCorruptKey;
+  delete db_;
+  db_ = nullptr;
+  ASSERT_OK(DestroyDB(dbname_, options));
+
+  std::shared_ptr<mock::MockTableFactory> mock =
+      std::make_shared<mock::MockTableFactory>();
+  mock->SetCorruptionMode(mode);
+  options.table_factory = mock;
+
+  ASSERT_OK(DB::Open(options, dbname_, &db_));
+  assert(db_ != nullptr);
+  Build(100, 2);
+
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  Status s = dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true);
+  ASSERT_NOK(s);
+  ASSERT_TRUE(s.IsCorruption());
+}
+
+TEST_F(CorruptionTest, CompactionKeyOrderCheck) {
+  Options options;
+  options.paranoid_file_checks = false;
+  options.create_if_missing = true;
+  options.check_flush_compaction_key_order = false;
+  delete db_;
+  db_ = nullptr;
+  ASSERT_OK(DestroyDB(dbname_, options));
+  std::shared_ptr<mock::MockTableFactory> mock =
+      std::make_shared<mock::MockTableFactory>();
+  options.table_factory = mock;
+  ASSERT_OK(DB::Open(options, dbname_, &db_));
+  assert(db_ != nullptr);
+  mock->SetCorruptionMode(mock::MockTableFactory::kCorruptReorderKey);
+  Build(100, 2);
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+
+  mock->SetCorruptionMode(mock::MockTableFactory::kCorruptNone);
+  ASSERT_OK(db_->SetOptions({{"check_flush_compaction_key_order", "true"}}));
+  ASSERT_NOK(dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true));
+}
+
+TEST_F(CorruptionTest, FlushKeyOrderCheck) {
+  Options options;
+  options.paranoid_file_checks = false;
+  options.create_if_missing = true;
+  ASSERT_OK(db_->SetOptions({{"check_flush_compaction_key_order", "true"}}));
+
+  ASSERT_OK(db_->Put(WriteOptions(), "foo1", "v1"));
+  ASSERT_OK(db_->Put(WriteOptions(), "foo2", "v1"));
+  ASSERT_OK(db_->Put(WriteOptions(), "foo3", "v1"));
+  ASSERT_OK(db_->Put(WriteOptions(), "foo4", "v1"));
+
+  int cnt = 0;
+  // Generate some out of order keys from the memtable
+  SyncPoint::GetInstance()->SetCallBack(
+      "MemTableIterator::Next:0", [&](void* arg) {
+        MemTableRep::Iterator* mem_iter =
+            static_cast<MemTableRep::Iterator*>(arg);
+        if (++cnt == 3) {
+          mem_iter->Prev();
+          mem_iter->Prev();
+        }
+      });
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
+  Status s = static_cast_with_check<DBImpl>(db_)->TEST_FlushMemTable();
+  ASSERT_NOK(s);
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+TEST_F(CorruptionTest, DisableKeyOrderCheck) {
+  Options options;
+  ASSERT_OK(db_->SetOptions({{"check_flush_compaction_key_order", "false"}}));
+  DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
+
+  SyncPoint::GetInstance()->SetCallBack(
+      "OutputValidator::Add:order_check",
+      [&](void* /*arg*/) { ASSERT_TRUE(false); });
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
+  ASSERT_OK(db_->Put(WriteOptions(), "foo1", "v1"));
+  ASSERT_OK(db_->Put(WriteOptions(), "foo3", "v1"));
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(db_->Put(WriteOptions(), "foo2", "v1"));
+  ASSERT_OK(db_->Put(WriteOptions(), "foo4", "v1"));
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr, nullptr, true));
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/cuckoo_table_db_test.cc

@@ -13,6 +13,7 @@
 #include "table/meta_blocks.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"
+#include "util/cast_util.h"
 #include "util/string_util.h"
 
 namespace ROCKSDB_NAMESPACE {
@@ -46,9 +47,7 @@ class CuckooTableDBTest : public testing::Test {
     return options;
   }
 
-  DBImpl* dbfull() {
-    return reinterpret_cast<DBImpl*>(db_);
-  }
+  DBImpl* dbfull() { return static_cast_with_check<DBImpl>(db_); }
 
   // The following util methods are copied from plain_table_db_test.
   void Reopen(Options* options = nullptr) {

db/db_block_cache_test.cc

@@ -7,10 +7,12 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 #include <cstdlib>
+
 #include "cache/lru_cache.h"
 #include "db/db_test_util.h"
 #include "port/stack_trace.h"
 #include "util/compression.h"
+#include "util/random.h"
 
 namespace ROCKSDB_NAMESPACE {
 
@@ -32,7 +34,8 @@ class DBBlockCacheTest : public DBTestBase {
   const size_t kNumBlocks = 10;
   const size_t kValueSize = 100;
 
-  DBBlockCacheTest() : DBTestBase("/db_block_cache_test") {}
+  DBBlockCacheTest()
+      : DBTestBase("/db_block_cache_test", /*env_do_fsync=*/true) {}
 
   BlockBasedTableOptions GetTableOptions() {
     BlockBasedTableOptions table_options;
@@ -47,7 +50,7 @@ class DBBlockCacheTest : public DBTestBase {
     options.avoid_flush_during_recovery = false;
     // options.compression = kNoCompression;
     options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
-    options.table_factory.reset(new BlockBasedTableFactory(table_options));
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
     return options;
   }
 
@@ -155,7 +158,7 @@ TEST_F(DBBlockCacheTest, IteratorBlockCacheUsage) {
 
   std::shared_ptr<Cache> cache = NewLRUCache(0, 0, false);
   table_options.block_cache = cache;
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   Reopen(options);
   RecordCacheCounters(options);
 
@@ -179,7 +182,7 @@ TEST_F(DBBlockCacheTest, TestWithoutCompressedBlockCache) {
 
   std::shared_ptr<Cache> cache = NewLRUCache(0, 0, false);
   table_options.block_cache = cache;
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   Reopen(options);
   RecordCacheCounters(options);
 
@@ -235,7 +238,7 @@ TEST_F(DBBlockCacheTest, TestWithCompressedBlockCache) {
   std::shared_ptr<Cache> compressed_cache = NewLRUCache(1 << 25, 0, false);
   table_options.block_cache = cache;
   table_options.block_cache_compressed = compressed_cache;
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   Reopen(options);
   RecordCacheCounters(options);
 
@@ -296,7 +299,7 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksOfNewTableAddedToCache) {
   BlockBasedTableOptions table_options;
   table_options.cache_index_and_filter_blocks = true;
   table_options.filter_policy.reset(NewBloomFilterPolicy(20));
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   CreateAndReopenWithCF({"pikachu"}, options);
 
   ASSERT_OK(Put(1, "key", "val"));
@@ -352,7 +355,7 @@ TEST_F(DBBlockCacheTest, FillCacheAndIterateDB) {
 
   std::shared_ptr<Cache> cache = NewLRUCache(10, 0, true);
   table_options.block_cache = cache;
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   Reopen(options);
   ASSERT_OK(Put("key1", "val1"));
   ASSERT_OK(Put("key2", "val2"));
@@ -390,7 +393,7 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksStats) {
   std::shared_ptr<Cache> cache = NewLRUCache(co);
   table_options.block_cache = cache;
   table_options.filter_policy.reset(NewBloomFilterPolicy(20, true));
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   CreateAndReopenWithCF({"pikachu"}, options);
 
   ASSERT_OK(Put(1, "longer_key", "val"));
@@ -471,7 +474,7 @@ TEST_F(DBBlockCacheTest, IndexAndFilterBlocksCachePriority) {
     table_options.filter_policy.reset(NewBloomFilterPolicy(20));
     table_options.cache_index_and_filter_blocks_with_high_priority =
         priority == Cache::Priority::HIGH ? true : false;
-    options.table_factory.reset(new BlockBasedTableFactory(table_options));
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
     DestroyAndReopen(options);
 
     MockCache::high_pri_insert_count = 0;
@@ -570,7 +573,7 @@ TEST_F(DBBlockCacheTest, AddRedundantStats) {
     table_options.cache_index_and_filter_blocks = true;
     table_options.block_cache = cache;
     table_options.filter_policy.reset(NewBloomFilterPolicy(50));
-    options.table_factory.reset(new BlockBasedTableFactory(table_options));
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
     DestroyAndReopen(options);
 
     // Create a new table.
@@ -659,7 +662,7 @@ TEST_F(DBBlockCacheTest, ParanoidFileChecks) {
   BlockBasedTableOptions table_options;
   table_options.cache_index_and_filter_blocks = false;
   table_options.filter_policy.reset(NewBloomFilterPolicy(20));
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   CreateAndReopenWithCF({"pikachu"}, options);
 
   ASSERT_OK(Put(1, "1_key", "val"));
@@ -764,7 +767,7 @@ TEST_F(DBBlockCacheTest, CompressedCache) {
     std::string str;
     for (int i = 0; i < num_iter; i++) {
       if (i % 4 == 0) {  // high compression ratio
-        str = RandomString(&rnd, 1000);
+        str = rnd.RandomString(1000);
       }
       values.push_back(str);
       ASSERT_OK(Put(1, Key(i), values[i]));
@@ -843,7 +846,7 @@ TEST_F(DBBlockCacheTest, CacheCompressionDict) {
     BlockBasedTableOptions table_options;
     table_options.cache_index_and_filter_blocks = true;
     table_options.block_cache.reset(new MockCache());
-    options.table_factory.reset(new BlockBasedTableFactory(table_options));
+    options.table_factory.reset(NewBlockBasedTableFactory(table_options));
     DestroyAndReopen(options);
 
     RecordCacheCountersForCompressionDict(options);
@@ -851,7 +854,7 @@ TEST_F(DBBlockCacheTest, CacheCompressionDict) {
     for (int i = 0; i < kNumFiles; ++i) {
       ASSERT_EQ(i, NumTableFilesAtLevel(0, 0));
       for (int j = 0; j < kNumEntriesPerFile; ++j) {
-        std::string value = RandomString(&rnd, kNumBytesPerEntry);
+        std::string value = rnd.RandomString(kNumBytesPerEntry);
         ASSERT_OK(Put(Key(j * kNumFiles + i), value.c_str()));
       }
       ASSERT_OK(Flush());
@@ -885,6 +888,163 @@ TEST_F(DBBlockCacheTest, CacheCompressionDict) {
 
 #endif  // ROCKSDB_LITE
 
+class DBBlockCachePinningTest
+    : public DBTestBase,
+      public testing::WithParamInterface<
+          std::tuple<bool, PinningTier, PinningTier, PinningTier>> {
+ public:
+  DBBlockCachePinningTest()
+      : DBTestBase("/db_block_cache_test", /*env_do_fsync=*/false) {}
+
+  void SetUp() override {
+    partition_index_and_filters_ = std::get<0>(GetParam());
+    top_level_index_pinning_ = std::get<1>(GetParam());
+    partition_pinning_ = std::get<2>(GetParam());
+    unpartitioned_pinning_ = std::get<3>(GetParam());
+  }
+
+  bool partition_index_and_filters_;
+  PinningTier top_level_index_pinning_;
+  PinningTier partition_pinning_;
+  PinningTier unpartitioned_pinning_;
+};
+
+TEST_P(DBBlockCachePinningTest, TwoLevelDB) {
+  // Creates one file in L0 and one file in L1. Both files have enough data that
+  // their index and filter blocks are partitioned. The L1 file will also have
+  // a compression dictionary (those are trained only during compaction), which
+  // must be unpartitioned.
+  const int kKeySize = 32;
+  const int kBlockSize = 128;
+  const int kNumBlocksPerFile = 128;
+  const int kNumKeysPerFile = kBlockSize * kNumBlocksPerFile / kKeySize;
+
+  Options options = CurrentOptions();
+  // `kNoCompression` makes the unit test more portable. But it relies on the
+  // current behavior of persisting/accessing dictionary even when there's no
+  // (de)compression happening, which seems fairly likely to change over time.
+  options.compression = kNoCompression;
+  options.compression_opts.max_dict_bytes = 4 << 10;
+  options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
+  BlockBasedTableOptions table_options;
+  table_options.block_cache = NewLRUCache(1 << 20 /* capacity */);
+  table_options.block_size = kBlockSize;
+  table_options.metadata_block_size = kBlockSize;
+  table_options.cache_index_and_filter_blocks = true;
+  table_options.metadata_cache_options.top_level_index_pinning =
+      top_level_index_pinning_;
+  table_options.metadata_cache_options.partition_pinning = partition_pinning_;
+  table_options.metadata_cache_options.unpartitioned_pinning =
+      unpartitioned_pinning_;
+  table_options.filter_policy.reset(
+      NewBloomFilterPolicy(10 /* bits_per_key */));
+  if (partition_index_and_filters_) {
+    table_options.index_type =
+        BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch;
+    table_options.partition_filters = true;
+  }
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
+  Reopen(options);
+
+  Random rnd(301);
+  for (int i = 0; i < 2; ++i) {
+    for (int j = 0; j < kNumKeysPerFile; ++j) {
+      ASSERT_OK(Put(Key(i * kNumKeysPerFile + j), rnd.RandomString(kKeySize)));
+    }
+    ASSERT_OK(Flush());
+    if (i == 0) {
+      // Prevent trivial move so file will be rewritten with dictionary and
+      // reopened with L1's pinning settings.
+      CompactRangeOptions cro;
+      cro.bottommost_level_compaction = BottommostLevelCompaction::kForce;
+      ASSERT_OK(db_->CompactRange(cro, nullptr, nullptr));
+    }
+  }
+
+  // Clear all unpinned blocks so unpinned blocks will show up as cache misses
+  // when reading a key from a file.
+  table_options.block_cache->EraseUnRefEntries();
+
+  // Get base cache values
+  uint64_t filter_misses = TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS);
+  uint64_t index_misses = TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS);
+  uint64_t compression_dict_misses =
+      TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_MISS);
+
+  // Read a key from the L0 file
+  Get(Key(kNumKeysPerFile));
+  uint64_t expected_filter_misses = filter_misses;
+  uint64_t expected_index_misses = index_misses;
+  uint64_t expected_compression_dict_misses = compression_dict_misses;
+  if (partition_index_and_filters_) {
+    if (top_level_index_pinning_ == PinningTier::kNone) {
+      ++expected_filter_misses;
+      ++expected_index_misses;
+    }
+    if (partition_pinning_ == PinningTier::kNone) {
+      ++expected_filter_misses;
+      ++expected_index_misses;
+    }
+  } else {
+    if (unpartitioned_pinning_ == PinningTier::kNone) {
+      ++expected_filter_misses;
+      ++expected_index_misses;
+    }
+  }
+  ASSERT_EQ(expected_filter_misses,
+            TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
+  ASSERT_EQ(expected_index_misses,
+            TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
+  ASSERT_EQ(expected_compression_dict_misses,
+            TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_MISS));
+
+  // Clear all unpinned blocks so unpinned blocks will show up as cache misses
+  // when reading a key from a file.
+  table_options.block_cache->EraseUnRefEntries();
+
+  // Read a key from the L1 file
+  Get(Key(0));
+  if (partition_index_and_filters_) {
+    if (top_level_index_pinning_ == PinningTier::kNone ||
+        top_level_index_pinning_ == PinningTier::kFlushedAndSimilar) {
+      ++expected_filter_misses;
+      ++expected_index_misses;
+    }
+    if (partition_pinning_ == PinningTier::kNone ||
+        partition_pinning_ == PinningTier::kFlushedAndSimilar) {
+      ++expected_filter_misses;
+      ++expected_index_misses;
+    }
+  } else {
+    if (unpartitioned_pinning_ == PinningTier::kNone ||
+        unpartitioned_pinning_ == PinningTier::kFlushedAndSimilar) {
+      ++expected_filter_misses;
+      ++expected_index_misses;
+    }
+  }
+  if (unpartitioned_pinning_ == PinningTier::kNone ||
+      unpartitioned_pinning_ == PinningTier::kFlushedAndSimilar) {
+    ++expected_compression_dict_misses;
+  }
+  ASSERT_EQ(expected_filter_misses,
+            TestGetTickerCount(options, BLOCK_CACHE_FILTER_MISS));
+  ASSERT_EQ(expected_index_misses,
+            TestGetTickerCount(options, BLOCK_CACHE_INDEX_MISS));
+  ASSERT_EQ(expected_compression_dict_misses,
+            TestGetTickerCount(options, BLOCK_CACHE_COMPRESSION_DICT_MISS));
+}
+
+INSTANTIATE_TEST_CASE_P(
+    DBBlockCachePinningTest, DBBlockCachePinningTest,
+    ::testing::Combine(
+        ::testing::Bool(),
+        ::testing::Values(PinningTier::kNone, PinningTier::kFlushedAndSimilar,
+                          PinningTier::kAll),
+        ::testing::Values(PinningTier::kNone, PinningTier::kFlushedAndSimilar,
+                          PinningTier::kAll),
+        ::testing::Values(PinningTier::kNone, PinningTier::kFlushedAndSimilar,
+                          PinningTier::kAll)));
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/db_bloom_filter_test.cc

@@ -23,7 +23,8 @@ using BFP = BloomFilterPolicy;
 
 class DBBloomFilterTest : public DBTestBase {
  public:
-  DBBloomFilterTest() : DBTestBase("/db_bloom_filter_test") {}
+  DBBloomFilterTest()
+      : DBTestBase("/db_bloom_filter_test", /*env_do_fsync=*/true) {}
 };
 
 class DBBloomFilterTestWithParam : public DBTestBase,
@@ -36,7 +37,8 @@ class DBBloomFilterTestWithParam : public DBTestBase,
   uint32_t format_version_;
 
  public:
-  DBBloomFilterTestWithParam() : DBTestBase("/db_bloom_filter_tests") {}
+  DBBloomFilterTestWithParam()
+      : DBTestBase("/db_bloom_filter_tests", /*env_do_fsync=*/true) {}
 
   ~DBBloomFilterTestWithParam() override {}
 
@@ -81,13 +83,16 @@ TEST_P(DBBloomFilterTestDefFormatVersion, KeyMayExist) {
     options_override.partition_filters = partition_filters_;
     options_override.metadata_block_size = 32;
     Options options = CurrentOptions(options_override);
-    if (partition_filters_ &&
-        static_cast<BlockBasedTableOptions*>(
-            options.table_factory->GetOptions())
-                ->index_type != BlockBasedTableOptions::kTwoLevelIndexSearch) {
-      // In the current implementation partitioned filters depend on partitioned
-      // indexes
-      continue;
+    if (partition_filters_) {
+      auto* table_options =
+          options.table_factory->GetOptions<BlockBasedTableOptions>();
+      if (table_options != nullptr &&
+          table_options->index_type !=
+              BlockBasedTableOptions::kTwoLevelIndexSearch) {
+        // In the current implementation partitioned filters depend on
+        // partitioned indexes
+        continue;
+      }
     }
     options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
     CreateAndReopenWithCF({"pikachu"}, options);
@@ -1039,7 +1044,7 @@ class DBBloomFilterTestVaryPrefixAndFormatVer
 
  public:
   DBBloomFilterTestVaryPrefixAndFormatVer()
-      : DBTestBase("/db_bloom_filter_tests") {}
+      : DBTestBase("/db_bloom_filter_tests", /*env_do_fsync=*/true) {}
 
   ~DBBloomFilterTestVaryPrefixAndFormatVer() override {}
 

db/db_compaction_filter_test.cc

@@ -21,7 +21,8 @@ static std::string NEW_VALUE = "NewValue";
 
 class DBTestCompactionFilter : public DBTestBase {
  public:
-  DBTestCompactionFilter() : DBTestBase("/db_compaction_filter_test") {}
+  DBTestCompactionFilter()
+      : DBTestBase("/db_compaction_filter_test", /*env_do_fsync=*/true) {}
 };
 
 // Param variant of DBTestBase::ChangeCompactOptions
@@ -126,22 +127,6 @@ class SkipEvenFilter : public CompactionFilter {
   const char* Name() const override { return "DeleteFilter"; }
 };
 
-class DelayFilter : public CompactionFilter {
- public:
-  explicit DelayFilter(DBTestBase* d) : db_test(d) {}
-  bool Filter(int /*level*/, const Slice& /*key*/, const Slice& /*value*/,
-              std::string* /*new_value*/,
-              bool* /*value_changed*/) const override {
-    db_test->env_->addon_time_.fetch_add(1000);
-    return true;
-  }
-
-  const char* Name() const override { return "DelayFilter"; }
-
- private:
-  DBTestBase* db_test;
-};
-
 class ConditionalFilter : public CompactionFilter {
  public:
   explicit ConditionalFilter(const std::string* filtered_value)
@@ -248,20 +233,6 @@ class SkipEvenFilterFactory : public CompactionFilterFactory {
   const char* Name() const override { return "SkipEvenFilterFactory"; }
 };
 
-class DelayFilterFactory : public CompactionFilterFactory {
- public:
-  explicit DelayFilterFactory(DBTestBase* d) : db_test(d) {}
-  std::unique_ptr<CompactionFilter> CreateCompactionFilter(
-      const CompactionFilter::Context& /*context*/) override {
-    return std::unique_ptr<CompactionFilter>(new DelayFilter(db_test));
-  }
-
-  const char* Name() const override { return "DelayFilterFactory"; }
-
- private:
-  DBTestBase* db_test;
-};
-
 class ConditionalFilterFactory : public CompactionFilterFactory {
  public:
   explicit ConditionalFilterFactory(const Slice& filtered_value)
@@ -336,13 +307,14 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
     InternalKeyComparator icmp(options.comparator);
     ReadRangeDelAggregator range_del_agg(&icmp,
                                          kMaxSequenceNumber /* upper_bound */);
+    ReadOptions read_options;
     ScopedArenaIterator iter(dbfull()->NewInternalIterator(
-        &arena, &range_del_agg, kMaxSequenceNumber, handles_[1]));
+        read_options, &arena, &range_del_agg, kMaxSequenceNumber, handles_[1]));
     iter->SeekToFirst();
     ASSERT_OK(iter->status());
     while (iter->Valid()) {
       ParsedInternalKey ikey(Slice(), 0, kTypeValue);
-      ASSERT_EQ(ParseInternalKey(iter->key(), &ikey), true);
+      ASSERT_OK(ParseInternalKey(iter->key(), &ikey));
       total++;
       if (ikey.sequence != 0) {
         count++;
@@ -426,13 +398,14 @@ TEST_F(DBTestCompactionFilter, CompactionFilter) {
     InternalKeyComparator icmp(options.comparator);
     ReadRangeDelAggregator range_del_agg(&icmp,
                                          kMaxSequenceNumber /* upper_bound */);
+    ReadOptions read_options;
     ScopedArenaIterator iter(dbfull()->NewInternalIterator(
-        &arena, &range_del_agg, kMaxSequenceNumber, handles_[1]));
+        read_options, &arena, &range_del_agg, kMaxSequenceNumber, handles_[1]));
     iter->SeekToFirst();
     ASSERT_OK(iter->status());
     while (iter->Valid()) {
       ParsedInternalKey ikey(Slice(), 0, kTypeValue);
-      ASSERT_EQ(ParseInternalKey(iter->key(), &ikey), true);
+      ASSERT_OK(ParseInternalKey(iter->key(), &ikey));
       ASSERT_NE(ikey.sequence, (unsigned)0);
       count++;
       iter->Next();
@@ -644,13 +617,14 @@ TEST_F(DBTestCompactionFilter, CompactionFilterContextManual) {
     InternalKeyComparator icmp(options.comparator);
     ReadRangeDelAggregator range_del_agg(&icmp,
                                          kMaxSequenceNumber /* snapshots */);
+    ReadOptions read_options;
     ScopedArenaIterator iter(dbfull()->NewInternalIterator(
-        &arena, &range_del_agg, kMaxSequenceNumber));
+        read_options, &arena, &range_del_agg, kMaxSequenceNumber));
     iter->SeekToFirst();
     ASSERT_OK(iter->status());
     while (iter->Valid()) {
       ParsedInternalKey ikey(Slice(), 0, kTypeValue);
-      ASSERT_EQ(ParseInternalKey(iter->key(), &ikey), true);
+      ASSERT_OK(ParseInternalKey(iter->key(), &ikey));
       total++;
       if (ikey.sequence != 0) {
         count++;

db/db_dynamic_level_test.cc

@@ -15,11 +15,13 @@
 #include "db/db_test_util.h"
 #include "port/port.h"
 #include "port/stack_trace.h"
+#include "util/random.h"
 
 namespace ROCKSDB_NAMESPACE {
 class DBTestDynamicLevel : public DBTestBase {
  public:
-  DBTestDynamicLevel() : DBTestBase("/db_dynamic_level_test") {}
+  DBTestDynamicLevel()
+      : DBTestBase("/db_dynamic_level_test", /*env_do_fsync=*/true) {}
 };
 
 TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase) {
@@ -80,9 +82,9 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase) {
 
       for (int i = 0; i < kNKeys; i++) {
         int key = keys[i];
-        ASSERT_OK(Put(Key(kNKeys + key), RandomString(&rnd, 102)));
-        ASSERT_OK(Put(Key(key), RandomString(&rnd, 102)));
-        ASSERT_OK(Put(Key(kNKeys * 2 + key), RandomString(&rnd, 102)));
+        ASSERT_OK(Put(Key(kNKeys + key), rnd.RandomString(102)));
+        ASSERT_OK(Put(Key(key), rnd.RandomString(102)));
+        ASSERT_OK(Put(Key(kNKeys * 2 + key), rnd.RandomString(102)));
         ASSERT_OK(Delete(Key(kNKeys + keys[i / 10])));
         env_->SleepForMicroseconds(5000);
       }
@@ -158,7 +160,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
   // Put about 28K to L0
   for (int i = 0; i < 70; i++) {
     ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 380)));
+                  rnd.RandomString(380)));
   }
   ASSERT_OK(dbfull()->SetOptions({
       {"disable_auto_compactions", "false"},
@@ -175,7 +177,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
   }));
   for (int i = 0; i < 70; i++) {
     ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 380)));
+                  rnd.RandomString(380)));
   }
 
   ASSERT_OK(dbfull()->SetOptions({
@@ -197,7 +199,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
   // Write about 40K more
   for (int i = 0; i < 100; i++) {
     ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 380)));
+                  rnd.RandomString(380)));
   }
   ASSERT_OK(dbfull()->SetOptions({
       {"disable_auto_compactions", "false"},
@@ -216,7 +218,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
   // Each file is about 11KB, with 9KB of data.
   for (int i = 0; i < 1300; i++) {
     ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 380)));
+                  rnd.RandomString(380)));
   }
 
   // Make sure that the compaction starts before the last bit of data is
@@ -257,7 +259,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
   TEST_SYNC_POINT("DynamicLevelMaxBytesBase2:1");
   for (int i = 0; i < 2; i++) {
     ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 380)));
+                  rnd.RandomString(380)));
   }
   TEST_SYNC_POINT("DynamicLevelMaxBytesBase2:2");
 
@@ -310,15 +312,15 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesCompactRange) {
 
   // Put about 7K to L0
   for (int i = 0; i < 140; i++) {
-    ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 80)));
+    ASSERT_OK(
+        Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))), rnd.RandomString(80)));
   }
   Flush();
   dbfull()->TEST_WaitForCompact();
   if (NumTableFilesAtLevel(0) == 0) {
     // Make sure level 0 is not empty
-    ASSERT_OK(Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))),
-                  RandomString(&rnd, 80)));
+    ASSERT_OK(
+        Put(Key(static_cast<int>(rnd.Uniform(kMaxKey))), rnd.RandomString(80)));
     Flush();
   }
 
@@ -382,7 +384,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBaseInc) {
   const int total_keys = 3000;
   const int random_part_size = 100;
   for (int i = 0; i < total_keys; i++) {
-    std::string value = RandomString(&rnd, random_part_size);
+    std::string value = rnd.RandomString(random_part_size);
     PutFixed32(&value, static_cast<uint32_t>(i));
     ASSERT_OK(Put(Key(i), value));
   }
@@ -441,8 +443,8 @@ TEST_F(DBTestDynamicLevel, DISABLED_MigrateToDynamicLevelMaxBytesBase) {
 
   int total_keys = 1000;
   for (int i = 0; i < total_keys; i++) {
-    ASSERT_OK(Put(Key(i), RandomString(&rnd, 102)));
-    ASSERT_OK(Put(Key(kMaxKey + i), RandomString(&rnd, 102)));
+    ASSERT_OK(Put(Key(i), rnd.RandomString(102)));
+    ASSERT_OK(Put(Key(kMaxKey + i), rnd.RandomString(102)));
     ASSERT_OK(Delete(Key(i / 10)));
   }
   verify_func(total_keys, false);
@@ -475,8 +477,8 @@ TEST_F(DBTestDynamicLevel, DISABLED_MigrateToDynamicLevelMaxBytesBase) {
 
   int total_keys2 = 2000;
   for (int i = total_keys; i < total_keys2; i++) {
-    ASSERT_OK(Put(Key(i), RandomString(&rnd, 102)));
-    ASSERT_OK(Put(Key(kMaxKey + i), RandomString(&rnd, 102)));
+    ASSERT_OK(Put(Key(i), rnd.RandomString(102)));
+    ASSERT_OK(Put(Key(kMaxKey + i), rnd.RandomString(102)));
     ASSERT_OK(Delete(Key(i / 10)));
   }
 

db/db_encryption_test.cc

@@ -16,7 +16,8 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBEncryptionTest : public DBTestBase {
  public:
-  DBEncryptionTest() : DBTestBase("/db_encryption_test") {}
+  DBEncryptionTest()
+      : DBTestBase("/db_encryption_test", /*env_do_fsync=*/true) {}
 };
 
 #ifndef ROCKSDB_LITE

db/db_filesnapshot.cc

@@ -23,57 +23,6 @@
 
 namespace ROCKSDB_NAMESPACE {
 
-Status DBImpl::DisableFileDeletions() {
-  InstrumentedMutexLock l(&mutex_);
-  ++disable_delete_obsolete_files_;
-  if (disable_delete_obsolete_files_ == 1) {
-    ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Disabled");
-  } else {
-    ROCKS_LOG_WARN(immutable_db_options_.info_log,
-                   "File Deletions Disabled, but already disabled. Counter: %d",
-                   disable_delete_obsolete_files_);
-  }
-  return Status::OK();
-}
-
-Status DBImpl::EnableFileDeletions(bool force) {
-  // Job id == 0 means that this is not our background process, but rather
-  // user thread
-  JobContext job_context(0);
-  bool file_deletion_enabled = false;
-  {
-    InstrumentedMutexLock l(&mutex_);
-    if (force) {
-      // if force, we need to enable file deletions right away
-      disable_delete_obsolete_files_ = 0;
-    } else if (disable_delete_obsolete_files_ > 0) {
-      --disable_delete_obsolete_files_;
-    }
-    if (disable_delete_obsolete_files_ == 0)  {
-      file_deletion_enabled = true;
-      FindObsoleteFiles(&job_context, true);
-      bg_cv_.SignalAll();
-    }
-  }
-  if (file_deletion_enabled) {
-    ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Enabled");
-    if (job_context.HaveSomethingToDelete()) {
-      PurgeObsoleteFiles(job_context);
-    }
-  } else {
-    ROCKS_LOG_WARN(immutable_db_options_.info_log,
-                   "File Deletions Enable, but not really enabled. Counter: %d",
-                   disable_delete_obsolete_files_);
-  }
-  job_context.Clean();
-  LogFlush(immutable_db_options_.info_log);
-  return Status::OK();
-}
-
-int DBImpl::IsFileDeletionsEnabled() const {
-  return !disable_delete_obsolete_files_;
-}
-
 Status DBImpl::GetLiveFiles(std::vector<std::string>& ret,
                             uint64_t* manifest_file_size,
                             bool flush_memtable) {
@@ -167,7 +116,7 @@ Status DBImpl::GetSortedWalFiles(VectorLogPtr& files) {
     // long as deletions are disabled (so the below loop must terminate).
     InstrumentedMutexLock l(&mutex_);
     while (disable_delete_obsolete_files_ > 0 &&
-           pending_purge_obsolete_files_ > 0) {
+           (pending_purge_obsolete_files_ > 0 || bg_purge_scheduled_ > 0)) {
       bg_cv_.Wait();
     }
   }

db/db_flush_test.cc

@@ -11,18 +11,19 @@
 
 #include "db/db_impl/db_impl.h"
 #include "db/db_test_util.h"
+#include "file/filename.h"
 #include "port/port.h"
 #include "port/stack_trace.h"
-#include "test_util/fault_injection_test_env.h"
 #include "test_util/sync_point.h"
 #include "util/cast_util.h"
 #include "util/mutexlock.h"
+#include "utilities/fault_injection_env.h"
 
 namespace ROCKSDB_NAMESPACE {
 
 class DBFlushTest : public DBTestBase {
  public:
-  DBFlushTest() : DBTestBase("/db_flush_test") {}
+  DBFlushTest() : DBTestBase("/db_flush_test", /*env_do_fsync=*/true) {}
 };
 
 class DBFlushDirectIOTest : public DBFlushTest,
@@ -62,7 +63,7 @@ TEST_F(DBFlushTest, FlushWhileWritingManifest) {
   ASSERT_OK(Put("bar", "v"));
   ASSERT_OK(dbfull()->Flush(no_wait));
   // If the issue is hit we will wait here forever.
-  dbfull()->TEST_WaitForFlushMemTable();
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
 #ifndef ROCKSDB_LITE
   ASSERT_EQ(2, TotalTableFiles());
 #endif  // ROCKSDB_LITE
@@ -87,9 +88,9 @@ TEST_F(DBFlushTest, SyncFail) {
   SyncPoint::GetInstance()->EnableProcessing();
 
   CreateAndReopenWithCF({"pikachu"}, options);
-  Put("key", "value");
+  ASSERT_OK(Put("key", "value"));
   auto* cfd =
-      reinterpret_cast<ColumnFamilyHandleImpl*>(db_->DefaultColumnFamily())
+      static_cast_with_check<ColumnFamilyHandleImpl>(db_->DefaultColumnFamily())
           ->cfd();
   FlushOptions flush_options;
   flush_options.wait = false;
@@ -106,7 +107,8 @@ TEST_F(DBFlushTest, SyncFail) {
   TEST_SYNC_POINT("DBFlushTest::SyncFail:2");
   fault_injection_env->SetFilesystemActive(true);
   // Now the background job will do the flush; wait for it.
-  dbfull()->TEST_WaitForFlushMemTable();
+  // Returns the IO error happend during flush.
+  ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable());
 #ifndef ROCKSDB_LITE
   ASSERT_EQ("", FilesPerLevel());  // flush failed.
 #endif                             // ROCKSDB_LITE
@@ -125,7 +127,7 @@ TEST_F(DBFlushTest, SyncSkip) {
   SyncPoint::GetInstance()->EnableProcessing();
 
   Reopen(options);
-  Put("key", "value");
+  ASSERT_OK(Put("key", "value"));
 
   FlushOptions flush_options;
   flush_options.wait = false;
@@ -135,7 +137,7 @@ TEST_F(DBFlushTest, SyncSkip) {
   TEST_SYNC_POINT("DBFlushTest::SyncSkip:2");
 
   // Now the background job will do the flush; wait for it.
-  dbfull()->TEST_WaitForFlushMemTable();
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
 
   Destroy(options);
 }
@@ -170,9 +172,9 @@ TEST_F(DBFlushTest, FlushInLowPriThreadPool) {
   ASSERT_OK(Put("key", "val"));
   for (int i = 0; i < 4; ++i) {
     ASSERT_OK(Put("key", "val"));
-    dbfull()->TEST_WaitForFlushMemTable();
+    ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
   }
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(4, num_flushes);
   ASSERT_EQ(1, num_compactions);
 }
@@ -305,7 +307,8 @@ TEST_F(DBFlushTest, ManualFlushFailsInReadOnlyMode) {
   // mode.
   fault_injection_env->SetFilesystemActive(false);
   ASSERT_OK(db_->ContinueBackgroundWork());
-  dbfull()->TEST_WaitForFlushMemTable();
+  // We ingested the error to env, so the returned status is not OK.
+  ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable());
 #ifndef ROCKSDB_LITE
   uint64_t num_bg_errors;
   ASSERT_TRUE(db_->GetIntProperty(DB::Properties::kBackgroundErrors,
@@ -379,9 +382,9 @@ TEST_F(DBFlushTest, FireOnFlushCompletedAfterCommittedResult) {
       DBImpl* db_impl = static_cast_with_check<DBImpl>(db);
       InstrumentedMutex* mutex = db_impl->mutex();
       mutex->Lock();
-      auto* cfd =
-          reinterpret_cast<ColumnFamilyHandleImpl*>(db->DefaultColumnFamily())
-              ->cfd();
+      auto* cfd = static_cast_with_check<ColumnFamilyHandleImpl>(
+                      db->DefaultColumnFamily())
+                      ->cfd();
       ASSERT_LT(seq, cfd->imm()->current()->GetEarliestSequenceNumber());
       mutex->Unlock();
     }
@@ -443,6 +446,169 @@ TEST_F(DBFlushTest, FireOnFlushCompletedAfterCommittedResult) {
 }
 #endif  // !ROCKSDB_LITE
 
+TEST_F(DBFlushTest, FlushWithBlob) {
+  constexpr uint64_t min_blob_size = 10;
+
+  Options options;
+  options.enable_blob_files = true;
+  options.min_blob_size = min_blob_size;
+  options.disable_auto_compactions = true;
+
+  Reopen(options);
+
+  constexpr char short_value[] = "short";
+  static_assert(sizeof(short_value) - 1 < min_blob_size,
+                "short_value too long");
+
+  constexpr char long_value[] = "long_value";
+  static_assert(sizeof(long_value) - 1 >= min_blob_size,
+                "long_value too short");
+
+  ASSERT_OK(Put("key1", short_value));
+  ASSERT_OK(Put("key2", long_value));
+
+  ASSERT_OK(Flush());
+
+  ASSERT_EQ(Get("key1"), short_value);
+
+  // TODO: enable once Get support is implemented for blobs
+  // ASSERT_EQ(Get("key2"), long_value);
+
+  VersionSet* const versions = dbfull()->TEST_GetVersionSet();
+  assert(versions);
+
+  ColumnFamilyData* const cfd = versions->GetColumnFamilySet()->GetDefault();
+  assert(cfd);
+
+  Version* const current = cfd->current();
+  assert(current);
+
+  const VersionStorageInfo* const storage_info = current->storage_info();
+  assert(storage_info);
+
+  const auto& l0_files = storage_info->LevelFiles(0);
+  ASSERT_EQ(l0_files.size(), 1);
+
+  const FileMetaData* const table_file = l0_files[0];
+  assert(table_file);
+
+  const auto& blob_files = storage_info->GetBlobFiles();
+  ASSERT_EQ(blob_files.size(), 1);
+
+  const auto& blob_file = blob_files.begin()->second;
+  assert(blob_file);
+
+  ASSERT_EQ(table_file->smallest.user_key(), "key1");
+  ASSERT_EQ(table_file->largest.user_key(), "key2");
+  ASSERT_EQ(table_file->fd.smallest_seqno, 1);
+  ASSERT_EQ(table_file->fd.largest_seqno, 2);
+  ASSERT_EQ(table_file->oldest_blob_file_number,
+            blob_file->GetBlobFileNumber());
+
+  ASSERT_EQ(blob_file->GetTotalBlobCount(), 1);
+
+#ifndef ROCKSDB_LITE
+  const InternalStats* const internal_stats = cfd->internal_stats();
+  assert(internal_stats);
+
+  const uint64_t expected_bytes =
+      table_file->fd.GetFileSize() + blob_file->GetTotalBlobBytes();
+
+  const auto& compaction_stats = internal_stats->TEST_GetCompactionStats();
+  ASSERT_FALSE(compaction_stats.empty());
+  ASSERT_EQ(compaction_stats[0].bytes_written, expected_bytes);
+  ASSERT_EQ(compaction_stats[0].num_output_files, 2);
+
+  const uint64_t* const cf_stats_value = internal_stats->TEST_GetCFStatsValue();
+  ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED], expected_bytes);
+#endif  // ROCKSDB_LITE
+}
+
+class DBFlushTestBlobError : public DBFlushTest,
+                             public testing::WithParamInterface<std::string> {
+ public:
+  DBFlushTestBlobError() : fault_injection_env_(env_) {}
+  ~DBFlushTestBlobError() { Close(); }
+
+  FaultInjectionTestEnv fault_injection_env_;
+};
+
+INSTANTIATE_TEST_CASE_P(DBFlushTestBlobError, DBFlushTestBlobError,
+                        ::testing::ValuesIn(std::vector<std::string>{
+                            "BlobFileBuilder::WriteBlobToFile:AddRecord",
+                            "BlobFileBuilder::WriteBlobToFile:AppendFooter"}));
+
+TEST_P(DBFlushTestBlobError, FlushError) {
+  Options options;
+  options.enable_blob_files = true;
+  options.disable_auto_compactions = true;
+  options.env = &fault_injection_env_;
+
+  Reopen(options);
+
+  ASSERT_OK(Put("key", "blob"));
+
+  SyncPoint::GetInstance()->SetCallBack(GetParam(), [this](void* /* arg */) {
+    fault_injection_env_.SetFilesystemActive(false, Status::IOError());
+  });
+  SyncPoint::GetInstance()->SetCallBack(
+      "BuildTable:BeforeFinishBuildTable", [this](void* /* arg */) {
+        fault_injection_env_.SetFilesystemActive(true);
+      });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_NOK(Flush());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+
+  VersionSet* const versions = dbfull()->TEST_GetVersionSet();
+  assert(versions);
+
+  ColumnFamilyData* const cfd = versions->GetColumnFamilySet()->GetDefault();
+  assert(cfd);
+
+  Version* const current = cfd->current();
+  assert(current);
+
+  const VersionStorageInfo* const storage_info = current->storage_info();
+  assert(storage_info);
+
+  const auto& l0_files = storage_info->LevelFiles(0);
+  ASSERT_TRUE(l0_files.empty());
+
+  const auto& blob_files = storage_info->GetBlobFiles();
+  ASSERT_TRUE(blob_files.empty());
+
+  // Make sure the files generated by the failed job have been deleted
+  std::vector<std::string> files;
+  ASSERT_OK(env_->GetChildren(dbname_, &files));
+  for (const auto& file : files) {
+    uint64_t number = 0;
+    FileType type = kTableFile;
+
+    if (!ParseFileName(file, &number, &type)) {
+      continue;
+    }
+
+    ASSERT_NE(type, kTableFile);
+    ASSERT_NE(type, kBlobFile);
+  }
+
+#ifndef ROCKSDB_LITE
+  const InternalStats* const internal_stats = cfd->internal_stats();
+  assert(internal_stats);
+
+  const auto& compaction_stats = internal_stats->TEST_GetCompactionStats();
+  ASSERT_FALSE(compaction_stats.empty());
+  ASSERT_EQ(compaction_stats[0].bytes_written, 0);
+  ASSERT_EQ(compaction_stats[0].num_output_files, 0);
+
+  const uint64_t* const cf_stats_value = internal_stats->TEST_GetCFStatsValue();
+  ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED], 0);
+#endif  // ROCKSDB_LITE
+}
+
 TEST_P(DBAtomicFlushTest, ManualAtomicFlush) {
   Options options = CurrentOptions();
   options.create_if_missing = true;
@@ -549,7 +715,8 @@ TEST_P(DBAtomicFlushTest, AtomicFlushRollbackSomeJobs) {
   fault_injection_env->SetFilesystemActive(false);
   TEST_SYNC_POINT("DBAtomicFlushTest::AtomicFlushRollbackSomeJobs:2");
   for (auto* cfh : handles_) {
-    dbfull()->TEST_WaitForFlushMemTable(cfh);
+    // Returns the IO error happend during flush.
+    ASSERT_NOK(dbfull()->TEST_WaitForFlushMemTable(cfh));
   }
   for (size_t i = 0; i != num_cfs; ++i) {
     auto cfh = static_cast<ColumnFamilyHandleImpl*>(handles_[i]);

db/db_impl/db_impl.h

@@ -55,6 +55,7 @@
 #include "rocksdb/write_buffer_manager.h"
 #include "table/scoped_arena_iterator.h"
 #include "trace_replay/block_cache_tracer.h"
+#include "trace_replay/io_tracer.h"
 #include "trace_replay/trace_replay.h"
 #include "util/autovector.h"
 #include "util/hash.h"
@@ -69,6 +70,10 @@ class ArenaWrappedDBIter;
 class InMemoryStatsHistoryIterator;
 class MemTable;
 class PersistentStatsHistoryIterator;
+class PeriodicWorkScheduler;
+#ifndef NDEBUG
+class PeriodicWorkTestScheduler;
+#endif  // !NDEBUG
 class TableCache;
 class TaskLimiterToken;
 class Version;
@@ -358,6 +363,12 @@ class DBImpl : public DB {
 
   virtual Status Close() override;
 
+  virtual Status DisableFileDeletions() override;
+
+  virtual Status EnableFileDeletions(bool force) override;
+
+  virtual bool IsFileDeletionsEnabled() const;
+
   Status GetStatsHistory(
       uint64_t start_time, uint64_t end_time,
       std::unique_ptr<StatsHistoryIterator>* stats_iterator) override;
@@ -365,9 +376,6 @@ class DBImpl : public DB {
 #ifndef ROCKSDB_LITE
   using DB::ResetStats;
   virtual Status ResetStats() override;
-  virtual Status DisableFileDeletions() override;
-  virtual Status EnableFileDeletions(bool force) override;
-  virtual int IsFileDeletionsEnabled() const;
   // All the returned filenames start with "/"
   virtual Status GetLiveFiles(std::vector<std::string>&,
                               uint64_t* manifest_file_size,
@@ -390,6 +398,9 @@ class DBImpl : public DB {
   virtual void GetLiveFilesMetaData(
       std::vector<LiveFileMetaData>* metadata) override;
 
+  virtual Status GetLiveFilesChecksumInfo(
+      FileChecksumList* checksum_list) override;
+
   // Obtains the meta data of the specified column family of the DB.
   // Status::NotFound() will be returned if the current DB does not have
   // any column family match the specified name.
@@ -439,6 +450,13 @@ class DBImpl : public DB {
   using DB::EndBlockCacheTrace;
   Status EndBlockCacheTrace() override;
 
+  using DB::StartIOTrace;
+  Status StartIOTrace(Env* env, const TraceOptions& options,
+                      std::unique_ptr<TraceWriter>&& trace_writer) override;
+
+  using DB::EndIOTrace;
+  Status EndIOTrace() override;
+
   using DB::GetPropertiesOfAllTables;
   virtual Status GetPropertiesOfAllTables(
       ColumnFamilyHandle* column_family,
@@ -588,8 +606,10 @@ class DBImpl : public DB {
   // the value and so will require PrepareValue() to be called before value();
   // allow_unprepared_value = false is convenient when this optimization is not
   // useful, e.g. when reading the whole column family.
+  // @param read_options Must outlive the returned iterator.
   InternalIterator* NewInternalIterator(
-      Arena* arena, RangeDelAggregator* range_del_agg, SequenceNumber sequence,
+      const ReadOptions& read_options, Arena* arena,
+      RangeDelAggregator* range_del_agg, SequenceNumber sequence,
       ColumnFamilyHandle* column_family = nullptr,
       bool allow_unprepared_value = false);
 
@@ -715,10 +735,14 @@ class DBImpl : public DB {
 
   const WriteController& write_controller() { return write_controller_; }
 
-  InternalIterator* NewInternalIterator(
-      const ReadOptions&, ColumnFamilyData* cfd, SuperVersion* super_version,
-      Arena* arena, RangeDelAggregator* range_del_agg, SequenceNumber sequence,
-      bool allow_unprepared_value);
+  // @param read_options Must outlive the returned iterator.
+  InternalIterator* NewInternalIterator(const ReadOptions& read_options,
+                                        ColumnFamilyData* cfd,
+                                        SuperVersion* super_version,
+                                        Arena* arena,
+                                        RangeDelAggregator* range_del_agg,
+                                        SequenceNumber sequence,
+                                        bool allow_unprepared_value);
 
   // hollow transactions shell used for recovery.
   // these will then be passed to TransactionDB so that
@@ -846,8 +870,8 @@ class DBImpl : public DB {
   InstrumentedMutex* mutex() const { return &mutex_; }
 
   // Initialize a brand new DB. The DB directory is expected to be empty before
-  // calling it.
-  Status NewDB();
+  // calling it. Push new manifest file name into `new_filenames`.
+  Status NewDB(std::vector<std::string>* new_filenames);
 
   // This is to be used only by internal rocksdb classes.
   static Status Open(const DBOptions& db_options, const std::string& name,
@@ -968,9 +992,7 @@ class DBImpl : public DB {
   int TEST_BGCompactionsAllowed() const;
   int TEST_BGFlushesAllowed() const;
   size_t TEST_GetWalPreallocateBlockSize(uint64_t write_buffer_size) const;
-  void TEST_WaitForDumpStatsRun(std::function<void()> callback) const;
-  void TEST_WaitForPersistStatsRun(std::function<void()> callback) const;
-  bool TEST_IsPersistentStatsEnabled() const;
+  void TEST_WaitForStatsDumpRun(std::function<void()> callback) const;
   size_t TEST_EstimateInMemoryStatsHistorySize() const;
 
   VersionSet* TEST_GetVersionSet() const { return versions_.get(); }
@@ -978,8 +1000,22 @@ class DBImpl : public DB {
   const std::unordered_set<uint64_t>& TEST_GetFilesGrabbedForPurge() const {
     return files_grabbed_for_purge_;
   }
+
+#ifndef ROCKSDB_LITE
+  PeriodicWorkTestScheduler* TEST_GetPeriodicWorkScheduler() const;
+#endif  // !ROCKSDB_LITE
+
 #endif  // NDEBUG
 
+  // persist stats to column family "_persistent_stats"
+  void PersistStats();
+
+  // dump rocksdb.stats to LOG
+  void DumpStats();
+
+  // flush LOG out of application buffer
+  void FlushInfoLog();
+
  protected:
   const std::string dbname_;
   std::string db_id_;
@@ -991,8 +1027,9 @@ class DBImpl : public DB {
   bool own_info_log_;
   const DBOptions initial_db_options_;
   Env* const env_;
-  std::shared_ptr<FileSystem> fs_;
+  std::shared_ptr<IOTracer> io_tracer_;
   const ImmutableDBOptions immutable_db_options_;
+  FileSystemPtr fs_;
   MutableDBOptions mutable_db_options_;
   Statistics* stats_;
   std::unordered_map<std::string, RecoveredTransaction*>
@@ -1152,8 +1189,8 @@ class DBImpl : public DB {
   // skipped.
   virtual Status Recover(
       const std::vector<ColumnFamilyDescriptor>& column_families,
-      bool read_only = false, bool error_if_log_file_exist = false,
-      bool error_if_data_exists_in_logs = false,
+      bool read_only = false, bool error_if_wal_file_exists = false,
+      bool error_if_data_exists_in_wals = false,
       uint64_t* recovered_seq = nullptr);
 
   virtual bool OwnTablesAndLogs() const { return true; }
@@ -1196,7 +1233,7 @@ class DBImpl : public DB {
   friend class StatsHistoryTest_PersistentStatsCreateColumnFamilies_Test;
 #ifndef NDEBUG
   friend class DBTest2_ReadCallbackTest_Test;
-  friend class WriteCallbackTest_WriteWithCallbackTest_Test;
+  friend class WriteCallbackPTest_WriteWithCallbackTest_Test;
   friend class XFTransactionWriteHandler;
   friend class DBBlobIndexTest;
   friend class WriteUnpreparedTransactionTest_RecoveryTest_Test;
@@ -1345,7 +1382,7 @@ class DBImpl : public DB {
   // Required: DB mutex held
   Status PersistentStatsProcessFormatVersion();
 
-  Status ResumeImpl();
+  Status ResumeImpl(DBRecoverContext context);
 
   void MaybeIgnoreError(Status* s) const;
 
@@ -1533,6 +1570,10 @@ class DBImpl : public DB {
                                 uint64_t* log_used,
                                 SequenceNumber* last_sequence, size_t seq_inc);
 
+  // Used by WriteImpl to update bg_error_ if paranoid check is enabled.
+  // Caller must hold mutex_.
+  void WriteStatusCheckOnLocked(const Status& status);
+
   // Used by WriteImpl to update bg_error_ if paranoid check is enabled.
   void WriteStatusCheck(const Status& status);
 
@@ -1614,18 +1655,12 @@ class DBImpl : public DB {
                               LogBuffer* log_buffer);
 
   // Schedule background tasks
-  void StartTimedTasks();
+  void StartPeriodicWorkScheduler();
 
   void PrintStatistics();
 
   size_t EstimateInMemoryStatsHistorySize() const;
 
-  // persist stats to column family "_persistent_stats"
-  void PersistStats();
-
-  // dump rocksdb.stats to LOG
-  void DumpStats();
-
   // Return the minimum empty level that could hold the total data in the
   // input level. Return the input level, if such level could not be found.
   int FindMinimumEmptyLevelFitting(ColumnFamilyData* cfd,
@@ -1704,8 +1739,8 @@ class DBImpl : public DB {
   size_t GetWalPreallocateBlockSize(uint64_t write_buffer_size) const;
   Env::WriteLifeTimeHint CalculateWALWriteHint() { return Env::WLTH_SHORT; }
 
-  Status CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
-                   size_t preallocate_block_size, log::Writer** new_log);
+  IOStatus CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
+                     size_t preallocate_block_size, log::Writer** new_log);
 
   // Validate self-consistency of DB options
   static Status ValidateOptions(const DBOptions& db_options);
@@ -1789,6 +1824,8 @@ class DBImpl : public DB {
       SuperVersion* sv, SequenceNumber snap_seqnum, ReadCallback* callback,
       bool* is_blob_index);
 
+  Status DisableFileDeletionsWithLock();
+
   // table_cache_ provides its own synchronization
   std::shared_ptr<Cache> table_cache_;
 
@@ -1806,7 +1843,14 @@ class DBImpl : public DB {
   InstrumentedMutex log_write_mutex_;
 
   std::atomic<bool> shutting_down_;
-  std::atomic<bool> manual_compaction_paused_;
+
+  // If zero, manual compactions are allowed to proceed. If non-zero, manual
+  // compactions may still be running, but will quickly fail with
+  // `Status::Incomplete`. The value indicates how many threads have paused
+  // manual compactions. It is accessed in read mode outside the DB mutex in
+  // compaction code paths.
+  std::atomic<int> manual_compaction_paused_;
+
   // This condition variable is signaled on these conditions:
   // * whenever bg_compaction_scheduled_ goes down to 0
   // * if AnyManualCompaction, whenever a compaction finishes, even if it hasn't
@@ -2069,13 +2113,13 @@ class DBImpl : public DB {
   // Only to be set during initialization
   std::unique_ptr<PreReleaseCallback> recoverable_state_pre_release_callback_;
 
-  // handle for scheduling stats dumping at fixed intervals
-  // REQUIRES: mutex locked
-  std::unique_ptr<ROCKSDB_NAMESPACE::RepeatableThread> thread_dump_stats_;
-
-  // handle for scheduling stats snapshoting at fixed intervals
-  // REQUIRES: mutex locked
-  std::unique_ptr<ROCKSDB_NAMESPACE::RepeatableThread> thread_persist_stats_;
+#ifndef ROCKSDB_LITE
+  // Scheduler to run DumpStats(), PersistStats(), and FlushInfoLog().
+  // Currently, it always use a global instance from
+  // PeriodicWorkScheduler::Default(). Only in unittest, it can be overrided by
+  // PeriodicWorkTestScheduler.
+  PeriodicWorkScheduler* periodic_work_scheduler_;
+#endif
 
   // When set, we use a separate queue for writes that don't write to memtable.
   // In 2PC these are the writes at Prepare phase.

db/db_impl/db_impl_compaction_flush.cc

@@ -6,11 +6,10 @@
 // Copyright (c) 2011 The LevelDB Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
-#include "db/db_impl/db_impl.h"
-
 #include <cinttypes>
 
 #include "db/builder.h"
+#include "db/db_impl/db_impl.h"
 #include "db/error_handler.h"
 #include "db/event_helpers.h"
 #include "file/sst_file_manager_impl.h"
@@ -126,7 +125,14 @@ IOStatus DBImpl::SyncClosedLogs(JobContext* job_context) {
     // "number < current_log_number".
     MarkLogsSynced(current_log_number - 1, true, io_s);
     if (!io_s.ok()) {
-      error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush);
+      if (total_log_size_ > 0) {
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush)
+            .PermitUncheckedError();
+      } else {
+        // If the WAL is empty, we use different error reason
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlushNoWAL)
+            .PermitUncheckedError();
+      }
       TEST_SYNC_POINT("DBImpl::SyncClosedLogs:Failed");
       return io_s;
     }
@@ -143,6 +149,7 @@ Status DBImpl::FlushMemTableToOutputFile(
     SnapshotChecker* snapshot_checker, LogBuffer* log_buffer,
     Env::Priority thread_pri) {
   mutex_.AssertHeld();
+  assert(cfd);
   assert(cfd->imm()->NumNotFlushed() != 0);
   assert(cfd->imm()->IsFlushPending());
 
@@ -155,7 +162,7 @@ Status DBImpl::FlushMemTableToOutputFile(
       GetCompressionFlush(*cfd->ioptions(), mutable_cf_options), stats_,
       &event_logger_, mutable_cf_options.report_bg_io_stats,
       true /* sync_output_directory */, true /* write_manifest */, thread_pri,
-      db_id_, db_session_id_);
+      io_tracer_, db_id_, db_session_id_);
   FileMetaData file_meta;
 
   TEST_SYNC_POINT("DBImpl::FlushMemTableToOutputFile:BeforePickMemtables");
@@ -168,7 +175,7 @@ Status DBImpl::FlushMemTableToOutputFile(
 #endif  // ROCKSDB_LITE
 
   Status s;
-  IOStatus io_s;
+  IOStatus io_s = IOStatus::OK();
   if (logfile_number_ > 0 &&
       versions_->GetColumnFamilySet()->NumberOfColumnFamilies() > 1) {
     // If there are more than one column families, we need to make sure that
@@ -179,6 +186,11 @@ Status DBImpl::FlushMemTableToOutputFile(
     // SyncClosedLogs() may unlock and re-lock the db_mutex.
     io_s = SyncClosedLogs(job_context);
     s = io_s;
+    if (!io_s.ok() && !io_s.IsShutdownInProgress() &&
+        !io_s.IsColumnFamilyDropped()) {
+      error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush)
+          .PermitUncheckedError();
+    }
   } else {
     TEST_SYNC_POINT("DBImpl::SyncClosedLogs:Skip");
   }
@@ -194,7 +206,9 @@ Status DBImpl::FlushMemTableToOutputFile(
   } else {
     flush_job.Cancel();
   }
-  io_s = flush_job.io_status();
+  if (io_s.ok()) {
+    io_s = flush_job.io_status();
+  }
 
   if (s.ok()) {
     InstallSuperVersionAndScheduleWork(cfd, superversion_context,
@@ -202,20 +216,62 @@ Status DBImpl::FlushMemTableToOutputFile(
     if (made_progress) {
       *made_progress = true;
     }
+
+    const std::string& column_family_name = cfd->GetName();
+
+    Version* const current = cfd->current();
+    assert(current);
+
+    const VersionStorageInfo* const storage_info = current->storage_info();
+    assert(storage_info);
+
     VersionStorageInfo::LevelSummaryStorage tmp;
     ROCKS_LOG_BUFFER(log_buffer, "[%s] Level summary: %s\n",
-                     cfd->GetName().c_str(),
-                     cfd->current()->storage_info()->LevelSummary(&tmp));
+                     column_family_name.c_str(),
+                     storage_info->LevelSummary(&tmp));
+
+    const auto& blob_files = storage_info->GetBlobFiles();
+    if (!blob_files.empty()) {
+      ROCKS_LOG_BUFFER(log_buffer,
+                       "[%s] Blob file summary: head=%" PRIu64 ", tail=%" PRIu64
+                       "\n",
+                       column_family_name.c_str(), blob_files.begin()->first,
+                       blob_files.rbegin()->first);
+    }
   }
 
   if (!s.ok() && !s.IsShutdownInProgress() && !s.IsColumnFamilyDropped()) {
     if (!io_s.ok() && !io_s.IsShutdownInProgress() &&
         !io_s.IsColumnFamilyDropped()) {
-      error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush);
+      // Error while writing to MANIFEST.
+      // In fact, versions_->io_status() can also be the result of renaming
+      // CURRENT file. With current code, it's just difficult to tell. So just
+      // be pessimistic and try write to a new MANIFEST.
+      // TODO: distinguish between MANIFEST write and CURRENT renaming
+      if (!versions_->io_status().ok()) {
+        // Should handle return error?
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kManifestWrite)
+            .PermitUncheckedError();
+      } else if (total_log_size_ > 0) {
+        // Should handle return error?
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush)
+            .PermitUncheckedError();
+      } else {
+        // If the WAL is empty, we use different error reason
+        // Should handle return error?
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlushNoWAL)
+            .PermitUncheckedError();
+      }
     } else {
       Status new_bg_error = s;
-      error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush);
+      // Should handle return error?
+      error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush)
+          .PermitUncheckedError();
     }
+  } else {
+    // If we got here, then we decided not to care about the i_os status (either
+    // from never needing it or ignoring the flush job status
+    io_s.PermitUncheckedError();
   }
   if (s.ok()) {
 #ifndef ROCKSDB_LITE
@@ -235,7 +291,9 @@ Status DBImpl::FlushMemTableToOutputFile(
         TEST_SYNC_POINT_CALLBACK(
             "DBImpl::FlushMemTableToOutputFile:MaxAllowedSpaceReached",
             &new_bg_error);
-        error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush);
+        // Should handle this error?
+        error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush)
+            .PermitUncheckedError();
       }
     }
 #endif  // ROCKSDB_LITE
@@ -346,7 +404,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
         data_dir, GetCompressionFlush(*cfd->ioptions(), mutable_cf_options),
         stats_, &event_logger_, mutable_cf_options.report_bg_io_stats,
         false /* sync_output_directory */, false /* write_manifest */,
-        thread_pri, db_id_, db_session_id_));
+        thread_pri, io_tracer_, db_id_, db_session_id_));
     jobs.back()->PickMemTable();
   }
 
@@ -531,16 +589,36 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
     assert(num_cfs ==
            static_cast<int>(job_context->superversion_contexts.size()));
     for (int i = 0; i != num_cfs; ++i) {
+      assert(cfds[i]);
+
       if (cfds[i]->IsDropped()) {
         continue;
       }
       InstallSuperVersionAndScheduleWork(cfds[i],
                                          &job_context->superversion_contexts[i],
                                          all_mutable_cf_options[i]);
+
+      const std::string& column_family_name = cfds[i]->GetName();
+
+      Version* const current = cfds[i]->current();
+      assert(current);
+
+      const VersionStorageInfo* const storage_info = current->storage_info();
+      assert(storage_info);
+
       VersionStorageInfo::LevelSummaryStorage tmp;
       ROCKS_LOG_BUFFER(log_buffer, "[%s] Level summary: %s\n",
-                       cfds[i]->GetName().c_str(),
-                       cfds[i]->current()->storage_info()->LevelSummary(&tmp));
+                       column_family_name.c_str(),
+                       storage_info->LevelSummary(&tmp));
+
+      const auto& blob_files = storage_info->GetBlobFiles();
+      if (!blob_files.empty()) {
+        ROCKS_LOG_BUFFER(log_buffer,
+                         "[%s] Blob file summary: head=%" PRIu64
+                         ", tail=%" PRIu64 "\n",
+                         column_family_name.c_str(), blob_files.begin()->first,
+                         blob_files.rbegin()->first);
+      }
     }
     if (made_progress) {
       *made_progress = true;
@@ -563,8 +641,9 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
             error_handler_.GetBGError().ok()) {
           Status new_bg_error =
               Status::SpaceLimit("Max allowed space was reached");
-          error_handler_.SetBGError(new_bg_error,
-                                    BackgroundErrorReason::kFlush);
+          // Should Handle this error?
+          error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush)
+              .PermitUncheckedError();
         }
       }
     }
@@ -575,10 +654,30 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
   // it is not because of CF drop.
   if (!s.ok() && !s.IsColumnFamilyDropped()) {
     if (!io_s.ok() && !io_s.IsColumnFamilyDropped()) {
-      error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush);
+      // Error while writing to MANIFEST.
+      // In fact, versions_->io_status() can also be the result of renaming
+      // CURRENT file. With current code, it's just difficult to tell. So just
+      // be pessimistic and try write to a new MANIFEST.
+      // TODO: distinguish between MANIFEST write and CURRENT renaming
+      if (!versions_->io_status().ok()) {
+        // Should Handle this error?
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kManifestWrite)
+            .PermitUncheckedError();
+      } else if (total_log_size_ > 0) {
+        // Should Handle this error?
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlush)
+            .PermitUncheckedError();
+      } else {
+        // If the WAL is empty, we use different error reason
+        // Should Handle this error?
+        error_handler_.SetBGError(io_s, BackgroundErrorReason::kFlushNoWAL)
+            .PermitUncheckedError();
+      }
     } else {
       Status new_bg_error = s;
-      error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush);
+      // Should Handle this error?
+      error_handler_.SetBGError(new_bg_error, BackgroundErrorReason::kFlush)
+          .PermitUncheckedError();
     }
   }
 
@@ -679,7 +778,7 @@ void DBImpl::NotifyOnFlushCompleted(
 Status DBImpl::CompactRange(const CompactRangeOptions& options,
                             ColumnFamilyHandle* column_family,
                             const Slice* begin, const Slice* end) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   auto cfd = cfh->cfd();
 
   if (options.target_path_id >= cfd->ioptions()->cf_paths.size()) {
@@ -774,6 +873,7 @@ Status DBImpl::CompactRange(const CompactRangeOptions& options,
       int output_level;
       for (int level = first_overlapped_level; level <= max_overlapped_level;
            level++) {
+        bool disallow_trivial_move = false;
         // in case the compaction is universal or if we're compacting the
         // bottom-most level, the output level will be the same as input one.
         // level 0 can never be the bottommost level (i.e. if all files are in
@@ -806,9 +906,19 @@ Status DBImpl::CompactRange(const CompactRangeOptions& options,
               level == 0) {
             output_level = ColumnFamilyData::kCompactToBaseLevel;
           }
+          // if it's a BottommostLevel compaction and `kForce*` compaction is
+          // set, disallow trivial move
+          if (level == max_overlapped_level &&
+              (options.bottommost_level_compaction ==
+                   BottommostLevelCompaction::kForce ||
+               options.bottommost_level_compaction ==
+                   BottommostLevelCompaction::kForceOptimized)) {
+            disallow_trivial_move = true;
+          }
         }
         s = RunManualCompaction(cfd, level, output_level, options, begin, end,
-                                exclusive, false, max_file_num_to_ignore);
+                                exclusive, disallow_trivial_move,
+                                max_file_num_to_ignore);
         if (!s.ok()) {
           break;
         }
@@ -828,13 +938,21 @@ Status DBImpl::CompactRange(const CompactRangeOptions& options,
   }
 
   if (options.change_level) {
+    TEST_SYNC_POINT("DBImpl::CompactRange:BeforeRefit:1");
+    TEST_SYNC_POINT("DBImpl::CompactRange:BeforeRefit:2");
+
     ROCKS_LOG_INFO(immutable_db_options_.info_log,
                    "[RefitLevel] waiting for background threads to stop");
+    DisableManualCompaction();
     s = PauseBackgroundWork();
     if (s.ok()) {
+      TEST_SYNC_POINT("DBImpl::CompactRange:PreRefitLevel");
       s = ReFitLevel(cfd, final_output_level, options.target_level);
+      TEST_SYNC_POINT("DBImpl::CompactRange:PostRefitLevel");
+      // ContinueBackgroundWork always return Status::OK().
+      assert(ContinueBackgroundWork().ok());
     }
-    ContinueBackgroundWork();
+    EnableManualCompaction();
   }
   LogFlush(immutable_db_options_.info_log);
 
@@ -869,7 +987,8 @@ Status DBImpl::CompactFiles(const CompactionOptions& compact_options,
     return Status::InvalidArgument("ColumnFamilyHandle must be non-null.");
   }
 
-  auto cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family)->cfd();
+  auto cfd =
+      static_cast_with_check<ColumnFamilyHandleImpl>(column_family)->cfd();
   assert(cfd);
 
   Status s;
@@ -940,7 +1059,7 @@ Status DBImpl::CompactFilesImpl(
   if (shutting_down_.load(std::memory_order_acquire)) {
     return Status::ShutdownInProgress();
   }
-  if (manual_compaction_paused_.load(std::memory_order_acquire)) {
+  if (manual_compaction_paused_.load(std::memory_order_acquire) > 0) {
     return Status::Incomplete(Status::SubCode::kManualCompactionPaused);
   }
 
@@ -1002,7 +1121,7 @@ Status DBImpl::CompactFilesImpl(
   assert(cfd->compaction_picker());
   c.reset(cfd->compaction_picker()->CompactFiles(
       compact_options, input_files, output_level, version->storage_info(),
-      *cfd->GetLatestMutableCFOptions(), output_path_id));
+      *cfd->GetLatestMutableCFOptions(), mutable_db_options_, output_path_id));
   // we already sanitized the set of input files and checked for conflicts
   // without releasing the lock, so we're guaranteed a compaction can be formed.
   assert(c != nullptr);
@@ -1032,8 +1151,8 @@ Status DBImpl::CompactFilesImpl(
       snapshot_checker, table_cache_, &event_logger_,
       c->mutable_cf_options()->paranoid_file_checks,
       c->mutable_cf_options()->report_bg_io_stats, dbname_,
-      &compaction_job_stats, Env::Priority::USER, &manual_compaction_paused_,
-      db_id_, db_session_id_);
+      &compaction_job_stats, Env::Priority::USER, io_tracer_,
+      &manual_compaction_paused_, db_id_, db_session_id_);
 
   // Creating a compaction influences the compaction score because the score
   // takes running compactions into account (by skipping files that are already
@@ -1054,10 +1173,15 @@ Status DBImpl::CompactFilesImpl(
 
   Status status = compaction_job.Install(*c->mutable_cf_options());
   if (status.ok()) {
+    assert(compaction_job.io_status().ok());
     InstallSuperVersionAndScheduleWork(c->column_family_data(),
                                        &job_context->superversion_contexts[0],
                                        *c->mutable_cf_options());
   }
+  // status above captures any error during compaction_job.Install, so its ok
+  // not check compaction_job.io_status() explicitly if we're not calling
+  // SetBGError
+  compaction_job.io_status().PermitUncheckedError();
   c->ReleaseCompactionFiles(s);
 #ifndef ROCKSDB_LITE
   // Need to make sure SstFileManager does its bookkeeping
@@ -1090,7 +1214,14 @@ Status DBImpl::CompactFilesImpl(
                    "[%s] [JOB %d] Compaction error: %s",
                    c->column_family_data()->GetName().c_str(),
                    job_context->job_id, status.ToString().c_str());
-    error_handler_.SetBGError(status, BackgroundErrorReason::kCompaction);
+    IOStatus io_s = compaction_job.io_status();
+    if (!io_s.ok()) {
+      error_handler_.SetBGError(io_s, BackgroundErrorReason::kCompaction)
+          .PermitUncheckedError();
+    } else {
+      error_handler_.SetBGError(status, BackgroundErrorReason::kCompaction)
+          .PermitUncheckedError();
+    }
   }
 
   if (output_file_names != nullptr) {
@@ -1156,7 +1287,7 @@ void DBImpl::NotifyOnCompactionBegin(ColumnFamilyData* cfd, Compaction* c,
     return;
   }
   if (c->is_manual_compaction() &&
-      manual_compaction_paused_.load(std::memory_order_acquire)) {
+      manual_compaction_paused_.load(std::memory_order_acquire) > 0) {
     return;
   }
   Version* current = cfd->current();
@@ -1166,46 +1297,11 @@ void DBImpl::NotifyOnCompactionBegin(ColumnFamilyData* cfd, Compaction* c,
   TEST_SYNC_POINT("DBImpl::NotifyOnCompactionBegin::UnlockMutex");
   {
     CompactionJobInfo info{};
-    info.cf_name = cfd->GetName();
-    info.status = st;
-    info.thread_id = env_->GetThreadID();
-    info.job_id = job_id;
-    info.base_input_level = c->start_level();
-    info.output_level = c->output_level();
-    info.stats = job_stats;
-    info.table_properties = c->GetOutputTableProperties();
-    info.compaction_reason = c->compaction_reason();
-    info.compression = c->output_compression();
-    for (size_t i = 0; i < c->num_input_levels(); ++i) {
-      for (const auto fmd : *c->inputs(i)) {
-        const FileDescriptor& desc = fmd->fd;
-        const uint64_t file_number = desc.GetNumber();
-        auto fn = TableFileName(c->immutable_cf_options()->cf_paths,
-                                file_number, desc.GetPathId());
-        info.input_files.push_back(fn);
-        info.input_file_infos.push_back(CompactionFileInfo{
-            static_cast<int>(i), file_number, fmd->oldest_blob_file_number});
-        if (info.table_properties.count(fn) == 0) {
-          std::shared_ptr<const TableProperties> tp;
-          auto s = current->GetTableProperties(&tp, fmd, &fn);
-          if (s.ok()) {
-            info.table_properties[fn] = tp;
-          }
-        }
-      }
-    }
-    for (const auto& newf : c->edit()->GetNewFiles()) {
-      const FileMetaData& meta = newf.second;
-      const FileDescriptor& desc = meta.fd;
-      const uint64_t file_number = desc.GetNumber();
-      info.output_files.push_back(TableFileName(
-          c->immutable_cf_options()->cf_paths, file_number, desc.GetPathId()));
-      info.output_file_infos.push_back(CompactionFileInfo{
-          newf.first, file_number, meta.oldest_blob_file_number});
-    }
+    BuildCompactionJobInfo(cfd, c, st, job_stats, job_id, current, &info);
     for (auto listener : immutable_db_options_.listeners) {
       listener->OnCompactionBegin(this, info);
     }
+    info.status.PermitUncheckedError();
   }
   mutex_.Lock();
   current->Unref();
@@ -1230,7 +1326,7 @@ void DBImpl::NotifyOnCompactionCompleted(
     return;
   }
   if (c->is_manual_compaction() &&
-      manual_compaction_paused_.load(std::memory_order_acquire)) {
+      manual_compaction_paused_.load(std::memory_order_acquire) > 0) {
     return;
   }
   Version* current = cfd->current();
@@ -1289,20 +1385,32 @@ Status DBImpl::ReFitLevel(ColumnFamilyData* cfd, int level, int target_level) {
   }
 
   auto* vstorage = cfd->current()->storage_info();
-  if (to_level > level) {
-    if (level == 0) {
-      return Status::NotSupported(
-          "Cannot change from level 0 to other levels.");
-    }
-    // Check levels are empty for a trivial move
-    for (int l = level + 1; l <= to_level; l++) {
-      if (vstorage->NumLevelFiles(l) > 0) {
+  if (to_level != level) {
+    if (to_level > level) {
+      if (level == 0) {
+        refitting_level_ = false;
         return Status::NotSupported(
-            "Levels between source and target are not empty for a move.");
+            "Cannot change from level 0 to other levels.");
+      }
+      // Check levels are empty for a trivial move
+      for (int l = level + 1; l <= to_level; l++) {
+        if (vstorage->NumLevelFiles(l) > 0) {
+          refitting_level_ = false;
+          return Status::NotSupported(
+              "Levels between source and target are not empty for a move.");
+        }
+      }
+    } else {
+      // to_level < level
+      // Check levels are empty for a trivial move
+      for (int l = to_level; l < level; l++) {
+        if (vstorage->NumLevelFiles(l) > 0) {
+          refitting_level_ = false;
+          return Status::NotSupported(
+              "Levels between source and target are not empty for a move.");
+        }
       }
     }
-  }
-  if (to_level != level) {
     ROCKS_LOG_DEBUG(immutable_db_options_.info_log,
                     "[%s] Before refitting:\n%s", cfd->GetName().c_str(),
                     cfd->current()->DebugString().data());
@@ -1343,7 +1451,7 @@ Status DBImpl::ReFitLevel(ColumnFamilyData* cfd, int level, int target_level) {
 }
 
 int DBImpl::NumberLevels(ColumnFamilyHandle* column_family) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   return cfh->cfd()->NumberLevels();
 }
 
@@ -1352,7 +1460,7 @@ int DBImpl::MaxMemCompactionLevel(ColumnFamilyHandle* /*column_family*/) {
 }
 
 int DBImpl::Level0StopWriteTrigger(ColumnFamilyHandle* column_family) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   InstrumentedMutexLock l(&mutex_);
   return cfh->cfd()
       ->GetSuperVersion()
@@ -1361,7 +1469,7 @@ int DBImpl::Level0StopWriteTrigger(ColumnFamilyHandle* column_family) {
 
 Status DBImpl::Flush(const FlushOptions& flush_options,
                      ColumnFamilyHandle* column_family) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   ROCKS_LOG_INFO(immutable_db_options_.info_log, "[%s] Manual flush start.",
                  cfh->GetName().c_str());
   Status s;
@@ -1512,9 +1620,9 @@ Status DBImpl::RunManualCompaction(
         scheduled ||
         (((manual.manual_end = &manual.tmp_storage1) != nullptr) &&
          ((compaction = manual.cfd->CompactRange(
-               *manual.cfd->GetLatestMutableCFOptions(), manual.input_level,
-               manual.output_level, compact_range_options, manual.begin,
-               manual.end, &manual.manual_end, &manual_conflict,
+               *manual.cfd->GetLatestMutableCFOptions(), mutable_db_options_,
+               manual.input_level, manual.output_level, compact_range_options,
+               manual.begin, manual.end, &manual.manual_end, &manual_conflict,
                max_file_num_to_ignore)) == nullptr &&
           manual_conflict))) {
       // exclusive manual compactions should not see a conflict during
@@ -1591,6 +1699,7 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
       return s;
     }
   }
+
   FlushRequest flush_req;
   {
     WriteContext context;
@@ -1607,7 +1716,11 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
     WaitForPendingWrites();
 
     if (!cfd->mem()->IsEmpty() || !cached_recoverable_state_empty_.load()) {
-      s = SwitchMemtable(cfd, &context);
+      if (flush_reason != FlushReason::kErrorRecoveryRetryFlush) {
+        s = SwitchMemtable(cfd, &context);
+      } else {
+        assert(cfd->imm()->NumNotFlushed() > 0);
+      }
     }
     if (s.ok()) {
       if (cfd->imm()->NumNotFlushed() != 0 || !cfd->mem()->IsEmpty() ||
@@ -1615,7 +1728,8 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
         flush_memtable_id = cfd->imm()->GetLatestMemTableID();
         flush_req.emplace_back(cfd, flush_memtable_id);
       }
-      if (immutable_db_options_.persist_stats_to_disk) {
+      if (immutable_db_options_.persist_stats_to_disk &&
+          flush_reason != FlushReason::kErrorRecoveryRetryFlush) {
         ColumnFamilyData* cfd_stats =
             versions_->GetColumnFamilySet()->GetColumnFamily(
                 kPersistentStatsColumnFamilyName);
@@ -1644,7 +1758,6 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
         }
       }
     }
-
     if (s.ok() && !flush_req.empty()) {
       for (auto& elem : flush_req) {
         ColumnFamilyData* loop_cfd = elem.first;
@@ -1680,8 +1793,10 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
       cfds.push_back(iter.first);
       flush_memtable_ids.push_back(&(iter.second));
     }
-    s = WaitForFlushMemTables(cfds, flush_memtable_ids,
-                              (flush_reason == FlushReason::kErrorRecovery));
+    s = WaitForFlushMemTables(
+        cfds, flush_memtable_ids,
+        (flush_reason == FlushReason::kErrorRecovery ||
+         flush_reason == FlushReason::kErrorRecoveryRetryFlush));
     InstrumentedMutexLock lock_guard(&mutex_);
     for (auto* tmp_cfd : cfds) {
       tmp_cfd->UnrefAndTryDelete();
@@ -1739,7 +1854,8 @@ Status DBImpl::AtomicFlushMemTables(
       }
     }
     for (auto cfd : cfds) {
-      if (cfd->mem()->IsEmpty() && cached_recoverable_state_empty_.load()) {
+      if ((cfd->mem()->IsEmpty() && cached_recoverable_state_empty_.load()) ||
+          flush_reason == FlushReason::kErrorRecoveryRetryFlush) {
         continue;
       }
       cfd->Ref();
@@ -1782,8 +1898,10 @@ Status DBImpl::AtomicFlushMemTables(
     for (auto& iter : flush_req) {
       flush_memtable_ids.push_back(&(iter.second));
     }
-    s = WaitForFlushMemTables(cfds, flush_memtable_ids,
-                              (flush_reason == FlushReason::kErrorRecovery));
+    s = WaitForFlushMemTables(
+        cfds, flush_memtable_ids,
+        (flush_reason == FlushReason::kErrorRecovery ||
+         flush_reason == FlushReason::kErrorRecoveryRetryFlush));
     InstrumentedMutexLock lock_guard(&mutex_);
     for (auto* cfd : cfds) {
       cfd->UnrefAndTryDelete();
@@ -1893,6 +2011,13 @@ Status DBImpl::WaitForFlushMemTables(
     if (!error_handler_.GetRecoveryError().ok()) {
       break;
     }
+    // If BGWorkStopped, which indicate that there is a BG error and
+    // 1) soft error but requires no BG work, 2) no in auto_recovery_
+    if (!resuming_from_bg_err && error_handler_.IsBGWorkStopped() &&
+        error_handler_.GetBGError().severity() < Status::Severity::kHardError) {
+      return error_handler_.GetBGError();
+    }
+
     // Number of column families that have been dropped.
     int num_dropped = 0;
     // Number of column families that have finished flush.
@@ -1941,11 +2066,21 @@ Status DBImpl::EnableAutoCompaction(
 }
 
 void DBImpl::DisableManualCompaction() {
-  manual_compaction_paused_.store(true, std::memory_order_release);
+  InstrumentedMutexLock l(&mutex_);
+  manual_compaction_paused_.fetch_add(1, std::memory_order_release);
+  // Wait for any pending manual compactions to finish (typically through
+  // failing with `Status::Incomplete`) prior to returning. This way we are
+  // guaranteed no pending manual compaction will commit while manual
+  // compactions are "disabled".
+  while (HasPendingManualCompaction()) {
+    bg_cv_.Wait();
+  }
 }
 
 void DBImpl::EnableManualCompaction() {
-  manual_compaction_paused_.store(false, std::memory_order_release);
+  InstrumentedMutexLock l(&mutex_);
+  assert(manual_compaction_paused_ > 0);
+  manual_compaction_paused_.fetch_sub(1, std::memory_order_release);
 }
 
 void DBImpl::MaybeScheduleFlushOrCompaction() {
@@ -2504,7 +2639,7 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
     if (shutting_down_.load(std::memory_order_acquire)) {
       status = Status::ShutdownInProgress();
     } else if (is_manual &&
-               manual_compaction_paused_.load(std::memory_order_acquire)) {
+               manual_compaction_paused_.load(std::memory_order_acquire) > 0) {
       status = Status::Incomplete(Status::SubCode::kManualCompactionPaused);
     }
   } else {
@@ -2546,12 +2681,13 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
     if (!c) {
       m->done = true;
       m->manual_end = nullptr;
-      ROCKS_LOG_BUFFER(log_buffer,
-                       "[%s] Manual compaction from level-%d from %s .. "
-                       "%s; nothing to do\n",
-                       m->cfd->GetName().c_str(), m->input_level,
-                       (m->begin ? m->begin->DebugString(true).c_str() : "(begin)"),
-                       (m->end ? m->end->DebugString(true).c_str() : "(end)"));
+      ROCKS_LOG_BUFFER(
+          log_buffer,
+          "[%s] Manual compaction from level-%d from %s .. "
+          "%s; nothing to do\n",
+          m->cfd->GetName().c_str(), m->input_level,
+          (m->begin ? m->begin->DebugString(true).c_str() : "(begin)"),
+          (m->end ? m->end->DebugString(true).c_str() : "(end)"));
     } else {
       // First check if we have enough room to do the compaction
       bool enough_room = EnoughRoomForCompaction(
@@ -2618,7 +2754,8 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
       // compaction is not necessary. Need to make sure mutex is held
       // until we make a copy in the following code
       TEST_SYNC_POINT("DBImpl::BackgroundCompaction():BeforePickCompaction");
-      c.reset(cfd->PickCompaction(*mutable_cf_options, log_buffer));
+      c.reset(cfd->PickCompaction(*mutable_cf_options, mutable_db_options_,
+                                  log_buffer));
       TEST_SYNC_POINT("DBImpl::BackgroundCompaction():AfterPickCompaction");
 
       if (c != nullptr) {
@@ -2689,7 +2826,6 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
     for (const auto& f : *c->inputs(0)) {
       c->edit()->DeleteFile(c->level(), f->fd.GetNumber());
     }
-    versions_->SetIOStatusOK();
     status = versions_->LogAndApply(c->column_family_data(),
                                     *c->mutable_cf_options(), c->edit(),
                                     &mutex_, directories_.GetDbDir());
@@ -2747,7 +2883,6 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
       }
     }
 
-    versions_->SetIOStatusOK();
     status = versions_->LogAndApply(c->column_family_data(),
                                     *c->mutable_cf_options(), c->edit(),
                                     &mutex_, directories_.GetDbDir());
@@ -2823,18 +2958,18 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
         earliest_write_conflict_snapshot, snapshot_checker, table_cache_,
         &event_logger_, c->mutable_cf_options()->paranoid_file_checks,
         c->mutable_cf_options()->report_bg_io_stats, dbname_,
-        &compaction_job_stats, thread_pri,
+        &compaction_job_stats, thread_pri, io_tracer_,
         is_manual ? &manual_compaction_paused_ : nullptr, db_id_,
         db_session_id_);
     compaction_job.Prepare();
 
     NotifyOnCompactionBegin(c->column_family_data(), c.get(), status,
                             compaction_job_stats, job_context->job_id);
-
     mutex_.Unlock();
     TEST_SYNC_POINT_CALLBACK(
         "DBImpl::BackgroundCompaction:NonTrivial:BeforeRun", nullptr);
-    compaction_job.Run();
+    // Should handle erorr?
+    compaction_job.Run().PermitUncheckedError();
     TEST_SYNC_POINT("DBImpl::BackgroundCompaction:NonTrivial:AfterRun");
     mutex_.Lock();
 
@@ -2852,6 +2987,8 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
 
   if (status.ok() && !io_s.ok()) {
     status = io_s;
+  } else {
+    io_s.PermitUncheckedError();
   }
 
   if (c != nullptr) {
@@ -2880,9 +3017,18 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
     ROCKS_LOG_WARN(immutable_db_options_.info_log, "Compaction error: %s",
                    status.ToString().c_str());
     if (!io_s.ok()) {
-      error_handler_.SetBGError(io_s, BackgroundErrorReason::kCompaction);
+      // Error while writing to MANIFEST.
+      // In fact, versions_->io_status() can also be the result of renaming
+      // CURRENT file. With current code, it's just difficult to tell. So just
+      // be pessimistic and try write to a new MANIFEST.
+      // TODO: distinguish between MANIFEST write and CURRENT renaming
+      auto err_reason = versions_->io_status().ok()
+                            ? BackgroundErrorReason::kCompaction
+                            : BackgroundErrorReason::kManifestWrite;
+      error_handler_.SetBGError(io_s, err_reason).PermitUncheckedError();
     } else {
-      error_handler_.SetBGError(status, BackgroundErrorReason::kCompaction);
+      error_handler_.SetBGError(status, BackgroundErrorReason::kCompaction)
+          .PermitUncheckedError();
     }
     if (c != nullptr && !is_manual && !error_handler_.IsBGWorkStopped()) {
       // Put this cfd back in the compaction queue so we can retry after some

db/db_impl/db_impl_debug.cc

@@ -12,6 +12,7 @@
 #include "db/column_family.h"
 #include "db/db_impl/db_impl.h"
 #include "db/error_handler.h"
+#include "db/periodic_work_scheduler.h"
 #include "monitoring/thread_status_updater.h"
 #include "util/cast_util.h"
 
@@ -47,7 +48,7 @@ int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes(
   if (column_family == nullptr) {
     cfd = default_cf_handle_->cfd();
   } else {
-    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+    auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
     cfd = cfh->cfd();
   }
   InstrumentedMutexLock l(&mutex_);
@@ -57,7 +58,7 @@ int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes(
 void DBImpl::TEST_GetFilesMetaData(
     ColumnFamilyHandle* column_family,
     std::vector<std::vector<FileMetaData>>* metadata) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   auto cfd = cfh->cfd();
   InstrumentedMutexLock l(&mutex_);
   metadata->resize(NumberLevels());
@@ -88,7 +89,7 @@ Status DBImpl::TEST_CompactRange(int level, const Slice* begin,
   if (column_family == nullptr) {
     cfd = default_cf_handle_->cfd();
   } else {
-    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+    auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
     cfd = cfh->cfd();
   }
   int output_level =
@@ -131,7 +132,7 @@ Status DBImpl::TEST_FlushMemTable(bool wait, bool allow_write_stall,
   if (cfh == nullptr) {
     cfd = default_cf_handle_->cfd();
   } else {
-    auto cfhi = reinterpret_cast<ColumnFamilyHandleImpl*>(cfh);
+    auto cfhi = static_cast_with_check<ColumnFamilyHandleImpl>(cfh);
     cfd = cfhi->cfd();
   }
   return FlushMemTable(cfd, fo, FlushReason::kTest);
@@ -152,7 +153,7 @@ Status DBImpl::TEST_WaitForFlushMemTable(ColumnFamilyHandle* column_family) {
   if (column_family == nullptr) {
     cfd = default_cf_handle_->cfd();
   } else {
-    auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+    auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
     cfd = cfh->cfd();
   }
   return WaitForFlushMemTable(cfd, nullptr, false);
@@ -242,7 +243,7 @@ Status DBImpl::TEST_GetLatestMutableCFOptions(
     ColumnFamilyHandle* column_family, MutableCFOptions* mutable_cf_options) {
   InstrumentedMutexLock l(&mutex_);
 
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   *mutable_cf_options = *cfh->cfd()->GetLatestMutableCFOptions();
   return Status::OK();
 }
@@ -271,21 +272,18 @@ size_t DBImpl::TEST_GetWalPreallocateBlockSize(
   return GetWalPreallocateBlockSize(write_buffer_size);
 }
 
-void DBImpl::TEST_WaitForDumpStatsRun(std::function<void()> callback) const {
-  if (thread_dump_stats_ != nullptr) {
-    thread_dump_stats_->TEST_WaitForRun(callback);
+#ifndef ROCKSDB_LITE
+void DBImpl::TEST_WaitForStatsDumpRun(std::function<void()> callback) const {
+  if (periodic_work_scheduler_ != nullptr) {
+    static_cast<PeriodicWorkTestScheduler*>(periodic_work_scheduler_)
+        ->TEST_WaitForRun(callback);
   }
 }
 
-void DBImpl::TEST_WaitForPersistStatsRun(std::function<void()> callback) const {
-  if (thread_persist_stats_ != nullptr) {
-    thread_persist_stats_->TEST_WaitForRun(callback);
-  }
-}
-
-bool DBImpl::TEST_IsPersistentStatsEnabled() const {
-  return thread_persist_stats_ && thread_persist_stats_->IsRunning();
+PeriodicWorkTestScheduler* DBImpl::TEST_GetPeriodicWorkScheduler() const {
+  return static_cast<PeriodicWorkTestScheduler*>(periodic_work_scheduler_);
 }
+#endif  // !ROCKSDB_LITE
 
 size_t DBImpl::TEST_EstimateInMemoryStatsHistorySize() const {
   return EstimateInMemoryStatsHistorySize();

db/db_impl/db_impl_experimental.cc

@@ -7,22 +7,22 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
 
-#include "db/db_impl/db_impl.h"
-
 #include <cinttypes>
 #include <vector>
 
 #include "db/column_family.h"
+#include "db/db_impl/db_impl.h"
 #include "db/job_context.h"
 #include "db/version_set.h"
 #include "rocksdb/status.h"
+#include "util/cast_util.h"
 
 namespace ROCKSDB_NAMESPACE {
 
 #ifndef ROCKSDB_LITE
 Status DBImpl::SuggestCompactRange(ColumnFamilyHandle* column_family,
                                    const Slice* begin, const Slice* end) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   auto cfd = cfh->cfd();
   InternalKey start_key, end_key;
   if (begin != nullptr) {

db/db_impl/db_impl_files.cc

@@ -36,6 +36,62 @@ uint64_t DBImpl::MinObsoleteSstNumberToKeep() {
   return std::numeric_limits<uint64_t>::max();
 }
 
+Status DBImpl::DisableFileDeletions() {
+  InstrumentedMutexLock l(&mutex_);
+  return DisableFileDeletionsWithLock();
+}
+
+Status DBImpl::DisableFileDeletionsWithLock() {
+  mutex_.AssertHeld();
+  ++disable_delete_obsolete_files_;
+  if (disable_delete_obsolete_files_ == 1) {
+    ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Disabled");
+  } else {
+    ROCKS_LOG_WARN(immutable_db_options_.info_log,
+                   "File Deletions Disabled, but already disabled. Counter: %d",
+                   disable_delete_obsolete_files_);
+  }
+  return Status::OK();
+}
+
+Status DBImpl::EnableFileDeletions(bool force) {
+  // Job id == 0 means that this is not our background process, but rather
+  // user thread
+  JobContext job_context(0);
+  int saved_counter;  // initialize on all paths
+  {
+    InstrumentedMutexLock l(&mutex_);
+    if (force) {
+      // if force, we need to enable file deletions right away
+      disable_delete_obsolete_files_ = 0;
+    } else if (disable_delete_obsolete_files_ > 0) {
+      --disable_delete_obsolete_files_;
+    }
+    saved_counter = disable_delete_obsolete_files_;
+    if (saved_counter == 0) {
+      FindObsoleteFiles(&job_context, true);
+      bg_cv_.SignalAll();
+    }
+  }
+  if (saved_counter == 0) {
+    ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Enabled");
+    if (job_context.HaveSomethingToDelete()) {
+      PurgeObsoleteFiles(job_context);
+    }
+  } else {
+    ROCKS_LOG_WARN(immutable_db_options_.info_log,
+                   "File Deletions Enable, but not really enabled. Counter: %d",
+                   saved_counter);
+  }
+  job_context.Clean();
+  LogFlush(immutable_db_options_.info_log);
+  return Status::OK();
+}
+
+bool DBImpl::IsFileDeletionsEnabled() const {
+  return 0 == disable_delete_obsolete_files_;
+}
+
 // * Returns the list of live files in 'sst_live' and 'blob_live'.
 // If it's doing full scan:
 // * Returns the list of all files in the filesystem in
@@ -134,7 +190,7 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
       // set of all files in the directory. We'll exclude files that are still
       // alive in the subsequent processings.
       std::vector<std::string> files;
-      env_->GetChildren(path, &files);  // Ignore errors
+      env_->GetChildren(path, &files).PermitUncheckedError();  // Ignore errors
       for (const std::string& file : files) {
         uint64_t number;
         FileType type;
@@ -159,7 +215,8 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
     if (immutable_db_options_.wal_dir != dbname_) {
       std::vector<std::string> log_files;
       env_->GetChildren(immutable_db_options_.wal_dir,
-                        &log_files);  // Ignore errors
+                        &log_files)
+          .PermitUncheckedError();  // Ignore errors
       for (const std::string& log_file : log_files) {
         job_context->full_scan_candidate_files.emplace_back(
             log_file, immutable_db_options_.wal_dir);
@@ -170,7 +227,8 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
         immutable_db_options_.db_log_dir != dbname_) {
       std::vector<std::string> info_log_files;
       // Ignore errors
-      env_->GetChildren(immutable_db_options_.db_log_dir, &info_log_files);
+      env_->GetChildren(immutable_db_options_.db_log_dir, &info_log_files)
+          .PermitUncheckedError();
       for (std::string& log_file : info_log_files) {
         job_context->full_scan_candidate_files.emplace_back(
             log_file, immutable_db_options_.db_log_dir);
@@ -391,7 +449,8 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
   // Close WALs before trying to delete them.
   for (const auto w : state.logs_to_free) {
     // TODO: maybe check the return value of Close.
-    w->Close();
+    auto s = w->Close();
+    s.PermitUncheckedError();
   }
 
   bool own_files = OwnTablesAndLogs();
@@ -510,7 +569,6 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
     if (!own_files) {
       continue;
     }
-    Status file_deletion_status;
     if (schedule_only) {
       InstrumentedMutexLock guard_lock(&mutex_);
       SchedulePendingPurge(fname, dir_to_sync, type, number, state.job_id);
@@ -704,9 +762,13 @@ Status DBImpl::FinishBestEffortsRecovery() {
   uint64_t next_file_number = versions_->current_next_file_number();
   uint64_t largest_file_number = next_file_number;
   std::set<std::string> files_to_delete;
+  Status s;
   for (const auto& path : paths) {
     std::vector<std::string> files;
-    env_->GetChildren(path, &files);
+    s = env_->GetChildren(path, &files);
+    if (!s.ok()) {
+      break;
+    }
     for (const auto& fname : files) {
       uint64_t number = 0;
       FileType type;
@@ -722,6 +784,10 @@ Status DBImpl::FinishBestEffortsRecovery() {
       }
     }
   }
+  if (!s.ok()) {
+    return s;
+  }
+
   if (largest_file_number > next_file_number) {
     versions_->next_file_number_.store(largest_file_number + 1);
   }
@@ -733,7 +799,7 @@ Status DBImpl::FinishBestEffortsRecovery() {
   assert(default_cfd);
   // Even if new_descriptor_log is false, we will still switch to a new
   // MANIFEST and update CURRENT file, since this is in recovery.
-  Status s = versions_->LogAndApply(
+  s = versions_->LogAndApply(
       default_cfd, *default_cfd->GetLatestMutableCFOptions(), &edit, &mutex_,
       directories_.GetDbDir(), /*new_descriptor_log*/ false);
   if (!s.ok()) {

db/db_impl/db_impl_open.cc

@@ -6,20 +6,20 @@
 // Copyright (c) 2011 The LevelDB Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
-#include "db/db_impl/db_impl.h"
-
 #include <cinttypes>
 
 #include "db/builder.h"
+#include "db/db_impl/db_impl.h"
 #include "db/error_handler.h"
+#include "db/periodic_work_scheduler.h"
 #include "env/composite_env_wrapper.h"
 #include "file/read_write_util.h"
 #include "file/sst_file_manager_impl.h"
 #include "file/writable_file_writer.h"
 #include "monitoring/persistent_stats_history.h"
 #include "options/options_helper.h"
+#include "rocksdb/table.h"
 #include "rocksdb/wal_filter.h"
-#include "table/block_based/block_based_table_factory.h"
 #include "test_util/sync_point.h"
 #include "util/rate_limiter.h"
 
@@ -90,12 +90,22 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
   }
 
   if (result.recycle_log_file_num &&
-      (result.wal_recovery_mode == WALRecoveryMode::kPointInTimeRecovery ||
+      (result.wal_recovery_mode ==
+           WALRecoveryMode::kTolerateCorruptedTailRecords ||
+       result.wal_recovery_mode == WALRecoveryMode::kPointInTimeRecovery ||
        result.wal_recovery_mode == WALRecoveryMode::kAbsoluteConsistency)) {
-    // kPointInTimeRecovery is inconsistent with recycle log file feature since
-    // we define the "end" of the log as the first corrupt record we encounter.
-    // kAbsoluteConsistency doesn't make sense because even a clean
-    // shutdown leaves old junk at the end of the log file.
+    // - kTolerateCorruptedTailRecords is inconsistent with recycle log file
+    //   feature. WAL recycling expects recovery success upon encountering a
+    //   corrupt record at the point where new data ends and recycled data
+    //   remains at the tail. However, `kTolerateCorruptedTailRecords` must fail
+    //   upon encountering any such corrupt record, as it cannot differentiate
+    //   between this and a real corruption, which would cause committed updates
+    //   to be truncated -- a violation of the recovery guarantee.
+    // - kPointInTimeRecovery and kAbsoluteConsistency are incompatible with
+    //   recycle log file feature temporarily due to a bug found introducing a
+    //   hole in the recovered data
+    //   (https://github.com/facebook/rocksdb/pull/7252#issuecomment-673766236).
+    //   Besides this bug, we believe the features are fundamentally compatible.
     result.recycle_log_file_num = 0;
   }
 
@@ -137,13 +147,13 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
     // DeleteScheduler::CleanupDirectory on the same dir later, it will be
     // safe
     std::vector<std::string> filenames;
-    result.env->GetChildren(result.wal_dir, &filenames);
+    result.env->GetChildren(result.wal_dir, &filenames).PermitUncheckedError();
     for (std::string& filename : filenames) {
       if (filename.find(".log.trash", filename.length() -
                                           std::string(".log.trash").length()) !=
           std::string::npos) {
         std::string trash_file = result.wal_dir + "/" + filename;
-        result.env->DeleteFile(trash_file);
+        result.env->DeleteFile(trash_file).PermitUncheckedError();
       }
     }
   }
@@ -175,12 +185,12 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src) {
 }
 
 namespace {
-Status SanitizeOptionsByTable(
+Status ValidateOptionsByTable(
     const DBOptions& db_opts,
     const std::vector<ColumnFamilyDescriptor>& column_families) {
   Status s;
   for (auto cf : column_families) {
-    s = cf.options.table_factory->SanitizeOptions(db_opts, cf.options);
+    s = ValidateOptions(db_opts, cf.options);
     if (!s.ok()) {
       return s;
     }
@@ -253,7 +263,7 @@ Status DBImpl::ValidateOptions(const DBOptions& db_options) {
   return Status::OK();
 }
 
-Status DBImpl::NewDB() {
+Status DBImpl::NewDB(std::vector<std::string>* new_filenames) {
   VersionEdit new_db;
   Status s = SetIdentityFile(env_, dbname_);
   if (!s.ok()) {
@@ -280,8 +290,8 @@ Status DBImpl::NewDB() {
     file->SetPreallocationBlockSize(
         immutable_db_options_.manifest_preallocation_size);
     std::unique_ptr<WritableFileWriter> file_writer(new WritableFileWriter(
-        std::move(file), manifest, file_options, env_, nullptr /* stats */,
-        immutable_db_options_.listeners));
+        std::move(file), manifest, file_options, env_, io_tracer_,
+        nullptr /* stats */, immutable_db_options_.listeners));
     log::Writer log(std::move(file_writer), 0, false);
     std::string record;
     new_db.EncodeTo(&record);
@@ -293,6 +303,10 @@ Status DBImpl::NewDB() {
   if (s.ok()) {
     // Make "CURRENT" file that points to the new manifest file.
     s = SetCurrentFile(fs_.get(), dbname_, 1, directories_.GetDbDir());
+    if (new_filenames) {
+      new_filenames->emplace_back(
+          manifest.substr(manifest.find_last_of("/\\") + 1));
+    }
   } else {
     fs_->DeleteFile(manifest, IOOptions(), nullptr);
   }
@@ -350,12 +364,13 @@ IOStatus Directories::SetDirectories(FileSystem* fs, const std::string& dbname,
 
 Status DBImpl::Recover(
     const std::vector<ColumnFamilyDescriptor>& column_families, bool read_only,
-    bool error_if_log_file_exist, bool error_if_data_exists_in_logs,
+    bool error_if_wal_file_exists, bool error_if_data_exists_in_wals,
     uint64_t* recovered_seq) {
   mutex_.AssertHeld();
 
   bool is_new_db = false;
   assert(db_lock_ == nullptr);
+  std::vector<std::string> files_in_dbname;
   if (!read_only) {
     Status s = directories_.SetDirectories(fs_.get(), dbname_,
                                            immutable_db_options_.wal_dir,
@@ -379,10 +394,12 @@ Status DBImpl::Recover(
       s = env_->FileExists(current_fname);
     } else {
       s = Status::NotFound();
-      std::vector<std::string> files;
-      // No need to check return value
-      env_->GetChildren(dbname_, &files);
-      for (const std::string& file : files) {
+      Status io_s = env_->GetChildren(dbname_, &files_in_dbname);
+      if (!io_s.ok()) {
+        s = io_s;
+        files_in_dbname.clear();
+      }
+      for (const std::string& file : files_in_dbname) {
         uint64_t number = 0;
         FileType type = kLogFile;  // initialize
         if (ParseFileName(file, &number, &type) && type == kDescriptorFile) {
@@ -396,7 +413,7 @@ Status DBImpl::Recover(
     }
     if (s.IsNotFound()) {
       if (immutable_db_options_.create_if_missing) {
-        s = NewDB();
+        s = NewDB(&files_in_dbname);
         is_new_db = true;
         if (!s.ok()) {
           return s;
@@ -438,6 +455,16 @@ Status DBImpl::Recover(
         }
       }
     }
+  } else if (immutable_db_options_.best_efforts_recovery) {
+    assert(files_in_dbname.empty());
+    Status s = env_->GetChildren(dbname_, &files_in_dbname);
+    if (s.IsNotFound()) {
+      return Status::InvalidArgument(dbname_,
+                                     "does not exist (open for read only)");
+    } else if (s.IsIOError()) {
+      return s;
+    }
+    assert(s.ok());
   }
   assert(db_id_.empty());
   Status s;
@@ -445,8 +472,9 @@ Status DBImpl::Recover(
   if (!immutable_db_options_.best_efforts_recovery) {
     s = versions_->Recover(column_families, read_only, &db_id_);
   } else {
-    s = versions_->TryRecover(column_families, read_only, &db_id_,
-                              &missing_table_file);
+    assert(!files_in_dbname.empty());
+    s = versions_->TryRecover(column_families, read_only, files_in_dbname,
+                              &db_id_, &missing_table_file);
     if (s.ok()) {
       // TryRecover may delete previous column_family_set_.
       column_family_memtables_.reset(
@@ -506,6 +534,7 @@ Status DBImpl::Recover(
     s = InitPersistStatsColumnFamily();
   }
 
+  std::vector<std::string> files_in_wal_dir;
   if (s.ok()) {
     // Initial max_total_in_memory_state_ before recovery logs. Log recovery
     // may check this value to decide whether to flush.
@@ -532,9 +561,8 @@ Status DBImpl::Recover(
     // Note that prev_log_number() is no longer used, but we pay
     // attention to it in case we are recovering a database
     // produced by an older version of rocksdb.
-    std::vector<std::string> filenames;
     if (!immutable_db_options_.best_efforts_recovery) {
-      s = env_->GetChildren(immutable_db_options_.wal_dir, &filenames);
+      s = env_->GetChildren(immutable_db_options_.wal_dir, &files_in_wal_dir);
     }
     if (s.IsNotFound()) {
       return Status::InvalidArgument("wal_dir not found",
@@ -544,15 +572,15 @@ Status DBImpl::Recover(
     }
 
     std::vector<uint64_t> logs;
-    for (size_t i = 0; i < filenames.size(); i++) {
+    for (const auto& file : files_in_wal_dir) {
       uint64_t number;
       FileType type;
-      if (ParseFileName(filenames[i], &number, &type) && type == kLogFile) {
+      if (ParseFileName(file, &number, &type) && type == kLogFile) {
         if (is_new_db) {
           return Status::Corruption(
               "While creating a new Db, wal_dir contains "
               "existing log file: ",
-              filenames[i]);
+              file);
         } else {
           logs.push_back(number);
         }
@@ -560,11 +588,11 @@ Status DBImpl::Recover(
     }
 
     if (logs.size() > 0) {
-      if (error_if_log_file_exist) {
+      if (error_if_wal_file_exists) {
         return Status::Corruption(
-            "The db was opened in readonly mode with error_if_log_file_exist"
-            "flag but a log file already exists");
-      } else if (error_if_data_exists_in_logs) {
+            "The db was opened in readonly mode with error_if_wal_file_exists"
+            "flag but a WAL file already exists");
+      } else if (error_if_data_exists_in_wals) {
         for (auto& log : logs) {
           std::string fname = LogFileName(immutable_db_options_.wal_dir, log);
           uint64_t bytes;
@@ -572,8 +600,8 @@ Status DBImpl::Recover(
           if (s.ok()) {
             if (bytes > 0) {
               return Status::Corruption(
-                  "error_if_data_exists_in_logs is set but there are data "
-                  " in log files.");
+                  "error_if_data_exists_in_wals is set but there are data "
+                  " in WAL files.");
             }
           }
         }
@@ -604,15 +632,24 @@ Status DBImpl::Recover(
     // to reflect the most recent OPTIONS file. It does not matter for regular
     // read-write db instance because options_file_number_ will later be
     // updated to versions_->NewFileNumber() in RenameTempFileToOptionsFile.
-    std::vector<std::string> file_names;
+    std::vector<std::string> filenames;
     if (s.ok()) {
-      s = env_->GetChildren(GetName(), &file_names);
+      const std::string normalized_dbname = NormalizePath(dbname_);
+      const std::string normalized_wal_dir =
+          NormalizePath(immutable_db_options_.wal_dir);
+      if (immutable_db_options_.best_efforts_recovery) {
+        filenames = std::move(files_in_dbname);
+      } else if (normalized_dbname == normalized_wal_dir) {
+        filenames = std::move(files_in_wal_dir);
+      } else {
+        s = env_->GetChildren(GetName(), &filenames);
+      }
     }
     if (s.ok()) {
       uint64_t number = 0;
       uint64_t options_file_number = 0;
       FileType type;
-      for (const auto& fname : file_names) {
+      for (const auto& fname : filenames) {
         if (ParseFileName(fname, &number, &type) && type == kOptionsFile) {
           options_file_number = std::max(number, options_file_number);
         }
@@ -620,7 +657,6 @@ Status DBImpl::Recover(
       versions_->options_file_number_ = options_file_number;
     }
   }
-
   return s;
 }
 
@@ -649,41 +685,56 @@ Status DBImpl::PersistentStatsProcessFormatVersion() {
         (kStatsCFCurrentFormatVersion < format_version_recovered &&
          kStatsCFCompatibleFormatVersion < compatible_version_recovered)) {
       if (!s_format.ok() || !s_compatible.ok()) {
-        ROCKS_LOG_INFO(
+        ROCKS_LOG_WARN(
             immutable_db_options_.info_log,
-            "Reading persistent stats version key failed. Format key: %s, "
-            "compatible key: %s",
+            "Recreating persistent stats column family since reading "
+            "persistent stats version key failed. Format key: %s, compatible "
+            "key: %s",
             s_format.ToString().c_str(), s_compatible.ToString().c_str());
       } else {
-        ROCKS_LOG_INFO(
+        ROCKS_LOG_WARN(
             immutable_db_options_.info_log,
-            "Disable persistent stats due to corrupted or incompatible format "
-            "version\n");
+            "Recreating persistent stats column family due to corrupted or "
+            "incompatible format version. Recovered format: %" PRIu64
+            "; recovered format compatible since: %" PRIu64 "\n",
+            format_version_recovered, compatible_version_recovered);
+      }
+      s = DropColumnFamily(persist_stats_cf_handle_);
+      if (s.ok()) {
+        s = DestroyColumnFamilyHandle(persist_stats_cf_handle_);
       }
-      DropColumnFamily(persist_stats_cf_handle_);
-      DestroyColumnFamilyHandle(persist_stats_cf_handle_);
       ColumnFamilyHandle* handle = nullptr;
-      ColumnFamilyOptions cfo;
-      OptimizeForPersistentStats(&cfo);
-      s = CreateColumnFamily(cfo, kPersistentStatsColumnFamilyName, &handle);
-      persist_stats_cf_handle_ = static_cast<ColumnFamilyHandleImpl*>(handle);
-      // should also persist version here because old stats CF is discarded
-      should_persist_format_version = true;
+      if (s.ok()) {
+        ColumnFamilyOptions cfo;
+        OptimizeForPersistentStats(&cfo);
+        s = CreateColumnFamily(cfo, kPersistentStatsColumnFamilyName, &handle);
+      }
+      if (s.ok()) {
+        persist_stats_cf_handle_ = static_cast<ColumnFamilyHandleImpl*>(handle);
+        // should also persist version here because old stats CF is discarded
+        should_persist_format_version = true;
+      }
     }
   }
-  if (s.ok() && should_persist_format_version) {
+  if (should_persist_format_version) {
     // Persistent stats CF being created for the first time, need to write
     // format version key
     WriteBatch batch;
-    batch.Put(persist_stats_cf_handle_, kFormatVersionKeyString,
-              ToString(kStatsCFCurrentFormatVersion));
-    batch.Put(persist_stats_cf_handle_, kCompatibleVersionKeyString,
-              ToString(kStatsCFCompatibleFormatVersion));
-    WriteOptions wo;
-    wo.low_pri = true;
-    wo.no_slowdown = true;
-    wo.sync = false;
-    s = Write(wo, &batch);
+    if (s.ok()) {
+      s = batch.Put(persist_stats_cf_handle_, kFormatVersionKeyString,
+                    ToString(kStatsCFCurrentFormatVersion));
+    }
+    if (s.ok()) {
+      s = batch.Put(persist_stats_cf_handle_, kCompatibleVersionKeyString,
+                    ToString(kStatsCFCompatibleFormatVersion));
+    }
+    if (s.ok()) {
+      WriteOptions wo;
+      wo.low_pri = true;
+      wo.no_slowdown = true;
+      wo.sync = false;
+      s = Write(wo, &batch);
+    }
   }
   mutex_.Lock();
   return s;
@@ -825,7 +876,8 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
         }
       }
       file_reader.reset(new SequentialFileReader(
-          std::move(file), fname, immutable_db_options_.log_readahead_size));
+          std::move(file), fname, immutable_db_options_.log_readahead_size,
+          io_tracer_));
     }
 
     // Create the log reader.
@@ -864,7 +916,11 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
                             Status::Corruption("log record too small"));
         continue;
       }
-      WriteBatchInternal::SetContents(&batch, record);
+
+      status = WriteBatchInternal::SetContents(&batch, record);
+      if (!status.ok()) {
+        return status;
+      }
       SequenceNumber sequence = WriteBatchInternal::Sequence(&batch);
 
       if (immutable_db_options_.wal_recovery_mode ==
@@ -1231,7 +1287,10 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
                                            MemTable* mem, VersionEdit* edit) {
   mutex_.AssertHeld();
   const uint64_t start_micros = env_->NowMicros();
+
   FileMetaData meta;
+  std::vector<BlobFileAddition> blob_file_additions;
+
   std::unique_ptr<std::list<uint64_t>::iterator> pending_outputs_inserted_elem(
       new std::list<uint64_t>::iterator(
           CaptureCurrentFileNumberInPendingOutputs()));
@@ -1255,7 +1314,8 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
         cfd->GetLatestMutableCFOptions()->paranoid_file_checks;
 
     int64_t _current_time = 0;
-    env_->GetCurrentTime(&_current_time);  // ignore error
+    env_->GetCurrentTime(&_current_time)
+        .PermitUncheckedError();  // ignore error
     const uint64_t current_time = static_cast<uint64_t>(_current_time);
     meta.oldest_ancester_time = current_time;
 
@@ -1277,20 +1337,23 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
       if (range_del_iter != nullptr) {
         range_del_iters.emplace_back(range_del_iter);
       }
+
       IOStatus io_s;
       s = BuildTable(
-          dbname_, env_, fs_.get(), *cfd->ioptions(), mutable_cf_options,
-          file_options_for_compaction_, cfd->table_cache(), iter.get(),
-          std::move(range_del_iters), &meta, cfd->internal_comparator(),
-          cfd->int_tbl_prop_collector_factories(), cfd->GetID(), cfd->GetName(),
-          snapshot_seqs, earliest_write_conflict_snapshot, snapshot_checker,
+          dbname_, versions_.get(), env_, fs_.get(), *cfd->ioptions(),
+          mutable_cf_options, file_options_for_compaction_, cfd->table_cache(),
+          iter.get(), std::move(range_del_iters), &meta, &blob_file_additions,
+          cfd->internal_comparator(), cfd->int_tbl_prop_collector_factories(),
+          cfd->GetID(), cfd->GetName(), snapshot_seqs,
+          earliest_write_conflict_snapshot, snapshot_checker,
           GetCompressionFlush(*cfd->ioptions(), mutable_cf_options),
           mutable_cf_options.sample_for_compression,
           mutable_cf_options.compression_opts, paranoid_file_checks,
           cfd->internal_stats(), TableFileCreationReason::kRecovery, &io_s,
-          &event_logger_, job_id, Env::IO_HIGH, nullptr /* table_properties */,
-          -1 /* level */, current_time, 0 /* oldest_key_time */, write_hint,
-          0 /* file_creation_time */, db_id_, db_session_id_);
+          io_tracer_, &event_logger_, job_id, Env::IO_HIGH,
+          nullptr /* table_properties */, -1 /* level */, current_time,
+          0 /* oldest_key_time */, write_hint, 0 /* file_creation_time */,
+          db_id_, db_session_id_);
       LogFlush(immutable_db_options_.info_log);
       ROCKS_LOG_DEBUG(immutable_db_options_.info_log,
                       "[%s] [WriteLevel0TableForRecovery]"
@@ -1304,23 +1367,39 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
 
   // Note that if file_size is zero, the file has been deleted and
   // should not be added to the manifest.
-  int level = 0;
-  if (s.ok() && meta.fd.GetFileSize() > 0) {
+  const bool has_output = meta.fd.GetFileSize() > 0;
+  assert(has_output || blob_file_additions.empty());
+
+  constexpr int level = 0;
+
+  if (s.ok() && has_output) {
     edit->AddFile(level, meta.fd.GetNumber(), meta.fd.GetPathId(),
                   meta.fd.GetFileSize(), meta.smallest, meta.largest,
                   meta.fd.smallest_seqno, meta.fd.largest_seqno,
                   meta.marked_for_compaction, meta.oldest_blob_file_number,
                   meta.oldest_ancester_time, meta.file_creation_time,
                   meta.file_checksum, meta.file_checksum_func_name);
+
+    edit->SetBlobFileAdditions(std::move(blob_file_additions));
   }
 
   InternalStats::CompactionStats stats(CompactionReason::kFlush, 1);
   stats.micros = env_->NowMicros() - start_micros;
-  stats.bytes_written = meta.fd.GetFileSize();
-  stats.num_output_files = 1;
+
+  if (has_output) {
+    stats.bytes_written = meta.fd.GetFileSize();
+
+    const auto& blobs = edit->GetBlobFileAdditions();
+    for (const auto& blob : blobs) {
+      stats.bytes_written += blob.GetTotalBlobBytes();
+    }
+
+    stats.num_output_files = static_cast<int>(blobs.size()) + 1;
+  }
+
   cfd->internal_stats()->AddCompactionStats(level, Env::Priority::USER, stats);
   cfd->internal_stats()->AddCFStats(InternalStats::BYTES_FLUSHED,
-                                    meta.fd.GetFileSize());
+                                    stats.bytes_written);
   RecordTick(stats_, COMPACT_WRITE_BYTES, meta.fd.GetFileSize());
   return s;
 }
@@ -1362,9 +1441,10 @@ Status DB::Open(const DBOptions& db_options, const std::string& dbname,
                       !kSeqPerBatch, kBatchPerTxn);
 }
 
-Status DBImpl::CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
-                         size_t preallocate_block_size, log::Writer** new_log) {
-  Status s;
+IOStatus DBImpl::CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
+                           size_t preallocate_block_size,
+                           log::Writer** new_log) {
+  IOStatus io_s;
   std::unique_ptr<FSWritableFile> lfile;
 
   DBOptions db_options =
@@ -1382,32 +1462,32 @@ Status DBImpl::CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
         LogFileName(immutable_db_options_.wal_dir, recycle_log_number);
     TEST_SYNC_POINT("DBImpl::CreateWAL:BeforeReuseWritableFile1");
     TEST_SYNC_POINT("DBImpl::CreateWAL:BeforeReuseWritableFile2");
-    s = fs_->ReuseWritableFile(log_fname, old_log_fname, opt_file_options,
-                               &lfile, /*dbg=*/nullptr);
+    io_s = fs_->ReuseWritableFile(log_fname, old_log_fname, opt_file_options,
+                                  &lfile, /*dbg=*/nullptr);
   } else {
-    s = NewWritableFile(fs_.get(), log_fname, &lfile, opt_file_options);
+    io_s = NewWritableFile(fs_.get(), log_fname, &lfile, opt_file_options);
   }
 
-  if (s.ok()) {
+  if (io_s.ok()) {
     lfile->SetWriteLifeTimeHint(CalculateWALWriteHint());
     lfile->SetPreallocationBlockSize(preallocate_block_size);
 
     const auto& listeners = immutable_db_options_.listeners;
-    std::unique_ptr<WritableFileWriter> file_writer(
-        new WritableFileWriter(std::move(lfile), log_fname, opt_file_options,
-                               env_, nullptr /* stats */, listeners));
+    std::unique_ptr<WritableFileWriter> file_writer(new WritableFileWriter(
+        std::move(lfile), log_fname, opt_file_options, env_, io_tracer_,
+        nullptr /* stats */, listeners));
     *new_log = new log::Writer(std::move(file_writer), log_file_num,
                                immutable_db_options_.recycle_log_file_num > 0,
                                immutable_db_options_.manual_wal_flush);
   }
-  return s;
+  return io_s;
 }
 
 Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
                     const std::vector<ColumnFamilyDescriptor>& column_families,
                     std::vector<ColumnFamilyHandle*>* handles, DB** dbptr,
                     const bool seq_per_batch, const bool batch_per_txn) {
-  Status s = SanitizeOptionsByTable(db_options, column_families);
+  Status s = ValidateOptionsByTable(db_options, column_families);
   if (!s.ok()) {
     return s;
   }
@@ -1501,7 +1581,7 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
               break;
             }
           } else {
-            s = Status::InvalidArgument("Column family not found: ", cf.name);
+            s = Status::InvalidArgument("Column family not found", cf.name);
             break;
           }
         }
@@ -1554,7 +1634,7 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
     }
   }
   if (s.ok() && impl->immutable_db_options_.persist_stats_to_disk) {
-    // try to read format version but no need to fail Open() even if it fails
+    // try to read format version
     s = impl->PersistentStatsProcessFormatVersion();
   }
 
@@ -1598,6 +1678,8 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
     *dbptr = impl;
     impl->opened_successfully_ = true;
     impl->MaybeScheduleFlushOrCompaction();
+  } else {
+    persist_options_status.PermitUncheckedError();
   }
   impl->mutex_.Unlock();
 
@@ -1646,20 +1728,24 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
     paths.erase(std::unique(paths.begin(), paths.end()), paths.end());
     for (auto& path : paths) {
       std::vector<std::string> existing_files;
-      impl->immutable_db_options_.env->GetChildren(path, &existing_files);
+      // TODO: Check for errors here?
+      impl->immutable_db_options_.env->GetChildren(path, &existing_files)
+          .PermitUncheckedError();
       for (auto& file_name : existing_files) {
         uint64_t file_number;
         FileType file_type;
         std::string file_path = path + "/" + file_name;
         if (ParseFileName(file_name, &file_number, &file_type) &&
             file_type == kTableFile) {
+          // TODO: Check for errors from OnAddFile?
           if (known_file_sizes.count(file_name)) {
             // We're assuming that each sst file name exists in at most one of
             // the paths.
             sfm->OnAddFile(file_path, known_file_sizes.at(file_name),
-                           /* compaction */ false);
+                           /* compaction */ false)
+                .PermitUncheckedError();
           } else {
-            sfm->OnAddFile(file_path);
+            sfm->OnAddFile(file_path).PermitUncheckedError();
           }
         }
       }
@@ -1673,6 +1759,7 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
     sfm->ReserveDiskBuffer(max_write_buffer_size,
                            impl->immutable_db_options_.db_paths[0].path);
   }
+
 #endif  // !ROCKSDB_LITE
 
   if (s.ok()) {
@@ -1687,11 +1774,14 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
           "DB::Open() failed --- Unable to persist Options file",
           persist_options_status.ToString());
     }
+  } else {
+    ROCKS_LOG_WARN(impl->immutable_db_options_.info_log,
+                   "Persisting Option File error: %s",
+                   persist_options_status.ToString().c_str());
   }
   if (s.ok()) {
-    impl->StartTimedTasks();
-  }
-  if (!s.ok()) {
+    impl->StartPeriodicWorkScheduler();
+  } else {
     for (auto* h : *handles) {
       delete h;
     }

db/db_impl/db_impl_readonly.cc

@@ -4,13 +4,14 @@
 //  (found in the LICENSE.Apache file in the root directory).
 
 #include "db/db_impl/db_impl_readonly.h"
-#include "db/arena_wrapped_db_iter.h"
 
+#include "db/arena_wrapped_db_iter.h"
 #include "db/compacted_db_impl.h"
 #include "db/db_impl/db_impl.h"
 #include "db/db_iter.h"
 #include "db/merge_context.h"
 #include "monitoring/perf_context_imp.h"
+#include "util/cast_util.h"
 
 namespace ROCKSDB_NAMESPACE {
 
@@ -35,7 +36,7 @@ Status DBImplReadOnly::Get(const ReadOptions& read_options,
   PERF_TIMER_GUARD(get_snapshot_time);
   Status s;
   SequenceNumber snapshot = versions_->LastSequence();
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   auto cfd = cfh->cfd();
   if (tracer_) {
     InstrumentedMutexLock lock(&trace_mutex_);
@@ -70,7 +71,7 @@ Status DBImplReadOnly::Get(const ReadOptions& read_options,
 
 Iterator* DBImplReadOnly::NewIterator(const ReadOptions& read_options,
                                       ColumnFamilyHandle* column_family) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   auto cfd = cfh->cfd();
   SuperVersion* super_version = cfd->GetSuperVersion()->Ref();
   SequenceNumber latest_snapshot = versions_->LastSequence();
@@ -85,10 +86,10 @@ Iterator* DBImplReadOnly::NewIterator(const ReadOptions& read_options,
       read_seq,
       super_version->mutable_cf_options.max_sequential_skip_in_iterations,
       super_version->version_number, read_callback);
-  auto internal_iter =
-      NewInternalIterator(read_options, cfd, super_version, db_iter->GetArena(),
-                          db_iter->GetRangeDelAggregator(), read_seq,
-                          /* allow_unprepared_value */ true);
+  auto internal_iter = NewInternalIterator(
+      db_iter->GetReadOptions(), cfd, super_version, db_iter->GetArena(),
+      db_iter->GetRangeDelAggregator(), read_seq,
+      /* allow_unprepared_value */ true);
   db_iter->SetIterUnderDBIter(internal_iter);
   return db_iter;
 }
@@ -111,16 +112,16 @@ Status DBImplReadOnly::NewIterators(
           : latest_snapshot;
 
   for (auto cfh : column_families) {
-    auto* cfd = reinterpret_cast<ColumnFamilyHandleImpl*>(cfh)->cfd();
+    auto* cfd = static_cast_with_check<ColumnFamilyHandleImpl>(cfh)->cfd();
     auto* sv = cfd->GetSuperVersion()->Ref();
     auto* db_iter = NewArenaWrappedDbIterator(
         env_, read_options, *cfd->ioptions(), sv->mutable_cf_options, read_seq,
         sv->mutable_cf_options.max_sequential_skip_in_iterations,
         sv->version_number, read_callback);
-    auto* internal_iter =
-        NewInternalIterator(read_options, cfd, sv, db_iter->GetArena(),
-                            db_iter->GetRangeDelAggregator(), read_seq,
-                            /* allow_unprepared_value */ true);
+    auto* internal_iter = NewInternalIterator(
+        db_iter->GetReadOptions(), cfd, sv, db_iter->GetArena(),
+        db_iter->GetRangeDelAggregator(), read_seq,
+        /* allow_unprepared_value */ true);
     db_iter->SetIterUnderDBIter(internal_iter);
     iterators->push_back(db_iter);
   }
@@ -150,7 +151,7 @@ Status OpenForReadOnlyCheckExistence(const DBOptions& db_options,
 }  // namespace
 
 Status DB::OpenForReadOnly(const Options& options, const std::string& dbname,
-                           DB** dbptr, bool /*error_if_log_file_exist*/) {
+                           DB** dbptr, bool /*error_if_wal_file_exists*/) {
   // If dbname does not exist in the file system, should not do anything
   Status s = OpenForReadOnlyCheckExistence(options, dbname);
   if (!s.ok()) {
@@ -187,7 +188,7 @@ Status DB::OpenForReadOnly(
     const DBOptions& db_options, const std::string& dbname,
     const std::vector<ColumnFamilyDescriptor>& column_families,
     std::vector<ColumnFamilyHandle*>* handles, DB** dbptr,
-    bool error_if_log_file_exist) {
+    bool error_if_wal_file_exists) {
   // If dbname does not exist in the file system, should not do anything
   Status s = OpenForReadOnlyCheckExistence(db_options, dbname);
   if (!s.ok()) {
@@ -196,14 +197,14 @@ Status DB::OpenForReadOnly(
 
   return DBImplReadOnly::OpenForReadOnlyWithoutCheck(
       db_options, dbname, column_families, handles, dbptr,
-      error_if_log_file_exist);
+      error_if_wal_file_exists);
 }
 
 Status DBImplReadOnly::OpenForReadOnlyWithoutCheck(
     const DBOptions& db_options, const std::string& dbname,
     const std::vector<ColumnFamilyDescriptor>& column_families,
     std::vector<ColumnFamilyHandle*>* handles, DB** dbptr,
-    bool error_if_log_file_exist) {
+    bool error_if_wal_file_exists) {
   *dbptr = nullptr;
   handles->clear();
 
@@ -211,14 +212,14 @@ Status DBImplReadOnly::OpenForReadOnlyWithoutCheck(
   DBImplReadOnly* impl = new DBImplReadOnly(db_options, dbname);
   impl->mutex_.Lock();
   Status s = impl->Recover(column_families, true /* read only */,
-                           error_if_log_file_exist);
+                           error_if_wal_file_exists);
   if (s.ok()) {
     // set column family handles
     for (auto cf : column_families) {
       auto cfd =
           impl->versions_->GetColumnFamilySet()->GetColumnFamily(cf.name);
       if (cfd == nullptr) {
-        s = Status::InvalidArgument("Column family not found: ", cf.name);
+        s = Status::InvalidArgument("Column family not found", cf.name);
         break;
       }
       handles->push_back(new ColumnFamilyHandleImpl(cfd, impl, &impl->mutex_));
@@ -236,7 +237,7 @@ Status DBImplReadOnly::OpenForReadOnlyWithoutCheck(
     *dbptr = impl;
     for (auto* h : *handles) {
       impl->NewThreadStatusCfInfo(
-          reinterpret_cast<ColumnFamilyHandleImpl*>(h)->cfd());
+          static_cast_with_check<ColumnFamilyHandleImpl>(h)->cfd());
     }
   } else {
     for (auto h : *handles) {
@@ -252,7 +253,7 @@ Status DBImplReadOnly::OpenForReadOnlyWithoutCheck(
 
 Status DB::OpenForReadOnly(const Options& /*options*/,
                            const std::string& /*dbname*/, DB** /*dbptr*/,
-                           bool /*error_if_log_file_exist*/) {
+                           bool /*error_if_wal_file_exists*/) {
   return Status::NotSupported("Not supported in ROCKSDB_LITE.");
 }
 
@@ -260,7 +261,7 @@ Status DB::OpenForReadOnly(
     const DBOptions& /*db_options*/, const std::string& /*dbname*/,
     const std::vector<ColumnFamilyDescriptor>& /*column_families*/,
     std::vector<ColumnFamilyHandle*>* /*handles*/, DB** /*dbptr*/,
-    bool /*error_if_log_file_exist*/) {
+    bool /*error_if_wal_file_exists*/) {
   return Status::NotSupported("Not supported in ROCKSDB_LITE.");
 }
 #endif  // !ROCKSDB_LITE

db/db_impl/db_impl_readonly.h

@@ -138,7 +138,7 @@ class DBImplReadOnly : public DBImpl {
       const DBOptions& db_options, const std::string& dbname,
       const std::vector<ColumnFamilyDescriptor>& column_families,
       std::vector<ColumnFamilyHandle*>* handles, DB** dbptr,
-      bool error_if_log_file_exist = false);
+      bool error_if_wal_file_exists = false);
   friend class DB;
 };
 }  // namespace ROCKSDB_NAMESPACE

db/db_impl/db_impl_secondary.cc

@@ -28,8 +28,8 @@ DBImplSecondary::~DBImplSecondary() {}
 
 Status DBImplSecondary::Recover(
     const std::vector<ColumnFamilyDescriptor>& column_families,
-    bool /*readonly*/, bool /*error_if_log_file_exist*/,
-    bool /*error_if_data_exists_in_logs*/, uint64_t*) {
+    bool /*readonly*/, bool /*error_if_wal_file_exists*/,
+    bool /*error_if_data_exists_in_wals*/, uint64_t*) {
   mutex_.AssertHeld();
 
   JobContext job_context(0);
@@ -153,7 +153,8 @@ Status DBImplSecondary::MaybeInitLogReader(
         return status;
       }
       file_reader.reset(new SequentialFileReader(
-          std::move(file), fname, immutable_db_options_.log_readahead_size));
+          std::move(file), fname, immutable_db_options_.log_readahead_size,
+          io_tracer_));
     }
 
     // Create the log reader.
@@ -191,6 +192,8 @@ Status DBImplSecondary::RecoverLogFiles(
     auto it  = log_readers_.find(log_number);
     assert(it != log_readers_.end());
     log::FragmentBufferedReader* reader = it->second->reader_;
+    Status* wal_read_status = it->second->status_;
+    assert(wal_read_status);
     // Manually update the file number allocation counter in VersionSet.
     versions_->MarkFileNumberUsed(log_number);
 
@@ -202,13 +205,16 @@ Status DBImplSecondary::RecoverLogFiles(
 
     while (reader->ReadRecord(&record, &scratch,
                               immutable_db_options_.wal_recovery_mode) &&
-           status.ok()) {
+           wal_read_status->ok() && status.ok()) {
       if (record.size() < WriteBatchInternal::kHeader) {
         reader->GetReporter()->Corruption(
             record.size(), Status::Corruption("log record too small"));
         continue;
       }
-      WriteBatchInternal::SetContents(&batch, record);
+      status = WriteBatchInternal::SetContents(&batch, record);
+      if (!status.ok()) {
+        break;
+      }
       SequenceNumber seq_of_batch = WriteBatchInternal::Sequence(&batch);
       std::vector<uint32_t> column_family_ids;
       status = CollectColumnFamilyIdsFromWriteBatch(batch, &column_family_ids);
@@ -294,6 +300,9 @@ Status DBImplSecondary::RecoverLogFiles(
         reader->GetReporter()->Corruption(record.size(), status);
       }
     }
+    if (status.ok() && !wal_read_status->ok()) {
+      status = *wal_read_status;
+    }
     if (!status.ok()) {
       return status;
     }
@@ -388,7 +397,7 @@ Iterator* DBImplSecondary::NewIterator(const ReadOptions& read_options,
         "ReadTier::kPersistedData is not yet supported in iterators."));
   }
   Iterator* result = nullptr;
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   auto cfd = cfh->cfd();
   ReadCallback* read_callback = nullptr;  // No read callback provided.
   if (read_options.tailing) {
@@ -415,10 +424,10 @@ ArenaWrappedDBIter* DBImplSecondary::NewIteratorImpl(
       snapshot,
       super_version->mutable_cf_options.max_sequential_skip_in_iterations,
       super_version->version_number, read_callback);
-  auto internal_iter =
-      NewInternalIterator(read_options, cfd, super_version, db_iter->GetArena(),
-                          db_iter->GetRangeDelAggregator(), snapshot,
-                          /* allow_unprepared_value */ true);
+  auto internal_iter = NewInternalIterator(
+      db_iter->GetReadOptions(), cfd, super_version, db_iter->GetArena(),
+      db_iter->GetRangeDelAggregator(), snapshot,
+      /* allow_unprepared_value */ true);
   db_iter->SetIterUnderDBIter(internal_iter);
   return db_iter;
 }
@@ -611,7 +620,7 @@ Status DB::OpenAsSecondary(
   impl->versions_.reset(new ReactiveVersionSet(
       dbname, &impl->immutable_db_options_, impl->file_options_,
       impl->table_cache_.get(), impl->write_buffer_manager_,
-      &impl->write_controller_));
+      &impl->write_controller_, impl->io_tracer_));
   impl->column_family_memtables_.reset(
       new ColumnFamilyMemTablesImpl(impl->versions_->GetColumnFamilySet()));
   impl->wal_in_db_path_ = IsWalDirSameAsDBPath(&impl->immutable_db_options_);
@@ -623,7 +632,7 @@ Status DB::OpenAsSecondary(
       auto cfd =
           impl->versions_->GetColumnFamilySet()->GetColumnFamily(cf.name);
       if (nullptr == cfd) {
-        s = Status::InvalidArgument("Column family not found: ", cf.name);
+        s = Status::InvalidArgument("Column family not found", cf.name);
         break;
       }
       handles->push_back(new ColumnFamilyHandleImpl(cfd, impl, &impl->mutex_));
@@ -642,7 +651,7 @@ Status DB::OpenAsSecondary(
     *dbptr = impl;
     for (auto h : *handles) {
       impl->NewThreadStatusCfInfo(
-          reinterpret_cast<ColumnFamilyHandleImpl*>(h)->cfd());
+          static_cast_with_check<ColumnFamilyHandleImpl>(h)->cfd());
     }
   } else {
     for (auto h : *handles) {

db/db_impl/db_impl_secondary.h

@@ -77,8 +77,8 @@ class DBImplSecondary : public DBImpl {
   // Recover by replaying MANIFEST and WAL. Also initialize manifest_reader_
   // and log_readers_ to facilitate future operations.
   Status Recover(const std::vector<ColumnFamilyDescriptor>& column_families,
-                 bool read_only, bool error_if_log_file_exist,
-                 bool error_if_data_exists_in_logs,
+                 bool read_only, bool error_if_wal_file_exists,
+                 bool error_if_data_exists_in_wals,
                  uint64_t* = nullptr) override;
 
   // Implementations of the DB interface

db/db_impl/db_impl_write.cc

@@ -6,14 +6,15 @@
 // Copyright (c) 2011 The LevelDB Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file. See the AUTHORS file for names of contributors.
-#include "db/db_impl/db_impl.h"
-
 #include <cinttypes>
+
+#include "db/db_impl/db_impl.h"
 #include "db/error_handler.h"
 #include "db/event_helpers.h"
 #include "monitoring/perf_context_imp.h"
 #include "options/options_helper.h"
 #include "test_util/sync_point.h"
+#include "util/cast_util.h"
 
 namespace ROCKSDB_NAMESPACE {
 // Convenience methods
@@ -24,7 +25,7 @@ Status DBImpl::Put(const WriteOptions& o, ColumnFamilyHandle* column_family,
 
 Status DBImpl::Merge(const WriteOptions& o, ColumnFamilyHandle* column_family,
                      const Slice& key, const Slice& val) {
-  auto cfh = reinterpret_cast<ColumnFamilyHandleImpl*>(column_family);
+  auto cfh = static_cast_with_check<ColumnFamilyHandleImpl>(column_family);
   if (!cfh->cfd()->ioptions()->merge_operator) {
     return Status::NotSupported("Provide a merge_operator when opening DB");
   } else {
@@ -76,7 +77,8 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
   if (tracer_) {
     InstrumentedMutexLock lock(&trace_mutex_);
     if (tracer_) {
-      tracer_->Write(my_batch);
+      // TODO: maybe handle the tracing status?
+      tracer_->Write(my_batch).PermitUncheckedError();
     }
   }
   if (write_options.sync && write_options.disableWAL) {
@@ -100,12 +102,10 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
   assert(!WriteBatchInternal::IsLatestPersistentState(my_batch) ||
          disable_memtable);
 
-  Status status;
-  IOStatus io_s;
   if (write_options.low_pri) {
-    status = ThrottleLowPriWritesIfNeeded(write_options, my_batch);
-    if (!status.ok()) {
-      return status;
+    Status s = ThrottleLowPriWritesIfNeeded(write_options, my_batch);
+    if (!s.ok()) {
+      return s;
     }
   }
 
@@ -125,13 +125,13 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
                                      ? batch_cnt
                                      // every key is a sub-batch consuming a seq
                                      : WriteBatchInternal::Count(my_batch);
-    uint64_t seq;
+    uint64_t seq = 0;
     // Use a write thread to i) optimize for WAL write, ii) publish last
     // sequence in in increasing order, iii) call pre_release_callback serially
-    status = WriteImplWALOnly(&write_thread_, write_options, my_batch, callback,
-                              log_used, log_ref, &seq, sub_batch_cnt,
-                              pre_release_callback, kDoAssignOrder,
-                              kDoPublishLastSeq, disable_memtable);
+    Status status = WriteImplWALOnly(
+        &write_thread_, write_options, my_batch, callback, log_used, log_ref,
+        &seq, sub_batch_cnt, pre_release_callback, kDoAssignOrder,
+        kDoPublishLastSeq, disable_memtable);
     TEST_SYNC_POINT("DBImpl::WriteImpl:UnorderedWriteAfterWriteWAL");
     if (!status.ok()) {
       return status;
@@ -163,6 +163,7 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
   StopWatch write_sw(env_, immutable_db_options_.statistics.get(), DB_WRITE);
 
   write_thread_.JoinBatchGroup(&w);
+  Status status;
   if (w.state == WriteThread::STATE_PARALLEL_MEMTABLE_WRITER) {
     // we are a non-leader in a parallel group
 
@@ -203,6 +204,8 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
       *seq_used = w.sequence;
     }
     // write is complete and leader has updated sequence
+    // Should we handle it?
+    status.PermitUncheckedError();
     return w.FinalStatus();
   }
   // else we are the leader of the write batch group
@@ -251,6 +254,7 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
   last_batch_group_size_ =
       write_thread_.EnterAsBatchGroupLeader(&w, &write_group);
 
+  IOStatus io_s;
   if (status.ok()) {
     // Rules for when we can update the memtable concurrently
     // 1. supported by memtable
@@ -469,6 +473,7 @@ Status DBImpl::PipelinedWriteImpl(const WriteOptions& write_options,
   WriteThread::Writer w(write_options, my_batch, callback, log_ref,
                         disable_memtable);
   write_thread_.JoinBatchGroup(&w);
+  TEST_SYNC_POINT("DBImplWrite::PipelinedWriteImpl:AfterJoinBatchGroup");
   if (w.state == WriteThread::STATE_GROUP_LEADER) {
     WriteThread::WriteGroup wal_write_group;
     if (w.callback && !w.callback->AllowWriteBatching()) {
@@ -622,8 +627,6 @@ Status DBImpl::UnorderedWriteMemtable(const WriteOptions& write_options,
         0 /*log_number*/, this, true /*concurrent_memtable_writes*/,
         seq_per_batch_, sub_batch_cnt, true /*batch_per_txn*/,
         write_options.memtable_insert_hint_per_batch);
-
-    WriteStatusCheck(w.status);
     if (write_options.disableWAL) {
       has_unpersisted_data_.store(true, std::memory_order_relaxed);
     }
@@ -638,6 +641,7 @@ Status DBImpl::UnorderedWriteMemtable(const WriteOptions& write_options,
     std::lock_guard<std::mutex> lck(switch_mutex_);
     switch_cv_.notify_all();
   }
+  WriteStatusCheck(w.status);
 
   if (!w.FinalStatus().ok()) {
     return w.FinalStatus();
@@ -688,7 +692,7 @@ Status DBImpl::WriteImplWALOnly(
       InstrumentedMutexLock l(&mutex_);
       bool need_log_sync = false;
       status = PreprocessWrite(write_options, &need_log_sync, &write_context);
-      WriteStatusCheck(status);
+      WriteStatusCheckOnLocked(status);
     }
     if (!status.ok()) {
       WriteThread::WriteGroup write_group;
@@ -829,13 +833,30 @@ Status DBImpl::WriteImplWALOnly(
   return status;
 }
 
+void DBImpl::WriteStatusCheckOnLocked(const Status& status) {
+  // Is setting bg_error_ enough here?  This will at least stop
+  // compaction and fail any further writes.
+  // Caller must hold mutex_.
+  assert(!status.IsIOFenced() || !error_handler_.GetBGError().ok());
+  mutex_.AssertHeld();
+  if (immutable_db_options_.paranoid_checks && !status.ok() &&
+      !status.IsBusy() && !status.IsIncomplete()) {
+    // Maybe change the return status to void?
+    error_handler_.SetBGError(status, BackgroundErrorReason::kWriteCallback)
+        .PermitUncheckedError();
+  }
+}
+
 void DBImpl::WriteStatusCheck(const Status& status) {
   // Is setting bg_error_ enough here?  This will at least stop
   // compaction and fail any further writes.
+  assert(!status.IsIOFenced() || !error_handler_.GetBGError().ok());
   if (immutable_db_options_.paranoid_checks && !status.ok() &&
       !status.IsBusy() && !status.IsIncomplete()) {
     mutex_.Lock();
-    error_handler_.SetBGError(status, BackgroundErrorReason::kWriteCallback);
+    // Maybe change the return status to void?
+    error_handler_.SetBGError(status, BackgroundErrorReason::kWriteCallback)
+        .PermitUncheckedError();
     mutex_.Unlock();
   }
 }
@@ -843,10 +864,13 @@ void DBImpl::WriteStatusCheck(const Status& status) {
 void DBImpl::IOStatusCheck(const IOStatus& io_status) {
   // Is setting bg_error_ enough here?  This will at least stop
   // compaction and fail any further writes.
-  if (immutable_db_options_.paranoid_checks && !io_status.ok() &&
-      !io_status.IsBusy() && !io_status.IsIncomplete()) {
+  if ((immutable_db_options_.paranoid_checks && !io_status.ok() &&
+       !io_status.IsBusy() && !io_status.IsIncomplete()) ||
+      io_status.IsIOFenced()) {
     mutex_.Lock();
-    error_handler_.SetBGError(io_status, BackgroundErrorReason::kWriteCallback);
+    // Maybe change the return status to void?
+    error_handler_.SetBGError(io_status, BackgroundErrorReason::kWriteCallback)
+        .PermitUncheckedError();
     mutex_.Unlock();
   }
 }
@@ -860,7 +884,9 @@ void DBImpl::MemTableInsertStatusCheck(const Status& status) {
   if (!status.ok()) {
     mutex_.Lock();
     assert(!error_handler_.IsBGWorkStopped());
-    error_handler_.SetBGError(status, BackgroundErrorReason::kMemTable);
+    // Maybe change the return status to void?
+    error_handler_.SetBGError(status, BackgroundErrorReason::kMemTable)
+        .PermitUncheckedError();
     mutex_.Unlock();
   }
 }
@@ -974,8 +1000,10 @@ WriteBatch* DBImpl::MergeBatch(const WriteThread::WriteGroup& write_group,
     merged_batch = tmp_batch;
     for (auto writer : write_group) {
       if (!writer->CallbackFailed()) {
-        WriteBatchInternal::Append(merged_batch, writer->batch,
-                                   /*WAL_only*/ true);
+        Status s = WriteBatchInternal::Append(merged_batch, writer->batch,
+                                              /*WAL_only*/ true);
+        // Always returns Status::OK.
+        assert(s.ok());
         if (WriteBatchInternal::IsLatestPersistentState(writer->batch)) {
           // We only need to cache the last of such write batch
           *to_be_cached_state = writer->batch;
@@ -1545,11 +1573,14 @@ Status DBImpl::TrimMemtableHistory(WriteContext* context) {
   }
   for (auto& cfd : cfds) {
     autovector<MemTable*> to_delete;
-    cfd->imm()->TrimHistory(&to_delete, cfd->mem()->ApproximateMemoryUsage());
+    bool trimmed = cfd->imm()->TrimHistory(
+        &to_delete, cfd->mem()->ApproximateMemoryUsage());
     if (!to_delete.empty()) {
       for (auto m : to_delete) {
         delete m;
       }
+    }
+    if (trimmed) {
       context->superversion_context.NewSuperVersion();
       assert(context->superversion_context.new_superversion.get() != nullptr);
       cfd->InstallSuperVersion(&context->superversion_context, &mutex_);
@@ -1637,6 +1668,7 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
   std::unique_ptr<WritableFile> lfile;
   log::Writer* new_log = nullptr;
   MemTable* new_mem = nullptr;
+  IOStatus io_s;
 
   // Recoverable state is persisted in WAL. After memtable switch, WAL might
   // be deleted, so we write the state to memtable to be persisted as well.
@@ -1682,8 +1714,11 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
   if (creating_new_log) {
     // TODO: Write buffer size passed in should be max of all CF's instead
     // of mutable_cf_options.write_buffer_size.
-    s = CreateWAL(new_log_number, recycle_log_number, preallocate_block_size,
-                  &new_log);
+    io_s = CreateWAL(new_log_number, recycle_log_number, preallocate_block_size,
+                     &new_log);
+    if (s.ok()) {
+      s = io_s;
+    }
   }
   if (s.ok()) {
     SequenceNumber seq = versions_->LastSequence();
@@ -1709,7 +1744,10 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
     if (!logs_.empty()) {
       // Alway flush the buffer of the last log before switching to a new one
       log::Writer* cur_log_writer = logs_.back().writer;
-      s = cur_log_writer->WriteBuffer();
+      io_s = cur_log_writer->WriteBuffer();
+      if (s.ok()) {
+        s = io_s;
+      }
       if (!s.ok()) {
         ROCKS_LOG_WARN(immutable_db_options_.info_log,
                        "[%s] Failed to switch from #%" PRIu64 " to #%" PRIu64
@@ -1744,7 +1782,16 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
     }
     // We may have lost data from the WritableFileBuffer in-memory buffer for
     // the current log, so treat it as a fatal error and set bg_error
-    error_handler_.SetBGError(s, BackgroundErrorReason::kMemTable);
+    // Should handle return error?
+    if (!io_s.ok()) {
+      // Should handle return error?
+      error_handler_.SetBGError(io_s, BackgroundErrorReason::kMemTable)
+          .PermitUncheckedError();
+    } else {
+      // Should handle return error?
+      error_handler_.SetBGError(s, BackgroundErrorReason::kMemTable)
+          .PermitUncheckedError();
+    }
     // Read back bg_error in order to get the right severity
     s = error_handler_.GetBGError();
     return s;
@@ -1777,6 +1824,10 @@ Status DBImpl::SwitchMemtable(ColumnFamilyData* cfd, WriteContext* context) {
   NotifyOnMemTableSealed(cfd, memtable_info);
   mutex_.Lock();
 #endif  // ROCKSDB_LITE
+  // It is possible that we got here without checking the value of i_os, but
+  // that is okay.  If we did, it most likely means that s was already an error.
+  // In any case, ignore any unchecked error for i_os here.
+  io_s.PermitUncheckedError();
   return s;
 }
 
@@ -1867,7 +1918,10 @@ Status DB::Delete(const WriteOptions& opt, ColumnFamilyHandle* column_family,
 Status DB::SingleDelete(const WriteOptions& opt,
                         ColumnFamilyHandle* column_family, const Slice& key) {
   WriteBatch batch;
-  batch.SingleDelete(column_family, key);
+  Status s = batch.SingleDelete(column_family, key);
+  if (!s.ok()) {
+    return s;
+  }
   return Write(opt, &batch);
 }
 
@@ -1875,7 +1929,10 @@ Status DB::DeleteRange(const WriteOptions& opt,
                        ColumnFamilyHandle* column_family,
                        const Slice& begin_key, const Slice& end_key) {
   WriteBatch batch;
-  batch.DeleteRange(column_family, begin_key, end_key);
+  Status s = batch.DeleteRange(column_family, begin_key, end_key);
+  if (!s.ok()) {
+    return s;
+  }
   return Write(opt, &batch);
 }
 

db/db_impl/db_secondary_test.cc

@@ -10,8 +10,8 @@
 #include "db/db_impl/db_impl_secondary.h"
 #include "db/db_test_util.h"
 #include "port/stack_trace.h"
-#include "test_util/fault_injection_test_env.h"
 #include "test_util/sync_point.h"
+#include "utilities/fault_injection_env.h"
 
 namespace ROCKSDB_NAMESPACE {
 
@@ -19,7 +19,7 @@ namespace ROCKSDB_NAMESPACE {
 class DBSecondaryTest : public DBTestBase {
  public:
   DBSecondaryTest()
-      : DBTestBase("/db_secondary_test"),
+      : DBTestBase("/db_secondary_test", /*env_do_fsync=*/true),
         secondary_path_(),
         handles_secondary_(),
         db_secondary_(nullptr) {
@@ -52,7 +52,7 @@ class DBSecondaryTest : public DBTestBase {
 
   void CloseSecondary() {
     for (auto h : handles_secondary_) {
-      db_secondary_->DestroyColumnFamilyHandle(h);
+      ASSERT_OK(db_secondary_->DestroyColumnFamilyHandle(h));
     }
     handles_secondary_.clear();
     delete db_secondary_;
@@ -97,7 +97,7 @@ void DBSecondaryTest::CheckFileTypeCounts(const std::string& dir,
                                           int expected_log, int expected_sst,
                                           int expected_manifest) const {
   std::vector<std::string> filenames;
-  env_->GetChildren(dir, &filenames);
+  ASSERT_OK(env_->GetChildren(dir, &filenames));
 
   int log_cnt = 0, sst_cnt = 0, manifest_cnt = 0;
   for (auto file : filenames) {
@@ -777,8 +777,8 @@ TEST_F(DBSecondaryTest, CatchUpAfterFlush) {
 
   WriteOptions write_opts;
   WriteBatch wb;
-  wb.Put("key0", "value0");
-  wb.Put("key1", "value1");
+  ASSERT_OK(wb.Put("key0", "value0"));
+  ASSERT_OK(wb.Put("key1", "value1"));
   ASSERT_OK(dbfull()->Write(write_opts, &wb));
   ReadOptions read_opts;
   std::unique_ptr<Iterator> iter1(db_secondary_->NewIterator(read_opts));
@@ -791,25 +791,27 @@ TEST_F(DBSecondaryTest, CatchUpAfterFlush) {
   ASSERT_FALSE(iter1->Valid());
   iter1->Seek("key1");
   ASSERT_FALSE(iter1->Valid());
+  ASSERT_OK(iter1->status());
   std::unique_ptr<Iterator> iter2(db_secondary_->NewIterator(read_opts));
   iter2->Seek("key0");
   ASSERT_TRUE(iter2->Valid());
   ASSERT_EQ("value0", iter2->value());
   iter2->Seek("key1");
   ASSERT_TRUE(iter2->Valid());
+  ASSERT_OK(iter2->status());
   ASSERT_EQ("value1", iter2->value());
 
   {
     WriteBatch wb1;
-    wb1.Put("key0", "value01");
-    wb1.Put("key1", "value11");
+    ASSERT_OK(wb1.Put("key0", "value01"));
+    ASSERT_OK(wb1.Put("key1", "value11"));
     ASSERT_OK(dbfull()->Write(write_opts, &wb1));
   }
 
   {
     WriteBatch wb2;
-    wb2.Put("key0", "new_value0");
-    wb2.Delete("key1");
+    ASSERT_OK(wb2.Put("key0", "new_value0"));
+    ASSERT_OK(wb2.Delete("key1"));
     ASSERT_OK(dbfull()->Write(write_opts, &wb2));
   }
 
@@ -823,6 +825,7 @@ TEST_F(DBSecondaryTest, CatchUpAfterFlush) {
   ASSERT_EQ("new_value0", iter3->value());
   iter3->Seek("key1");
   ASSERT_FALSE(iter3->Valid());
+  ASSERT_OK(iter3->status());
 }
 
 TEST_F(DBSecondaryTest, CheckConsistencyWhenOpen) {

db/db_info_dumper.cc

@@ -53,16 +53,24 @@ void DumpDBFileSummary(const ImmutableDBOptions& options,
         Header(options.info_log, "IDENTITY file:  %s\n", file.c_str());
         break;
       case kDescriptorFile:
-        env->GetFileSize(dbname + "/" + file, &file_size);
-        Header(options.info_log, "MANIFEST file:  %s size: %" PRIu64 " Bytes\n",
-               file.c_str(), file_size);
+        if (env->GetFileSize(dbname + "/" + file, &file_size).ok()) {
+          Header(options.info_log,
+                 "MANIFEST file:  %s size: %" PRIu64 " Bytes\n", file.c_str(),
+                 file_size);
+        } else {
+          Error(options.info_log, "Error when reading MANIFEST file: %s/%s\n",
+                dbname.c_str(), file.c_str());
+        }
         break;
       case kLogFile:
-        env->GetFileSize(dbname + "/" + file, &file_size);
-        char str[16];
-        snprintf(str, sizeof(str), "%" PRIu64, file_size);
-        wal_info.append(file).append(" size: ").
-            append(str).append(" ; ");
+        if (env->GetFileSize(dbname + "/" + file, &file_size).ok()) {
+          char str[16];
+          snprintf(str, sizeof(str), "%" PRIu64, file_size);
+          wal_info.append(file).append(" size: ").append(str).append(" ; ");
+        } else {
+          Error(options.info_log, "Error when reading LOG file: %s/%s\n",
+                dbname.c_str(), file.c_str());
+        }
         break;
       case kTableFile:
         if (++file_num < 10) {
@@ -111,11 +119,14 @@ void DumpDBFileSummary(const ImmutableDBOptions& options,
     for (const std::string& file : files) {
       if (ParseFileName(file, &number, &type)) {
         if (type == kLogFile) {
-          env->GetFileSize(options.wal_dir + "/" + file, &file_size);
-          char str[16];
-          snprintf(str, sizeof(str), "%" PRIu64, file_size);
-          wal_info.append(file).append(" size: ").
-              append(str).append(" ; ");
+          if (env->GetFileSize(options.wal_dir + "/" + file, &file_size).ok()) {
+            char str[16];
+            snprintf(str, sizeof(str), "%" PRIu64, file_size);
+            wal_info.append(file).append(" size: ").append(str).append(" ; ");
+          } else {
+            Error(options.info_log, "Error when reading LOG file %s/%s\n",
+                  options.wal_dir.c_str(), file.c_str());
+          }
         }
       }
     }

db/db_inplace_update_test.cc

@@ -13,7 +13,8 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBTestInPlaceUpdate : public DBTestBase {
  public:
-  DBTestInPlaceUpdate() : DBTestBase("/db_inplace_update_test") {}
+  DBTestInPlaceUpdate()
+      : DBTestBase("/db_inplace_update_test", /*env_do_fsync=*/true) {}
 };
 
 TEST_F(DBTestInPlaceUpdate, InPlaceUpdate) {