bump version and update HISTORY.md for 6.15.5

Summary:
Explicitly reject all range deletions on `TransactionDB` or `OptimisticTransactionDB`, except when the user provides sufficient promises that allow us to proceed safely. The necessary promises are described in the API doc for `TransactionDB::DeleteRange()`. There is currently no way to provide enough promises to make it safe in `OptimisticTransactionDB`.

Fixes https://github.com/facebook/rocksdb/issues/7913.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/7929

Test Plan: unit tests covering the cases it's permitted/rejected

Reviewed By: ltamasi

Differential Revision: D26240254

Pulled By: ajkr

fbshipit-source-id: 2834a0ce64cc3e4c3799e35b885a5e79c2f4f6d9

.circleci/config.yml

@@ -2,154 +2,283 @@ 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
+
+  install-cmake-on-macos:
+    steps:
+      - run:
+          name: Install cmake on macos
+          command: |
+            HOMEBREW_NO_AUTO_UPDATE=1 brew install cmake
+
+  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:
+    parameters:
+      python-version:
+        default: "3.5.9"
+        type: string
+    steps:
+      - checkout
+      - run: pyenv install --skip-existing <<parameters.python-version>>
+      - run: pyenv global <<parameters.python-version>>
+      - 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" >> $BASH_ENV
+
+  pre-steps-macos:
+      steps:
+        - pre-steps:
+            python-version: "3.6.0"
+
+  post-steps:
+    steps:
+      - slack/status: *notify-on-master-failure
+      - store_test_results: # store test result if there's any
+          path: /tmp/test-results
+      - store_artifacts: # store LOG for debugging if there's any
+          path: LOG
+
+  install-clang-10:
+    steps:
+      - run:
+          name: Install Clang 10
+          command: |
+            echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
+            echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
+            echo "APT::Acquire::Retries \"10\";" | sudo tee -a /etc/apt/apt.conf.d/80-retries # llvm.org unreliable
+            sudo apt-get update -y && sudo apt-get install -y clang-10
+
+  install-gflags:
+    steps:
+      - run:
+          name: Install gflags
+          command: |
+            sudo apt-get update -y && sudo apt-get install -y libgflags-dev
+  
+  install-gflags-on-macos:
+    steps:
+      - run:
+          name: Install gflags on macos
+          command: |
+            HOMEBREW_NO_AUTO_UPDATE=1 brew install gflags
+
+  install-gtest-parallel:
+    steps:
+      - run:
+          name: Install gtest-parallel
+          command: |
+            git clone --single-branch --branch master --depth 1 https://github.com/google/gtest-parallel.git ~/gtest-parallel
+            echo "export PATH=$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: 9.4.1
+    steps:
+      - increase-max-open-files-on-macos
+      - install-pyenv-on-macos
+      - install-gflags-on-macos
+      - pre-steps-macos
+      - run: ulimit -S -n 1048576 && OPT=-DCIRCLECI make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-macos-cmake:
+    macos:
+      xcode: 9.4.1
+    steps:
+      - increase-max-open-files-on-macos
+      - install-pyenv-on-macos
+      - install-cmake-on-macos
+      - install-gflags-on-macos
+      - pre-steps-macos
+      - run: ulimit -S -n 1048576 && (mkdir build && cd build && cmake -DWITH_GFLAGS=0 .. && make V=1 -j32) | .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 && sudo apt-get install -y libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 PRINT_PARALLEL_OUTPUTS=1 GTEST_THROW_ON_FAILURE=0 GTEST_OUTPUT="xml:/tmp/test-results/" make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
-      - store_test_results:
-          path: /tmp/test-results
+      - pre-steps
+      - install-gflags
+      - run: make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-mem-env:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - run: MEM_ENV=1 make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-encrypted-env:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: 2xlarge
+    steps:
+      - pre-steps
+      - install-gflags
+      - run: ENCRYPTED_ENV=1 make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
+      - post-steps
 
   build-linux-shared_lib-alt_namespace-status_checked:
     machine:
-      image: ubuntu-1604:201903-01
+      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 && sudo apt-get install -y libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 PRINT_PARALLEL_OUTPUTS=1 ASSERT_STATUS_CHECKED=1 LIB_MODE=shared OPT="-DROCKSDB_NAMESPACE=alternative_rocksdb_ns" GTEST_THROW_ON_FAILURE=0 GTEST_OUTPUT="xml:/tmp/test-results/" make V=1 -j32 all check_some | .circleci/cat_ignore_eagain
-      - store_test_results:
-          path: /tmp/test-results
+      - pre-steps
+      - install-gflags
+      - run: ASSERT_STATUS_CHECKED=1 TEST_UINT128_COMPAT=1 ROCKSDB_MODIFY_NPHASH=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 V=1 -j32 release | .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
+      - 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: sudo apt-get update -y && sudo apt-get install -y libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 PRINT_PARALLEL_OUTPUTS=1 LITE=1 GTEST_THROW_ON_FAILURE=0 GTEST_OUTPUT="xml:/tmp/test-results/" make V=1 J=32 -j32 check | .circleci/cat_ignore_eagain
-      - store_test_results:
-          path: /tmp/test-results
-
-  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: LITE=1 make V=1 -j32 release | .circleci/cat_ignore_eagain
+      - 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 && 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
-
-  build-linux-clang10-no-test:
-    machine:
-      image: ubuntu-1604:201903-01
-    resource_class: 2xlarge
-    steps:
-      - checkout # check out the code in the project directory
-      - run: echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: sudo apt-get update -y && sudo apt-get install -y clang-10 libgflags-dev
-      - run: CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 make V=1 -j32 all | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out
+      - post-steps
 
   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: pyenv global 3.5.2
-      - run: echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: sudo apt-get update -y && sudo apt-get install -y clang-10 libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 COMPILE_WITH_ASAN=1 CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 PRINT_PARALLEL_OUTPUTS=1 GTEST_THROW_ON_FAILURE=0 GTEST_OUTPUT="xml:/tmp/test-results/" make V=1 -j32 check | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out
-      - store_test_results:
-          path: /tmp/test-results
+      - 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-1604:201903-01
+      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: echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: sudo apt-get update -y && sudo apt-get install -y clang-10 libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 COMPILE_WITH_TSAN=1 CC=clang-10 CXX=clang++-10 ROCKSDB_DISABLE_ALIGNED_NEW=1 USE_CLANG=1 PRINT_PARALLEL_OUTPUTS=1 EXCLUDE_TESTS_REGEX="TransactionStressTest|SnapshotConcurrentAccess|SeqAdvanceConcurrent|DeadlockStress|MultiThreadedDBTest.MultiThreaded|WriteUnpreparedStressTest.ReadYourOwnWriteStress|DBAsBaseDB/TransactionStressTest|FlushCloseWALFiles|BackgroundPurgeCFDropTest" GTEST_THROW_ON_FAILURE=0 GTEST_OUTPUT="xml:/tmp/test-results/" make V=1 -j32 check | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out. Exclude FlushCloseWALFiles and BackgroundPurgeCFDropTest for occasional failures.
-      - store_test_results:
-          path: /tmp/test-results
+      - 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-1604:201903-01
+      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: echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: sudo apt-get update -y && sudo apt-get install -y clang-10 libgflags-dev
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 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 PRINT_PARALLEL_OUTPUTS=1 make V=1 -j32 ubsan_check | .circleci/cat_ignore_eagain # aligned new doesn't work for reason we haven't figured out
+      - 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-1604:201903-01
+      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: echo "deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: echo "deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-10 main" | sudo tee -a /etc/apt/sources.list
-      - run: sudo apt-get update -y && sudo apt-get install -y clang-10 libgflags-dev clang-tools-10
-      - run: SKIP_FORMAT_BUCK_CHECKS=1 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 PRINT_PARALLEL_OUTPUTS=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.
+      - 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 V=1 -j32) | .circleci/cat_ignore_eagain
+      - post-steps
 
   build-linux-unity:
     docker: # executor type
@@ -159,6 +288,17 @@ jobs:
       - 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
@@ -219,43 +359,156 @@ jobs:
           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-1604:201903-01
-    resource_class: 2xlarge
+      image: ubuntu-1604:202007-01
+    resource_class: large
+    environment:
+      JAVA_HOME: /usr/lib/jvm/java-1.8.0-openjdk-amd64
     steps:
-      - checkout
-      - run: pyenv global 3.5.2
-      - run: sudo apt-get update -y && sudo apt-get install -y libgflags-dev
+      - pre-steps
+      - install-gflags
       - run:
-          name: "Build RocksDBJava"
+          name: "Set Java Environment"
           command: |
-            export JAVA_HOME=/usr/lib/jvm/jdk1.8.0
-            export PATH=$JAVA_HOME/bin:$PATH
             echo "JAVA_HOME=${JAVA_HOME}"
+            echo 'export PATH=$JAVA_HOME/bin:$PATH' >> $BASH_ENV
             which java && java -version
             which javac && javac -version
-            SKIP_FORMAT_BUCK_CHECKS=1 PRINT_PARALLEL_OUTPUTS=1 make V=1 J=32 -j32 rocksdbjava jtest | .circleci/cat_ignore_eagain
+      - run:
+          name: "Build RocksDBJava Shared Library"
+          command: make V=1 J=8 -j8 rocksdbjava | .circleci/cat_ignore_eagain
+      - run:
+          name: "Test RocksDBJava"
+          command: make V=1 J=8 -j8 jtest | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-linux-java-static:
+    machine:
+      image: ubuntu-1604:202007-01
+    resource_class: large
+    environment:
+      JAVA_HOME: /usr/lib/jvm/java-1.8.0-openjdk-amd64
+    steps:
+      - pre-steps
+      - install-gflags
+      - run:
+          name: "Set Java Environment"
+          command: |
+            echo "JAVA_HOME=${JAVA_HOME}"
+            echo 'export PATH=$JAVA_HOME/bin:$PATH' >> $BASH_ENV
+            which java && java -version
+            which javac && javac -version
+      - run:
+          name: "Build RocksDBJava Static Library"
+          command: make V=1 J=8 -j8 rocksdbjavastatic | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-macos-java:
+    macos:
+      xcode: 9.4.1
+    resource_class: medium
+    environment:
+      JAVA_HOME: /Library/Java/JavaVirtualMachines/jdk1.8.0_172.jdk/Contents/Home
+    steps:
+      - increase-max-open-files-on-macos
+      - install-pyenv-on-macos
+      - install-gflags-on-macos
+      - pre-steps-macos
+      - run:
+          name: "Set Java Environment"
+          command: |
+            echo "JAVA_HOME=${JAVA_HOME}"
+            echo 'export PATH=$JAVA_HOME/bin:$PATH' >> $BASH_ENV
+            which java && java -version
+            which javac && javac -version
+      - run:
+          name: "Build RocksDBJava Shared Library"
+          command: make V=1 J=8 -j8 rocksdbjava | .circleci/cat_ignore_eagain
+      - run:
+          name: "Test RocksDBJava"
+          command: make V=1 J=8 -j8 jtest | .circleci/cat_ignore_eagain
+      - post-steps
+
+  build-macos-java-static:
+    macos:
+      xcode: 9.4.1
+    resource_class: medium
+    environment:
+      JAVA_HOME: /Library/Java/JavaVirtualMachines/jdk1.8.0_172.jdk/Contents/Home
+    steps:
+      - increase-max-open-files-on-macos
+      - install-pyenv-on-macos
+      - install-gflags-on-macos
+      - install-cmake-on-macos
+      - pre-steps-macos
+      - run:
+          name: "Set Java Environment"
+          command: |
+            echo "JAVA_HOME=${JAVA_HOME}"
+            echo 'export PATH=$JAVA_HOME/bin:$PATH' >> $BASH_ENV
+            which java && java -version
+            which javac && javac -version
+      - run:
+          name: "Build RocksDBJava Static Library"
+          command: make V=1 J=8 -j8 rocksdbjavastatic | .circleci/cat_ignore_eagain
+      - post-steps
 
   build-examples:
     machine:
-      image: ubuntu-1604:201903-01
+      image: ubuntu-1604:202007-01
     resource_class: medium
     steps:
-      - checkout # check out the code in the project directory
-      - run: pyenv global 3.5.2
-      - run: sudo apt-get update -y && sudo apt-get install -y libgflags-dev
+      - 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-mem-env:
+    jobs:
+      - build-linux-mem-env
+  build-linux-encrypted-env:
+    jobs:
+      - build-linux-encrypted-env
   build-linux-shared_lib-alt_namespace-status_checked:
     jobs:
       - build-linux-shared_lib-alt_namespace-status_checked
@@ -265,15 +518,15 @@ workflows:
   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:
-    jobs:
-      - build-linux-clang10-no-test
   build-linux-clang10-asan:
     jobs:
       - build-linux-clang10-asan
@@ -312,6 +565,29 @@ workflows:
   build-java:
     jobs:
       - build-linux-java
+      - build-linux-java-static
+      - build-macos-java
+      - build-macos-java-static
   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/vs2015_install.ps1

@@ -10,7 +10,7 @@ $process = Start-Process "${PWD}\vs_installer.exe" -ArgumentList $VS_INSTALL_ARG
 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]."
+    echo "VS 2015 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."

.github/workflows/sanity_check.yml

@@ -32,7 +32,7 @@ jobs:
     - name: Download clang-format-diff.py
       uses: wei/wget@v1
       with:
-        args: https://raw.githubusercontent.com/llvm-mirror/clang/master/tools/clang-format/clang-format-diff.py
+        args: https://raw.githubusercontent.com/llvm/llvm-project/master/clang/tools/clang-format/clang-format-diff.py
 
     - name: Check format
       run: VERBOSE_CHECK=1 make check-format

.gitignore

@@ -1,4 +1,5 @@
 make_config.mk
+rocksdb.pc
 
 *.a
 *.arc
@@ -7,6 +8,7 @@ make_config.mk
 *.gcda
 *.gcno
 *.o
+*.o.tmp
 *.so
 *.so.*
 *_test
@@ -52,6 +54,7 @@ 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
@@ -224,33 +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
-      sudo snap install cmake --beta --classic;
+      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
@@ -269,19 +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 LIB_MODE=shared 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

@@ -254,12 +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")
@@ -373,6 +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")
@@ -546,10 +570,13 @@ 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_cache.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
@@ -595,6 +622,8 @@ 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
@@ -607,6 +636,7 @@ 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
@@ -617,6 +647,7 @@ 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
@@ -655,6 +686,8 @@ set(SOURCES
         monitoring/thread_status_util.cc
         monitoring/thread_status_util_debug.cc
         options/cf_options.cc
+        options/configurable.cc
+        options/customizable.cc
         options/db_options.cc
         options/options.cc
         options/options_helper.cc
@@ -705,6 +738,7 @@ set(SOURCES
         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
@@ -713,12 +747,14 @@ 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
@@ -775,6 +811,9 @@ 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_manager.cc
+        utilities/transactions/lock/point/point_lock_tracker.cc
+        utilities/transactions/lock/point/point_lock_manager.cc
         utilities/transactions/optimistic_transaction_db_impl.cc
         utilities/transactions/optimistic_transaction.cc
         utilities/transactions/pessimistic_transaction.cc
@@ -782,7 +821,6 @@ set(SOURCES
         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
@@ -992,10 +1030,11 @@ option(WITH_ALL_TESTS "Build all test, rather than a small subset" ON)
 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
@@ -1006,7 +1045,11 @@ 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_cache_test.cc
         db/blob/blob_file_garbage_test.cc
+        db/blob/blob_file_reader_test.cc
+        db/blob/db_blob_basic_test.cc
         db/blob/db_blob_index_test.cc
         db/column_family_test.cc
         db/compact_files_test.cc
@@ -1066,6 +1109,7 @@ 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
@@ -1076,6 +1120,7 @@ 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
@@ -1083,6 +1128,7 @@ 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
@@ -1096,6 +1142,8 @@ if(WITH_TESTS)
         monitoring/iostats_context_test.cc
         monitoring/statistics_test.cc
         monitoring/stats_history_test.cc
+        options/configurable_test.cc
+        options/customizable_test.cc
         options/options_settable_test.cc
         options/options_test.cc
         table/block_based/block_based_filter_block_test.cc
@@ -1112,7 +1160,9 @@ 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
@@ -1130,6 +1180,7 @@ if(WITH_TESTS)
         util/random_test.cc
         util/rate_limiter_test.cc
         util/repeatable_thread_test.cc
+        util/ribbon_test.cc
         util/slice_test.cc
         util/slice_transform_test.cc
         util/timer_queue_test.cc
@@ -1156,7 +1207,7 @@ if(WITH_TESTS)
         utilities/table_properties_collectors/compact_on_deletion_collector_test.cc
         utilities/transactions/optimistic_transaction_test.cc
         utilities/transactions/transaction_test.cc
-        utilities/transactions/transaction_lock_mgr_test.cc
+        utilities/transactions/lock/point/point_lock_manager_test.cc
         utilities/transactions/write_prepared_transaction_test.cc
         utilities/transactions/write_unprepared_transaction_test.cc
         utilities/ttl/ttl_test.cc
@@ -1200,7 +1251,7 @@ 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 ${GFLAGS_LIB} ${ROCKSDB_LIB})
+      target_link_libraries(${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})
@@ -1224,9 +1275,8 @@ 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()
@@ -1237,7 +1287,7 @@ if(WITH_BENCHMARK_TOOLS)
     tools/db_bench.cc
     tools/db_bench_tool.cc)
   target_link_libraries(db_bench
-    ${ROCKSDB_LIB} ${GFLAGS_LIB})
+    ${ROCKSDB_LIB} ${THIRDPARTY_LIBS})
 
   add_executable(cache_bench
     cache/cache_bench.cc)

HISTORY.md

@@ -1,5 +1,120 @@
 # Rocksdb Change Log
-## Unreleased
+## 6.15.5 (02/05/2021)
+### Bug Fixes
+* Since 6.15.0, `TransactionDB` returns error `Status`es from calls to `DeleteRange()` and calls to `Write()` where the `WriteBatch` contains a range deletion. Previously such operations may have succeeded while not providing the expected transactional guarantees. There are certain cases where range deletion can still be used on such DBs; see the API doc on `TransactionDB::DeleteRange()` for details.
+* `OptimisticTransactionDB` now returns error `Status`es from calls to `DeleteRange()` and calls to `Write()` where the `WriteBatch` contains a range deletion. Previously such operations may have succeeded while not providing the expected transactional guarantees.
+
+## 6.15.4 (01/21/2021)
+### Bug Fixes
+* Fix a race condition between DB startups and shutdowns in managing the periodic background worker threads. One effect of this race condition could be the process being terminated.
+
+## 6.15.3 (01/07/2021)
+### Bug Fixes
+* For Java builds, fix errors due to missing compression library includes.
+
+## 6.15.2 (12/22/2020)
+### Bug Fixes
+* Fix failing RocksJava test compilation and add CI jobs
+* Fix jemalloc compilation issue on macOS
+* Fix build issues - compatibility with older gcc, older jemalloc libraries, docker warning when building i686 binaries
+
+## 6.15.1 (12/01/2020)
+### Bug Fixes
+* Truncated WALs ending in incomplete records can no longer produce gaps in the recovered data when `WALRecoveryMode::kPointInTimeRecovery` is used. Gaps are still possible when WALs are truncated exactly on record boundaries.
+* Fix a bug where compressed blocks read by MultiGet are not inserted into the compressed block cache when use_direct_reads = true.
+
+## 6.15.0 (11/13/2020)
+### 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.
+* Since 6.14, fix a bug that could cause a stalled write to crash with mixed of slowdown and no_slowdown writes (`WriteOptions.no_slowdown=true`).
+* Fixed a bug which causes hang in closing DB when refit level is set in opt build. It was because ContinueBackgroundWork() was called in assert statement which is a no op. It was introduced in 6.14.
+* Fixed a bug which causes Get() to return incorrect result when a key's merge operand is applied twice. This can occur if the thread performing Get() runs concurrently with a background flush thread and another thread writing to the MANIFEST file (PR6069).
+* Reverted a behavior change silently introduced in 6.14.2, in which the effects of the `ignore_unknown_options` flag (used in option parsing/loading functions) changed.
+* Reverted a behavior change silently introduced in 6.14, in which options parsing/loading functions began returning `NotFound` instead of `InvalidArgument` for option names not available in the present version.
+* Fixed MultiGet bugs it doesn't return valid data with user defined timestamp.
+* Fixed a potential bug caused by evaluating `TableBuilder::NeedCompact()` before `TableBuilder::Finish()` in compaction job. For example, the `NeedCompact()` method of `CompactOnDeletionCollector` returned by built-in `CompactOnDeletionCollectorFactory` requires `BlockBasedTable::Finish()` to return the correct result. The bug can cause a compaction-generated file not to be marked for future compaction based on deletion ratio.
+* Fixed a seek issue with prefix extractor and timestamp.
+* Fixed a bug of encoding and parsing BlockBasedTableOptions::read_amp_bytes_per_bit as a 64-bit integer.
+* Fixed the logic of populating native data structure for `read_amp_bytes_per_bit` during OPTIONS file parsing on big-endian architecture. Without this fix, original code introduced in PR7659, when running on big-endian machine, can mistakenly store read_amp_bytes_per_bit (an uint32) in little endian format. Future access to `read_amp_bytes_per_bit` will give wrong values. Little endian architecture is not affected.
+
+
+### 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.
+* Add new API `DB::VerifyFileChecksums` to verify SST file checksum with corresponding entries in the MANIFEST if present. Current implementation requires scanning and recomputing file checksums.
+
+### Behavior Changes
+* The dictionary compression settings specified in `ColumnFamilyOptions::compression_opts` now additionally affect files generated by flush and compaction to non-bottommost level. Previously those settings at most affected files generated by compaction to bottommost level, depending on whether `ColumnFamilyOptions::bottommost_compression_opts` overrode them. Users who relied on dictionary compression settings in `ColumnFamilyOptions::compression_opts` affecting only the bottommost level can keep the behavior by moving their dictionary settings to `ColumnFamilyOptions::bottommost_compression_opts` and setting its `enabled` flag.
+* When the `enabled` flag is set in `ColumnFamilyOptions::bottommost_compression_opts`, those compression options now take effect regardless of the value in `ColumnFamilyOptions::bottommost_compression`. Previously, those compression options only took effect when `ColumnFamilyOptions::bottommost_compression != kDisableCompressionOption`. Now, they additionally take effect when `ColumnFamilyOptions::bottommost_compression == kDisableCompressionOption` (such a setting causes bottommost compression type to fall back to `ColumnFamilyOptions::compression_per_level` if configured, and otherwise fall back to `ColumnFamilyOptions::compression`).
+
+### New Features
+* An EXPERIMENTAL new Bloom alternative that saves about 30% space compared to Bloom filters, with about 3-4x construction time and similar query times is available using NewExperimentalRibbonFilterPolicy.
+
+## 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.
+* SST files have a new table property called db_host_id, which is set to the hostname by default. A new option in DBOptions, db_host_id, allows the property value to be overridden with a user specified string, or disable it completely by making the option string empty.
+* Methods to create customizable extensions -- such as TableFactory -- are exposed directly through the Customizable base class (from which these objects inherit).  This change will allow these Customizable classes to be loaded and configured in a standard way (via CreateFromString).  More information on how to write and use Customizable classes is in the customizable.h header file.
+
+## 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
@@ -10,7 +125,6 @@
 * `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.
@@ -26,11 +140,12 @@
 * 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.
-* `BackupTableNameOption BackupableDBOptions::share_files_with_checksum_naming` is added, where `BackupTableNameOption` is an `enum` type with two enumerators `kChecksumAndFileSize` and `kOptionalChecksumAndDbSessionId`. By default, `BackupableDBOptions::share_files_with_checksum_naming` is set to `kOptionalChecksumAndDbSessionId`. In the default case, backup table filenames generated by this version of RocksDB are of the form either `<file_number>_<crc32c>_<db_session_id>.sst` or `<file_number>_<db_session_id>.sst` as opposed to `<file_number>_<crc32c>_<file_size>.sst`. Specifically, table filenames are of the form `<file_number>_<crc32c>_<db_session_id>.sst` if `DBOptions::file_checksum_gen_factory` is set to `GetFileChecksumGenCrc32cFactory()`. Futhermore, the checksum value `<crc32c>` appeared in the filenames is hexadecimal-encoded, instead of being decimal-encoded `uint32_t` value. If `DBOptions::file_checksum_gen_factory` is `nullptr`, the table filenames are of the form `<file_number>_<db_session_id>.sst`. The new default behavior fixes the backup file name collision problem, which might be possible at large scale, but the option `kChecksumAndFileSize` is added to allow use of old naming in case it is needed. Moreover, for table files generated prior to this version of RocksDB, using `kOptionalChecksumAndDbSessionId` will fall back on `kChecksumAndFileSize`. In these cases, the checksum value `<crc32c>` in the filenames `<file_number>_<crc32c>_<file_size>.sst` is decimal-encoded `uint32_t` value as before. This default behavior change is not an upgrade issue, because previous versions of RocksDB can read, restore, and delete backups using new names, and it's OK for a backup directory to use a mixture of table file naming schemes. Note that `share_files_with_checksum_naming` comes into effect only when both `share_files_with_checksum` and `share_table_files` are true.
+* `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).
@@ -38,7 +153,7 @@
 ### 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 `kOptionalChecksumAndDbSessionId`, except on SST files generated before this version of RocksDB, which fall back on using `kChecksumAndFileSize`.
+* 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

Makefile

@@ -99,14 +99,14 @@ endif
 
 ifneq ($(findstring rocksdbjava, $(MAKECMDGOALS)),)
 	LIB_MODE=shared
-        ifneq ($(findstring rocksdbjavastatic, $(MAKECMDGOALS)),)
+	ifneq ($(findstring rocksdbjavastatic, $(MAKECMDGOALS)),)
 		OBJ_DIR=jls
-	        ifneq ($(DEBUG_LEVEL),2)
-	            DEBUG_LEVEL=0
-                endif
-                ifeq ($(MAKECMDGOALS),rocksdbjavastaticpublish)
-	            DEBUG_LEVEL=0
-                endif
+		ifneq ($(DEBUG_LEVEL),2)
+			DEBUG_LEVEL=0
+		endif
+		ifeq ($(MAKECMDGOALS),rocksdbjavastaticpublish)
+			DEBUG_LEVEL=0
+		endif
 	else
 		OBJ_DIR=jl
 	endif
@@ -131,11 +131,14 @@ endif
 # Figure out optimize level.
 ifneq ($(DEBUG_LEVEL), 2)
 ifeq ($(LITE), 0)
-	OPT += -O2
+	OPTIMIZE_LEVEL ?= -O2
 else
-	OPT += -Os
+	OPTIMIZE_LEVEL ?= -Os
 endif
 endif
+# `OPTIMIZE_LEVEL` is empty when the user does not set it and `DEBUG_LEVEL=2`.
+# In that case, the compiler default (`-O0` for gcc and clang) will be used.
+OPT += $(OPTIMIZE_LEVEL)
 
 # compile with -O2 if debug level is not 2
 ifneq ($(DEBUG_LEVEL), 2)
@@ -262,7 +265,7 @@ dummy := $(shell (export ROCKSDB_ROOT="$(CURDIR)"; \
 include make_config.mk
 
 export JAVAC_ARGS
-CLEAN_FILES += make_config.mk
+CLEAN_FILES += make_config.mk rocksdb.pc
 
 ifeq ($(V), 1)
 $(info $(shell uname -a))
@@ -407,6 +410,14 @@ ifdef TEST_CACHE_LINE_SIZE
   PLATFORM_CCFLAGS += -DTEST_CACHE_LINE_SIZE=$(TEST_CACHE_LINE_SIZE)
   PLATFORM_CXXFLAGS += -DTEST_CACHE_LINE_SIZE=$(TEST_CACHE_LINE_SIZE)
 endif
+ifdef TEST_UINT128_COMPAT
+  PLATFORM_CCFLAGS += -DTEST_UINT128_COMPAT=1
+  PLATFORM_CXXFLAGS += -DTEST_UINT128_COMPAT=1
+endif
+ifdef ROCKSDB_MODIFY_NPHASH
+  PLATFORM_CCFLAGS += -DROCKSDB_MODIFY_NPHASH=1
+  PLATFORM_CXXFLAGS += -DROCKSDB_MODIFY_NPHASH=1
+endif
 
 # This (the first rule) must depend on "all".
 default: all
@@ -414,6 +425,11 @@ default: all
 WARNING_FLAGS = -W -Wextra -Wall -Wsign-compare -Wshadow \
   -Wunused-parameter
 
+ifdef USE_CLANG
+	# Used by some teams in Facebook
+	WARNING_FLAGS += -Wshift-sign-overflow
+endif
+
 ifeq ($(PLATFORM), OS_OPENBSD)
 	WARNING_FLAGS += -Wno-unused-lambda-capture
 endif
@@ -541,13 +557,14 @@ PARALLEL_TEST = \
 	persistent_cache_test \
 	table_test \
 	transaction_test \
-	transaction_lock_mgr_test \
+	point_lock_manager_test \
 	write_prepared_transaction_test \
 	write_unprepared_transaction_test \
 
 ifeq ($(USE_FOLLY_DISTRIBUTED_MUTEX),1)
 	TESTS += folly_synchronization_distributed_mutex_test
 	PARALLEL_TEST += folly_synchronization_distributed_mutex_test
+	TESTS_PASSING_ASC = folly_synchronization_distributed_mutex_test
 endif
 
 # options_settable_test doesn't pass with UBSAN as we use hack in the test
@@ -560,8 +577,13 @@ ifdef ASSERT_STATUS_CHECKED
 	TESTS_PASSING_ASC = \
 		arena_test \
 		autovector_test \
+		cache_test \
+		lru_cache_test \
 		blob_file_addition_test \
+		blob_file_builder_test \
+		blob_file_cache_test \
 		blob_file_garbage_test \
+		blob_file_reader_test \
 		bloom_test \
 		cassandra_format_test \
 		cassandra_row_merge_test \
@@ -570,13 +592,26 @@ ifdef ASSERT_STATUS_CHECKED
 		coding_test \
 		crc32c_test \
 		dbformat_test \
+		db_basic_test \
+		db_blob_basic_test \
+		db_flush_test \
+		db_with_timestamp_basic_test \
+		db_with_timestamp_compaction_test \
+		db_options_test \
+		db_properties_test \
+		db_secondary_test \
+		options_file_test \
 		defer_test \
+		filename_test \
 		dynamic_bloom_test \
 		env_basic_test \
 		env_test \
 		env_logger_test \
 		event_logger_test \
+		error_handler_fs_test \
+		auto_roll_logger_test \
 		file_indexer_test \
+		flush_job_test \
 		hash_table_test \
 		hash_test \
 		heap_test \
@@ -584,48 +619,125 @@ ifdef ASSERT_STATUS_CHECKED
 		inlineskiplist_test \
 		io_posix_test \
 		iostats_context_test \
+		ldb_cmd_test \
 		memkind_kmem_allocator_test \
 		merger_test \
 		mock_env_test \
 		object_registry_test \
+		prefix_test \
+		plain_table_db_test \
+		repair_test \
+		configurable_test \
+		customizable_test \
 		options_settable_test \
 		options_test \
 		random_test \
 		range_del_aggregator_test \
+		sst_file_reader_test \
 		range_tombstone_fragmenter_test \
 		repeatable_thread_test \
+		ribbon_test \
 		skiplist_test \
 		slice_test \
+		sst_dump_test \
 		statistics_test \
+		stats_history_test \
 		thread_local_test \
+		trace_analyzer_test \
 		env_timed_test \
+		filelock_test \
 		timer_queue_test \
 		timer_test \
+		options_util_test \
+		persistent_cache_test \
 		util_merge_operators_test \
+		block_cache_trace_analyzer_test \
+		block_cache_tracer_test \
+		cache_simulator_test \
+		sim_cache_test \
+		version_builder_test \
 		version_edit_test \
 		work_queue_test \
 		write_controller_test \
+		compaction_iterator_test \
+		compaction_job_test \
+		compaction_job_stats_test \
+	        io_tracer_test \
+		merge_helper_test \
+		memtable_list_test \
+		flush_job_test \
+		block_based_filter_block_test \
+		block_fetcher_test \
+		full_filter_block_test \
+		partitioned_filter_block_test \
+		column_family_test \
+		file_reader_writer_test \
+		corruption_test \
+		db_universal_compaction_test \
+		import_column_family_test \
+		memory_test \
+		table_test \
 
 ifeq ($(USE_FOLLY_DISTRIBUTED_MUTEX),1)
 TESTS_PASSING_ASC += folly_synchronization_distributed_mutex_test
 endif
 
 	# Enable building all unit tests, but use check_some to run only tests
-	# known to pass ASC
-	SUBSET := $(TESTS_PASSING_ASC)
+	# known to pass ASC (ASSERT_STATUS_CHECKED)
+	ROCKSDBTESTS_SUBSET ?= $(TESTS_PASSING_ASC)
 	# Alternate: only build unit tests known to pass ASC, and run them
 	# with make check
 	#TESTS := $(filter $(TESTS_PASSING_ASC),$(TESTS))
 	#PARALLEL_TEST := $(filter $(TESTS_PASSING_ASC),$(PARALLEL_TEST))
 else
-	SUBSET := $(TESTS)
+	ROCKSDBTESTS_SUBSET ?= $(TESTS)
 endif
+# Not necessarily well thought out or up-to-date, but matches old list
+TESTS_PLATFORM_DEPENDENT := \
+	db_basic_test \
+	db_blob_basic_test \
+	db_with_timestamp_basic_test \
+	db_encryption_test \
+	db_test2 \
+	external_sst_file_basic_test \
+	auto_roll_logger_test \
+	bloom_test \
+	dynamic_bloom_test \
+	c_test \
+	checkpoint_test \
+	crc32c_test \
+	coding_test \
+	inlineskiplist_test \
+	env_basic_test \
+	env_test \
+	env_logger_test \
+	io_posix_test \
+	hash_test \
+	random_test \
+	ribbon_test \
+	thread_local_test \
+	work_queue_test \
+	rate_limiter_test \
+	perf_context_test \
+	iostats_context_test \
+	db_wal_test \
+
+# Sort ROCKSDBTESTS_SUBSET for filtering, except db_test is special (expensive)
+# so is placed first (out-of-order)
+ROCKSDBTESTS_SUBSET := $(filter db_test, $(ROCKSDBTESTS_SUBSET)) $(sort $(filter-out db_test, $(ROCKSDBTESTS_SUBSET)))
+
 ifdef ROCKSDBTESTS_START
-        SUBSET := $(shell echo $(SUBSET) | sed 's/^.*$(ROCKSDBTESTS_START)/$(ROCKSDBTESTS_START)/')
+        ROCKSDBTESTS_SUBSET := $(shell echo $(ROCKSDBTESTS_SUBSET) | sed 's/^.*$(ROCKSDBTESTS_START)/$(ROCKSDBTESTS_START)/')
 endif
 
 ifdef ROCKSDBTESTS_END
-        SUBSET := $(shell echo $(SUBSET) | sed 's/$(ROCKSDBTESTS_END).*//')
+        ROCKSDBTESTS_SUBSET := $(shell echo $(ROCKSDBTESTS_SUBSET) | sed 's/$(ROCKSDBTESTS_END).*//')
+endif
+
+ifeq ($(ROCKSDBTESTS_PLATFORM_DEPENDENT), only)
+        ROCKSDBTESTS_SUBSET := $(filter $(TESTS_PLATFORM_DEPENDENT), $(ROCKSDBTESTS_SUBSET))
+else ifeq ($(ROCKSDBTESTS_PLATFORM_DEPENDENT), exclude)
+        ROCKSDBTESTS_SUBSET := $(filter-out $(TESTS_PLATFORM_DEPENDENT), $(ROCKSDBTESTS_SUBSET))
 endif
 
 # bench_tool_analyer main is in bench_tool_analyzer_tool, or this would be simpler...
@@ -718,7 +830,7 @@ endif  # PLATFORM_SHARED_EXT
 
 .PHONY: blackbox_crash_test check clean coverage crash_test ldb_tests package \
 	release tags tags0 valgrind_check whitebox_crash_test format static_lib shared_lib all \
-	dbg rocksdbjavastatic rocksdbjava install install-static install-shared uninstall \
+	dbg rocksdbjavastatic rocksdbjava gen-pc install install-static install-shared uninstall \
 	analyze tools tools_lib \
 	blackbox_crash_test_with_atomic_flush whitebox_crash_test_with_atomic_flush  \
 	blackbox_crash_test_with_txn whitebox_crash_test_with_txn \
@@ -727,7 +839,7 @@ endif  # PLATFORM_SHARED_EXT
 
 all: $(LIBRARY) $(BENCHMARKS) tools tools_lib test_libs $(TESTS)
 
-all_but_some_tests: $(LIBRARY) $(BENCHMARKS) tools tools_lib test_libs $(SUBSET)
+all_but_some_tests: $(LIBRARY) $(BENCHMARKS) tools tools_lib test_libs $(ROCKSDBTESTS_SUBSET)
 
 static_lib: $(STATIC_LIBRARY)
 
@@ -945,8 +1057,8 @@ ifndef SKIP_FORMAT_BUCK_CHECKS
 endif
 
 # TODO add ldb_tests
-check_some: $(SUBSET)
-	for t in $(SUBSET); do echo "===== Running $$t (`date`)"; ./$$t || exit 1; done
+check_some: $(ROCKSDBTESTS_SUBSET)
+	for t in $(ROCKSDBTESTS_SUBSET); do echo "===== Running $$t (`date`)"; ./$$t || exit 1; done
 
 .PHONY: ldb_tests
 ldb_tests: ldb
@@ -1050,6 +1162,9 @@ ubsan_crash_test_with_best_efforts_recovery: clean
 valgrind_test:
 	ROCKSDB_VALGRIND_RUN=1 DISABLE_JEMALLOC=1 $(MAKE) valgrind_check
 
+valgrind_test_some:
+	ROCKSDB_VALGRIND_RUN=1 DISABLE_JEMALLOC=1 $(MAKE) valgrind_check_some
+
 valgrind_check: $(TESTS)
 	$(MAKE) DRIVER="$(VALGRIND_VER) $(VALGRIND_OPTS)" gen_parallel_tests
 	$(AM_V_GEN)if test "$(J)" != 1                                  \
@@ -1068,6 +1183,14 @@ valgrind_check: $(TESTS)
 		done; \
 	fi
 
+valgrind_check_some: $(ROCKSDBTESTS_SUBSET)
+	for t in $(ROCKSDBTESTS_SUBSET); do \
+		$(VALGRIND_VER) $(VALGRIND_OPTS) ./$$t; \
+		ret_code=$$?; \
+		if [ $$ret_code -ne 0 ]; then \
+			exit $$ret_code; \
+		fi; \
+	done
 
 ifneq ($(PAR_TEST),)
 parloop:
@@ -1304,6 +1427,9 @@ hash_test: $(OBJ_DIR)/util/hash_test.o $(TEST_LIBRARY) $(LIBRARY)
 random_test: $(OBJ_DIR)/util/random_test.o  $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+ribbon_test: $(OBJ_DIR)/util/ribbon_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 option_change_migration_test: $(OBJ_DIR)/utilities/option_change_migration/option_change_migration_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
@@ -1349,6 +1475,9 @@ slice_transform_test: $(OBJ_DIR)/util/slice_transform_test.o $(TEST_LIBRARY) $(L
 db_basic_test: $(OBJ_DIR)/db/db_basic_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+db_blob_basic_test: $(OBJ_DIR)/db/blob/db_blob_basic_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 db_with_timestamp_basic_test: $(OBJ_DIR)/db/db_with_timestamp_basic_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
@@ -1525,6 +1654,9 @@ compact_on_deletion_collector_test: $(OBJ_DIR)/utilities/table_properties_collec
 wal_manager_test: $(OBJ_DIR)/db/wal_manager_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+wal_edit_test: $(OBJ_DIR)/db/wal_edit_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 dbformat_test: $(OBJ_DIR)/db/dbformat_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
@@ -1666,6 +1798,12 @@ thread_list_test: $(OBJ_DIR)/util/thread_list_test.o $(TEST_LIBRARY) $(LIBRARY)
 compact_files_test: $(OBJ_DIR)/db/compact_files_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+configurable_test: options/configurable_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
+customizable_test: options/customizable_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 options_test: $(OBJ_DIR)/options/options_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
@@ -1717,7 +1855,7 @@ write_callback_test: $(OBJ_DIR)/db/write_callback_test.o $(TEST_LIBRARY) $(LIBRA
 heap_test: $(OBJ_DIR)/util/heap_test.o $(GTEST)
 	$(AM_LINK)
 
-transaction_lock_mgr_test: utilities/transactions/transaction_lock_mgr_test.o $(TEST_LIBRARY) $(LIBRARY)
+point_lock_manager_test: utilities/transactions/lock/point/point_lock_manager_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
 transaction_test: $(OBJ_DIR)/utilities/transactions/transaction_test.o $(TEST_LIBRARY) $(LIBRARY)
@@ -1792,52 +1930,94 @@ defer_test: $(OBJ_DIR)/util/defer_test.o $(TEST_LIBRARY) $(LIBRARY)
 blob_file_addition_test: $(OBJ_DIR)/db/blob/blob_file_addition_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+blob_file_builder_test: $(OBJ_DIR)/db/blob/blob_file_builder_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
+blob_file_cache_test: $(OBJ_DIR)/db/blob/blob_file_cache_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 blob_file_garbage_test: $(OBJ_DIR)/db/blob/blob_file_garbage_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+blob_file_reader_test: $(OBJ_DIR)/db/blob/blob_file_reader_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 timer_test: $(OBJ_DIR)/util/timer_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+periodic_work_scheduler_test: $(OBJ_DIR)/db/periodic_work_scheduler_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 testutil_test: $(OBJ_DIR)/test_util/testutil_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
 io_tracer_test: $(OBJ_DIR)/trace_replay/io_tracer_test.o $(OBJ_DIR)/trace_replay/io_tracer.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)
 
+prefetch_test: $(OBJ_DIR)/file/prefetch_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
+io_tracer_parser_test: $(OBJ_DIR)/tools/io_tracer_parser_test.o $(OBJ_DIR)/tools/io_tracer_parser_tool.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
+io_tracer_parser: $(OBJ_DIR)/tools/io_tracer_parser.o $(TOOLS_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 #-------------------------------------------------
 # make install related stuff
-INSTALL_PATH ?= /usr/local
+PREFIX ?= /usr/local
+LIBDIR ?= $(PREFIX)/lib
+INSTALL_LIBDIR = $(DESTDIR)$(LIBDIR)
 
 uninstall:
-	rm -rf $(INSTALL_PATH)/include/rocksdb \
-	  $(INSTALL_PATH)/lib/$(LIBRARY) \
-	  $(INSTALL_PATH)/lib/$(SHARED4) \
-	  $(INSTALL_PATH)/lib/$(SHARED3) \
-	  $(INSTALL_PATH)/lib/$(SHARED2) \
-	  $(INSTALL_PATH)/lib/$(SHARED1)
+	rm -rf $(DESTDIR)$(PREFIX)/include/rocksdb \
+	  $(INSTALL_LIBDIR)/$(LIBRARY) \
+	  $(INSTALL_LIBDIR)/$(SHARED4) \
+	  $(INSTALL_LIBDIR)/$(SHARED3) \
+	  $(INSTALL_LIBDIR)/$(SHARED2) \
+	  $(INSTALL_LIBDIR)/$(SHARED1) \
+	  $(INSTALL_LIBDIR)/pkgconfig/rocksdb.pc
 
-install-headers:
-	install -d $(INSTALL_PATH)/lib
+install-headers: gen-pc
+	install -d $(INSTALL_LIBDIR)
+	install -d $(INSTALL_LIBDIR)/pkgconfig
 	for header_dir in `$(FIND) "include/rocksdb" -type d`; do \
-		install -d $(INSTALL_PATH)/$$header_dir; \
+		install -d $(DESTDIR)/$(PREFIX)/$$header_dir; \
 	done
 	for header in `$(FIND) "include/rocksdb" -type f -name *.h`; do \
-		install -C -m 644 $$header $(INSTALL_PATH)/$$header; \
+		install -C -m 644 $$header $(DESTDIR)/$(PREFIX)/$$header; \
 	done
+	install -C -m 644 rocksdb.pc $(INSTALL_LIBDIR)/pkgconfig/rocksdb.pc
 
 install-static: install-headers $(LIBRARY)
-	install -C -m 755 $(LIBRARY) $(INSTALL_PATH)/lib
+	install -d $(INSTALL_LIBDIR)
+	install -C -m 755 $(LIBRARY) $(INSTALL_LIBDIR)
 
 install-shared: install-headers $(SHARED4)
-	install -C -m 755 $(SHARED4) $(INSTALL_PATH)/lib && \
-		ln -fs $(SHARED4) $(INSTALL_PATH)/lib/$(SHARED3) && \
-		ln -fs $(SHARED4) $(INSTALL_PATH)/lib/$(SHARED2) && \
-		ln -fs $(SHARED4) $(INSTALL_PATH)/lib/$(SHARED1)
+	install -d $(INSTALL_LIBDIR)
+	install -C -m 755 $(SHARED4) $(INSTALL_LIBDIR)
+	ln -fs $(SHARED4) $(INSTALL_LIBDIR)/$(SHARED3)
+	ln -fs $(SHARED4) $(INSTALL_LIBDIR)/$(SHARED2)
+	ln -fs $(SHARED4) $(INSTALL_LIBDIR)/$(SHARED1)
 
 # install static by default + install shared if it exists
 install: install-static
 	[ -e $(SHARED4) ] && $(MAKE) install-shared || :
 
+# Generate the pkg-config file
+gen-pc:
+	-echo 'prefix=$(PREFIX)' > rocksdb.pc
+	-echo 'exec_prefix=$${prefix}' >> rocksdb.pc
+	-echo 'includedir=$${prefix}/include' >> rocksdb.pc
+	-echo 'libdir=$(LIBDIR)' >> rocksdb.pc
+	-echo '' >> rocksdb.pc
+	-echo 'Name: rocksdb' >> rocksdb.pc
+	-echo 'Description: An embeddable persistent key-value store for fast storage' >> rocksdb.pc
+	-echo Version: $(shell ./build_tools/version.sh full) >> rocksdb.pc
+	-echo 'Libs: -L$${libdir} $(EXEC_LDFLAGS) -lrocksdb' >> rocksdb.pc
+	-echo 'Libs.private: $(PLATFORM_LDFLAGS)' >> rocksdb.pc
+	-echo 'Cflags: -I$${includedir} $(PLATFORM_CXXFLAGS)' >> rocksdb.pc
+
 #-------------------------------------------------
 
 
@@ -1874,10 +2054,11 @@ ifneq (,$(filter ppc% arm64 aarch64 sparc64, $(MACHINE)))
 else
 	ROCKSDBJNILIB = librocksdbjni-linux$(ARCH)$(JNI_LIBC_POSTFIX).so
 endif
-ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-linux$(ARCH)$(JNI_LIBC_POSTFIX).jar
-ROCKSDB_JAR_ALL = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH).jar
-ROCKSDB_JAVADOCS_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-javadoc.jar
-ROCKSDB_SOURCES_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-sources.jar
+ROCKSDB_JAVA_VERSION ?= $(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)
+ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_JAVA_VERSION)-linux$(ARCH)$(JNI_LIBC_POSTFIX).jar
+ROCKSDB_JAR_ALL = rocksdbjni-$(ROCKSDB_JAVA_VERSION).jar
+ROCKSDB_JAVADOCS_JAR = rocksdbjni-$(ROCKSDB_JAVA_VERSION)-javadoc.jar
+ROCKSDB_SOURCES_JAR = rocksdbjni-$(ROCKSDB_JAVA_VERSION)-sources.jar
 SHA256_CMD = sha256sum
 
 ZLIB_VER ?= 1.2.11
@@ -1892,14 +2073,14 @@ SNAPPY_DOWNLOAD_BASE ?= https://github.com/google/snappy/archive
 LZ4_VER ?= 1.9.2
 LZ4_SHA256 ?= 658ba6191fa44c92280d4aa2c271b0f4fbc0e34d249578dd05e50e76d0e5efcc
 LZ4_DOWNLOAD_BASE ?= https://github.com/lz4/lz4/archive
-ZSTD_VER ?= 1.4.4
-ZSTD_SHA256 ?= a364f5162c7d1a455cc915e8e3cf5f4bd8b75d09bc0f53965b0c9ca1383c52c8
+ZSTD_VER ?= 1.4.5
+ZSTD_SHA256 ?= 734d1f565c42f691f8420c8d06783ad818060fc390dee43ae0a89f86d0a4f8c2
 ZSTD_DOWNLOAD_BASE ?= https://github.com/facebook/zstd/archive
 CURL_SSL_OPTS ?= --tlsv1
 
 ifeq ($(PLATFORM), OS_MACOSX)
 	ROCKSDBJNILIB = librocksdbjni-osx.jnilib
-	ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-osx.jar
+	ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_JAVA_VERSION)-osx.jar
 	SHA256_CMD = openssl sha256 -r
 ifneq ("$(wildcard $(JAVA_HOME)/include/darwin)","")
 	JAVA_INCLUDE = -I$(JAVA_HOME)/include -I $(JAVA_HOME)/include/darwin
@@ -1910,7 +2091,7 @@ endif
 ifeq ($(PLATFORM), OS_FREEBSD)
 	JAVA_INCLUDE = -I$(JAVA_HOME)/include -I$(JAVA_HOME)/include/freebsd
 	ROCKSDBJNILIB = librocksdbjni-freebsd$(ARCH).so
-	ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-freebsd$(ARCH).jar
+	ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_JAVA_VERSION)-freebsd$(ARCH).jar
 endif
 ifeq ($(PLATFORM), OS_SOLARIS)
 	ROCKSDBJNILIB = librocksdbjni-solaris$(ARCH).so
@@ -1925,78 +2106,78 @@ ifeq ($(PLATFORM), OS_AIX)
 	SNAPPY_MAKE_TARGET = libsnappy.la
 endif
 ifeq ($(PLATFORM), OS_OPENBSD)
-        JAVA_INCLUDE = -I$(JAVA_HOME)/include -I$(JAVA_HOME)/include/openbsd
+	JAVA_INCLUDE = -I$(JAVA_HOME)/include -I$(JAVA_HOME)/include/openbsd
 	ROCKSDBJNILIB = librocksdbjni-openbsd$(ARCH).so
-        ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-openbsd$(ARCH).jar
+	ROCKSDB_JAR = rocksdbjni-$(ROCKSDB_JAVA_VERSION)-openbsd$(ARCH).jar
 endif
 
-libz.a:
-	-rm -rf zlib-$(ZLIB_VER)
-ifeq (,$(wildcard ./zlib-$(ZLIB_VER).tar.gz))
+zlib-$(ZLIB_VER).tar.gz:
 	curl --fail --output zlib-$(ZLIB_VER).tar.gz --location ${ZLIB_DOWNLOAD_BASE}/zlib-$(ZLIB_VER).tar.gz
-endif
 	ZLIB_SHA256_ACTUAL=`$(SHA256_CMD) zlib-$(ZLIB_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(ZLIB_SHA256)" != "$$ZLIB_SHA256_ACTUAL" ]; then \
 		echo zlib-$(ZLIB_VER).tar.gz checksum mismatch, expected=\"$(ZLIB_SHA256)\" actual=\"$$ZLIB_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
+
+libz.a: zlib-$(ZLIB_VER).tar.gz
+	-rm -rf zlib-$(ZLIB_VER)
 	tar xvzf zlib-$(ZLIB_VER).tar.gz
 	cd zlib-$(ZLIB_VER) && CFLAGS='-fPIC ${EXTRA_CFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' ./configure --static && $(MAKE)
 	cp zlib-$(ZLIB_VER)/libz.a .
 
-libbz2.a:
-	-rm -rf bzip2-$(BZIP2_VER)
-ifeq (,$(wildcard ./bzip2-$(BZIP2_VER).tar.gz))
+bzip2-$(BZIP2_VER).tar.gz:
 	curl --fail --output bzip2-$(BZIP2_VER).tar.gz --location ${CURL_SSL_OPTS} ${BZIP2_DOWNLOAD_BASE}/bzip2-$(BZIP2_VER).tar.gz
-endif
 	BZIP2_SHA256_ACTUAL=`$(SHA256_CMD) bzip2-$(BZIP2_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(BZIP2_SHA256)" != "$$BZIP2_SHA256_ACTUAL" ]; then \
 		echo bzip2-$(BZIP2_VER).tar.gz checksum mismatch, expected=\"$(BZIP2_SHA256)\" actual=\"$$BZIP2_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
+
+libbz2.a: bzip2-$(BZIP2_VER).tar.gz
+	-rm -rf bzip2-$(BZIP2_VER)
 	tar xvzf bzip2-$(BZIP2_VER).tar.gz
 	cd bzip2-$(BZIP2_VER) && $(MAKE) CFLAGS='-fPIC -O2 -g -D_FILE_OFFSET_BITS=64 ${EXTRA_CFLAGS}' AR='ar ${EXTRA_ARFLAGS}'
 	cp bzip2-$(BZIP2_VER)/libbz2.a .
 
-libsnappy.a:
-	-rm -rf snappy-$(SNAPPY_VER)
-ifeq (,$(wildcard ./snappy-$(SNAPPY_VER).tar.gz))
+snappy-$(SNAPPY_VER).tar.gz:
 	curl --fail --output snappy-$(SNAPPY_VER).tar.gz --location ${CURL_SSL_OPTS} ${SNAPPY_DOWNLOAD_BASE}/$(SNAPPY_VER).tar.gz
-endif
 	SNAPPY_SHA256_ACTUAL=`$(SHA256_CMD) snappy-$(SNAPPY_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(SNAPPY_SHA256)" != "$$SNAPPY_SHA256_ACTUAL" ]; then \
 		echo snappy-$(SNAPPY_VER).tar.gz checksum mismatch, expected=\"$(SNAPPY_SHA256)\" actual=\"$$SNAPPY_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
+
+libsnappy.a: snappy-$(SNAPPY_VER).tar.gz
+	-rm -rf snappy-$(SNAPPY_VER)
 	tar xvzf snappy-$(SNAPPY_VER).tar.gz
 	mkdir snappy-$(SNAPPY_VER)/build
 	cd snappy-$(SNAPPY_VER)/build && CFLAGS='${EXTRA_CFLAGS}' CXXFLAGS='${EXTRA_CXXFLAGS}' LDFLAGS='${EXTRA_LDFLAGS}' cmake -DCMAKE_POSITION_INDEPENDENT_CODE=ON .. && $(MAKE) ${SNAPPY_MAKE_TARGET}
 	cp snappy-$(SNAPPY_VER)/build/libsnappy.a .
 
-liblz4.a:
-	-rm -rf lz4-$(LZ4_VER)
-ifeq (,$(wildcard ./lz4-$(LZ4_VER).tar.gz))
+lz4-$(LZ4_VER).tar.gz:
 	curl --fail --output lz4-$(LZ4_VER).tar.gz --location ${CURL_SSL_OPTS} ${LZ4_DOWNLOAD_BASE}/v$(LZ4_VER).tar.gz
-endif
 	LZ4_SHA256_ACTUAL=`$(SHA256_CMD) lz4-$(LZ4_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(LZ4_SHA256)" != "$$LZ4_SHA256_ACTUAL" ]; then \
 		echo lz4-$(LZ4_VER).tar.gz checksum mismatch, expected=\"$(LZ4_SHA256)\" actual=\"$$LZ4_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
+
+liblz4.a: lz4-$(LZ4_VER).tar.gz
+	-rm -rf lz4-$(LZ4_VER)
 	tar xvzf lz4-$(LZ4_VER).tar.gz
 	cd lz4-$(LZ4_VER)/lib && $(MAKE) CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' all
 	cp lz4-$(LZ4_VER)/lib/liblz4.a .
 
-libzstd.a:
-	-rm -rf zstd-$(ZSTD_VER)
-ifeq (,$(wildcard ./zstd-$(ZSTD_VER).tar.gz))
+zstd-$(ZSTD_VER).tar.gz:
 	curl --fail --output zstd-$(ZSTD_VER).tar.gz --location ${CURL_SSL_OPTS} ${ZSTD_DOWNLOAD_BASE}/v$(ZSTD_VER).tar.gz
-endif
 	ZSTD_SHA256_ACTUAL=`$(SHA256_CMD) zstd-$(ZSTD_VER).tar.gz | cut -d ' ' -f 1`; \
 	if [ "$(ZSTD_SHA256)" != "$$ZSTD_SHA256_ACTUAL" ]; then \
 		echo zstd-$(ZSTD_VER).tar.gz checksum mismatch, expected=\"$(ZSTD_SHA256)\" actual=\"$$ZSTD_SHA256_ACTUAL\"; \
 		exit 1; \
 	fi
+
+libzstd.a: zstd-$(ZSTD_VER).tar.gz
+	-rm -rf zstd-$(ZSTD_VER)
 	tar xvzf zstd-$(ZSTD_VER).tar.gz
 	cd zstd-$(ZSTD_VER)/lib && DESTDIR=. PREFIX= $(MAKE) CFLAGS='-fPIC -O2 ${EXTRA_CFLAGS}' install
 	cp zstd-$(ZSTD_VER)/lib/libzstd.a .
@@ -2007,14 +2188,23 @@ JAVA_COMPRESSIONS = libz.a libbz2.a libsnappy.a liblz4.a libzstd.a
 endif
 
 JAVA_STATIC_FLAGS = -DZLIB -DBZIP2 -DSNAPPY -DLZ4 -DZSTD
-JAVA_STATIC_INCLUDES = -I./zlib-$(ZLIB_VER) -I./bzip2-$(BZIP2_VER) -I./snappy-$(SNAPPY_VER) -I./lz4-$(LZ4_VER)/lib -I./zstd-$(ZSTD_VER)/lib/include
-ifneq ($(findstring rocksdbjavastatic, $(MAKECMDGOALS)),)
+JAVA_STATIC_INCLUDES = -I./zlib-$(ZLIB_VER) -I./bzip2-$(BZIP2_VER) -I./snappy-$(SNAPPY_VER) -I./snappy-$(SNAPPY_VER)/build -I./lz4-$(LZ4_VER)/lib -I./zstd-$(ZSTD_VER)/lib -I./zstd-$(ZSTD_VER)/lib/dictBuilder
+
+ifneq ($(findstring rocksdbjavastatic, $(filter-out rocksdbjavastatic_deps, $(MAKECMDGOALS))),)
 CXXFLAGS += $(JAVA_STATIC_FLAGS) $(JAVA_STATIC_INCLUDES)
-CFLAGS +=  $(JAVA_STATIC_FLAGS) $(JAVA_STATIC_INCLUDES)
+CFLAGS += $(JAVA_STATIC_FLAGS) $(JAVA_STATIC_INCLUDES)
 endif
-rocksdbjavastatic: $(LIB_OBJECTS) $(JAVA_COMPRESSIONS)
-	cd java;$(MAKE) javalib;
-	rm -f ./java/target/$(ROCKSDBJNILIB)
+rocksdbjavastatic:
+ifeq ($(JAVA_HOME),)
+	$(error JAVA_HOME is not set)
+endif
+	$(MAKE) rocksdbjavastatic_deps
+	$(MAKE) rocksdbjavastatic_libobjects
+	$(MAKE) rocksdbjavastatic_javalib
+
+rocksdbjavastatic_javalib:
+	cd java;$(MAKE) javalib
+	rm -f java/target/$(ROCKSDBJNILIB)
 	$(CXX) $(CXXFLAGS) -I./java/. $(JAVA_INCLUDE) -shared -fPIC \
 	  -o ./java/target/$(ROCKSDBJNILIB) $(JNI_NATIVE_SOURCES) \
 	  $(LIB_OBJECTS) $(COVERAGEFLAGS) \
@@ -2027,21 +2217,30 @@ rocksdbjavastatic: $(LIB_OBJECTS) $(JAVA_COMPRESSIONS)
 	cd java/target/classes;jar -uf ../$(ROCKSDB_JAR) org/rocksdb/*.class org/rocksdb/util/*.class
 	cd java/target/apidocs;jar -cf ../$(ROCKSDB_JAVADOCS_JAR) *
 	cd java/src/main/java;jar -cf ../../../target/$(ROCKSDB_SOURCES_JAR) org
+	openssl sha1 java/target/$(ROCKSDB_JAR) | sed 's/.*= \([0-9a-f]*\)/\1/' > java/target/$(ROCKSDB_JAR).sha1
+	openssl sha1 java/target/$(ROCKSDB_JAVADOCS_JAR) | sed 's/.*= \([0-9a-f]*\)/\1/' > java/target/$(ROCKSDB_JAVADOCS_JAR).sha1
+	openssl sha1 java/target/$(ROCKSDB_SOURCES_JAR) | sed 's/.*= \([0-9a-f]*\)/\1/' > java/target/$(ROCKSDB_SOURCES_JAR).sha1
+
+rocksdbjavastatic_deps: $(JAVA_COMPRESSIONS)
+
+rocksdbjavastatic_libobjects: $(LIB_OBJECTS)
 
 rocksdbjavastaticrelease: rocksdbjavastatic
 	cd java/crossbuild && (vagrant destroy -f || true) && vagrant up linux32 && vagrant halt linux32 && vagrant up linux64 && vagrant halt linux64 && vagrant up linux64-musl && vagrant halt linux64-musl
 	cd java;jar -cf target/$(ROCKSDB_JAR_ALL) HISTORY*.md
 	cd java/target;jar -uf $(ROCKSDB_JAR_ALL) librocksdbjni-*.so librocksdbjni-*.jnilib
 	cd java/target/classes;jar -uf ../$(ROCKSDB_JAR_ALL) org/rocksdb/*.class org/rocksdb/util/*.class
+	openssl sha1 java/target/$(ROCKSDB_JAR_ALL) | sed 's/.*= \([0-9a-f]*\)/\1/' > java/target/$(ROCKSDB_JAR_ALL).sha1
 
 rocksdbjavastaticreleasedocker: rocksdbjavastatic rocksdbjavastaticdockerx86 rocksdbjavastaticdockerx86_64 rocksdbjavastaticdockerx86musl rocksdbjavastaticdockerx86_64musl
 	cd java;jar -cf target/$(ROCKSDB_JAR_ALL) HISTORY*.md
 	cd java/target;jar -uf $(ROCKSDB_JAR_ALL) librocksdbjni-*.so librocksdbjni-*.jnilib
 	cd java/target/classes;jar -uf ../$(ROCKSDB_JAR_ALL) org/rocksdb/*.class org/rocksdb/util/*.class
+	openssl sha1 java/target/$(ROCKSDB_JAR_ALL) | sed 's/.*= \([0-9a-f]*\)/\1/' > java/target/$(ROCKSDB_JAR_ALL).sha1
 
 rocksdbjavastaticdockerx86:
 	mkdir -p java/target
-	docker run --rm --name rocksdb_linux_x86-be --attach stdin --attach stdout --attach stderr --volume $(HOME)/.m2:/root/.m2:ro --volume `pwd`:/rocksdb-host:ro --volume /rocksdb-local-build --volume `pwd`/java/target:/rocksdb-java-target --env DEBUG_LEVEL=$(DEBUG_LEVEL) evolvedbinary/rocksjava:centos6_x86-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh
+	docker run --rm --name rocksdb_linux_x86-be --platform linux/386 --attach stdin --attach stdout --attach stderr --volume $(HOME)/.m2:/root/.m2:ro --volume `pwd`:/rocksdb-host:ro --volume /rocksdb-local-build --volume `pwd`/java/target:/rocksdb-java-target --env DEBUG_LEVEL=$(DEBUG_LEVEL) evolvedbinary/rocksjava:centos6_x86-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh
 
 rocksdbjavastaticdockerx86_64:
 	mkdir -p java/target
@@ -2057,7 +2256,7 @@ rocksdbjavastaticdockerarm64v8:
 
 rocksdbjavastaticdockerx86musl:
 	mkdir -p java/target
-	docker run --rm --name rocksdb_linux_x86-musl-be --attach stdin --attach stdout --attach stderr --volume $(HOME)/.m2:/root/.m2:ro --volume `pwd`:/rocksdb-host:ro --volume /rocksdb-local-build --volume `pwd`/java/target:/rocksdb-java-target --env DEBUG_LEVEL=$(DEBUG_LEVEL) evolvedbinary/rocksjava:alpine3_x86-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh
+	docker run --rm --name rocksdb_linux_x86-musl-be --platform linux/386 --attach stdin --attach stdout --attach stderr --volume $(HOME)/.m2:/root/.m2:ro --volume `pwd`:/rocksdb-host:ro --volume /rocksdb-local-build --volume `pwd`/java/target:/rocksdb-java-target --env DEBUG_LEVEL=$(DEBUG_LEVEL) evolvedbinary/rocksjava:alpine3_x86-be /rocksdb-host/java/crossbuild/docker-build-linux-centos.sh
 
 rocksdbjavastaticdockerx86_64musl:
 	mkdir -p java/target
@@ -2075,36 +2274,36 @@ rocksdbjavastaticpublish: rocksdbjavastaticrelease rocksdbjavastaticpublishcentr
 
 rocksdbjavastaticpublishdocker: rocksdbjavastaticreleasedocker rocksdbjavastaticpublishcentral
 
-rocksdbjavastaticpublishcentral:
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-javadoc.jar -Dclassifier=javadoc
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-sources.jar -Dclassifier=sources
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-linux64.jar -Dclassifier=linux64
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-linux32.jar -Dclassifier=linux32
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-linux64-musl.jar -Dclassifier=linux64-musl
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-linux32-musl.jar -Dclassifier=linux32-musl
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-osx.jar -Dclassifier=osx
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH)-win64.jar -Dclassifier=win64
-	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH).jar
+rocksdbjavastaticpublishcentral: rocksdbjavageneratepom
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-javadoc.jar -Dclassifier=javadoc
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-sources.jar -Dclassifier=sources
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-linux64.jar -Dclassifier=linux64
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-linux32.jar -Dclassifier=linux32
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-linux64-musl.jar -Dclassifier=linux64-musl
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-linux32-musl.jar -Dclassifier=linux32-musl
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-osx.jar -Dclassifier=osx
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION)-win64.jar -Dclassifier=win64
+	mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/pom.xml -Dfile=java/target/rocksdbjni-$(ROCKSDB_JAVA_VERSION).jar
+
+rocksdbjavageneratepom:
+	cd java;cat pom.xml.template | sed 's/\$${ROCKSDB_JAVA_VERSION}/$(ROCKSDB_JAVA_VERSION)/' > pom.xml
 
 # A version of each $(LIBOBJECTS) compiled with -fPIC
 
 jl/%.o: %.cc
 	$(AM_V_CC)mkdir -p $(@D) && $(CXX) $(CXXFLAGS) -fPIC -c $< -o $@ $(COVERAGEFLAGS)
 
-jl/crc32c_ppc.o: util/crc32c_ppc.c
-	$(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@
-
-jl/crc32c_ppc_asm.o: util/crc32c_ppc_asm.S
-	$(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@
-
-
 rocksdbjava: $(LIB_OBJECTS)
+ifeq ($(JAVA_HOME),)
+	$(error JAVA_HOME is not set)
+endif
 	$(AM_V_GEN)cd java;$(MAKE) javalib;
 	$(AM_V_at)rm -f ./java/target/$(ROCKSDBJNILIB)
 	$(AM_V_at)$(CXX) $(CXXFLAGS) -I./java/. $(JAVA_INCLUDE) -shared -fPIC -o ./java/target/$(ROCKSDBJNILIB) $(JNI_NATIVE_SOURCES) $(LIB_OBJECTS) $(JAVA_LDFLAGS) $(COVERAGEFLAGS)
 	$(AM_V_at)cd java;jar -cf target/$(ROCKSDB_JAR) HISTORY*.md
 	$(AM_V_at)cd java/target;jar -uf $(ROCKSDB_JAR) $(ROCKSDBJNILIB)
 	$(AM_V_at)cd java/target/classes;jar -uf ../$(ROCKSDB_JAR) org/rocksdb/*.class org/rocksdb/util/*.class
+	$(AM_V_at)openssl sha1 java/target/$(ROCKSDB_JAR) | sed 's/.*= \([0-9a-f]*\)/\1/' > java/target/$(ROCKSDB_JAR).sha1
 
 jclean:
 	cd java;$(MAKE) clean;
@@ -2159,7 +2358,7 @@ ifeq ($(HAVE_POWER8),1)
 $(OBJ_DIR)/util/crc32c_ppc.o: util/crc32c_ppc.c
 	$(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@
 
-+$(OBJ_DIR)/util/crc32c_ppc_asm.o: util/crc32c_ppc_asm.S
+$(OBJ_DIR)/util/crc32c_ppc_asm.o: util/crc32c_ppc_asm.S
 	$(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@
 endif
 $(OBJ_DIR)/%.o: %.cc
@@ -2205,7 +2404,7 @@ $(OBJ_DIR)/%.c.d: %.c
 	@$(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) \
 	  -MM -MT'$@' -MT'$(<:.c=.o)' "$<" -o '$@'
 
-+$(OBJ_DIR)/%.S.d: %.S
+$(OBJ_DIR)/%.S.d: %.S
 	@$(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) \
 	  -MM -MT'$@' -MT'$(<:.S=.o)' "$<" -o '$@'
 
@@ -2217,6 +2416,9 @@ else
 depend: $(DEPFILES)
 endif
 
+build_subset_tests: $(ROCKSDBTESTS_SUBSET)
+	$(AM_V_GEN)if [ -n "$${ROCKSDBTESTS_SUBSET_TESTS_TO_FILE}" ]; then echo "$(ROCKSDBTESTS_SUBSET)" > "$${ROCKSDBTESTS_SUBSET_TESTS_TO_FILE}"; else echo "$(ROCKSDBTESTS_SUBSET)"; fi
+
 # if the make goal is either "clean" or "format", we shouldn't
 # try to import the *.d files.
 # TODO(kailiu) The unfamiliarity of Make's conditions leads to the ugly
@@ -2225,6 +2427,8 @@ ifneq ($(MAKECMDGOALS),clean)
 ifneq ($(MAKECMDGOALS),format)
 ifneq ($(MAKECMDGOALS),jclean)
 ifneq ($(MAKECMDGOALS),jtest)
+ifneq ($(MAKECMDGOALS),rocksdbjavastatic)
+ifneq ($(MAKECMDGOALS),rocksdbjavastatic_deps)
 ifneq ($(MAKECMDGOALS),package)
 ifneq ($(MAKECMDGOALS),analyze)
 -include $(DEPFILES)
@@ -2234,3 +2438,5 @@ endif
 endif
 endif
 endif
+endif
+endif

README.md

@@ -3,7 +3,7 @@
 [![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)
+[![PPC64le Build Status](http://140-211-168-68-openstack.osuosl.org:8080/buildStatus/icon?job=rocksdb&style=plastic)](http://140-211-168-68-openstack.osuosl.org:8080/job/rocksdb)
 
 RocksDB is developed and maintained by Facebook Database Engineering Team.
 It is built on earlier work on [LevelDB](https://github.com/google/leveldb) by Sanjay Ghemawat (sanjay@google.com)

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 +134,13 @@ 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_cache.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",
@@ -171,6 +186,8 @@ 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",
@@ -183,6 +200,7 @@ 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",
@@ -236,12 +254,20 @@ cpp_library(
         "monitoring/thread_status_util.cc",
         "monitoring/thread_status_util_debug.cc",
         "options/cf_options.cc",
+        "options/configurable.cc",
+        "options/customizable.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",
@@ -287,12 +313,14 @@ cpp_library(
         "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",
@@ -357,6 +385,9 @@ 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_manager.cc",
+        "utilities/transactions/lock/point/point_lock_manager.cc",
+        "utilities/transactions/lock/point/point_lock_tracker.cc",
         "utilities/transactions/optimistic_transaction.cc",
         "utilities/transactions/optimistic_transaction_db_impl.cc",
         "utilities/transactions/pessimistic_transaction.cc",
@@ -364,7 +395,6 @@ cpp_library(
         "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",
@@ -382,6 +412,297 @@ 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_cache.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/customizable.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_manager.cc",
+        "utilities/transactions/lock/point/point_lock_manager.cc",
+        "utilities/transactions/lock/point/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_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(
@@ -389,6 +710,7 @@ cpp_library(
     srcs = [
         "db/db_test_util.cc",
         "table/mock_table.cc",
+        "test_util/mock_time_env.cc",
         "test_util/testharness.cc",
         "test_util/testutil.cc",
         "tools/block_cache_analyzer/block_cache_trace_analyzer.cc",
@@ -405,6 +727,7 @@ cpp_library(
     external_deps = ROCKSDB_EXTERNAL_DEPS + [
         ("googletest", None, "gtest"),
     ],
+    link_whole = False,
 )
 
 cpp_library(
@@ -423,6 +746,7 @@ cpp_library(
     preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
     deps = [":rocksdb_lib"],
     external_deps = ROCKSDB_EXTERNAL_DEPS,
+    link_whole = False,
 )
 
 cpp_library(
@@ -451,26 +775,24 @@ 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"],
-    )
+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 else None
 
-if not is_opt_mode:
-    custom_unittest(
-        "c_test",
-        command = [
-            native.package_name() + "/buckifier/rocks_test_runner.sh",
-            "$(location :{})".format("c_test_bin"),
-        ],
-        type = "simple",
-    )
+custom_unittest(
+    "c_test",
+    command = [
+        native.package_name() + "/buckifier/rocks_test_runner.sh",
+        "$(location :{})".format("c_test_bin"),
+    ],
+    type = "simple",
+) if not is_opt_mode else None
 
 cpp_library(
     name = "env_basic_test_lib",
@@ -483,6 +805,7 @@ 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]
@@ -529,6 +852,20 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "blob_file_builder_test",
+        "db/blob/blob_file_builder_test.cc",
+        "serial",
+        [],
+        [],
+    ],
+    [
+        "blob_file_cache_test",
+        "db/blob/blob_file_cache_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "blob_file_garbage_test",
         "db/blob/blob_file_garbage_test.cc",
@@ -536,6 +873,13 @@ 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",
@@ -704,6 +1048,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "configurable_test",
+        "options/configurable_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "corruption_test",
         "db/corruption_test.cc",
@@ -739,6 +1090,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "customizable_test",
+        "options/customizable_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "data_block_hash_index_test",
         "table/block_based/data_block_hash_index_test.cc",
@@ -753,6 +1111,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "db_blob_basic_test",
+        "db/blob/db_blob_basic_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "db_blob_index_test",
         "db/blob/db_blob_index_test.cc",
@@ -1173,6 +1538,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "io_tracer_parser_test",
+        "tools/io_tracer_parser_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "io_tracer_test",
         "trace_replay/io_tracer_test.cc",
@@ -1341,6 +1713,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "periodic_work_scheduler_test",
+        "db/periodic_work_scheduler_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "persistent_cache_test",
         "utilities/persistent_cache/persistent_cache_test.cc",
@@ -1355,6 +1734,20 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "point_lock_manager_test",
+        "utilities/transactions/lock/point/point_lock_manager_test.cc",
+        "parallel",
+        [],
+        [],
+    ],
+    [
+        "prefetch_test",
+        "file/prefetch_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "prefix_test",
         "db/prefix_test.cc",
@@ -1418,6 +1811,13 @@ ROCKS_TESTS = [
         [],
         [],
     ],
+    [
+        "ribbon_test",
+        "util/ribbon_test.cc",
+        "serial",
+        [],
+        [],
+    ],
     [
         "sim_cache_test",
         "utilities/simulator_cache/sim_cache_test.cc",
@@ -1537,13 +1937,6 @@ ROCKS_TESTS = [
         [],
         [],
     ],
-    [
-        "transaction_lock_mgr_test",
-        "utilities/transactions/transaction_lock_mgr_test.cc",
-        "parallel",
-        [],
-        [],
-    ],
     [
         "transaction_test",
         "utilities/transactions/transaction_test.cc",
@@ -1660,7 +2053,7 @@ 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 + [

buckifier/buckify_rocksdb.py

@@ -136,6 +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",

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):
@@ -66,38 +67,36 @@ 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("""
-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"],
-    )
+        self.targets_file.write(b"""
+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 else None
 
-if not is_opt_mode:
-    custom_unittest(
-        "c_test",
-        command = [
-            native.package_name() + "/buckifier/rocks_test_runner.sh",
-            "$(location :{})".format("c_test_bin"),
-        ],
-        type = "simple",
-    )
+custom_unittest(
+    "c_test",
+    command = [
+        native.package_name() + "/buckifier/rocks_test_runner.sh",
+        "$(location :{})".format("c_test_bin"),
+    ],
+    type = "simple",
+) if not is_opt_mode else None
 """)
 
     def register_test(self,
@@ -119,5 +118,5 @@ if not is_opt_mode:
         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
 
@@ -657,11 +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 $TRY_SSE42 -x c++ - -o /dev/null 2>/dev/null <<EOF
@@ -710,6 +725,34 @@ elif test "$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() {

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_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/format-diff.sh

@@ -52,15 +52,15 @@ else
     else
       echo "You didn't have clang-format-diff.py and/or clang-format available in your computer!"
       echo "You can download clang-format-diff.py by running: "
-      echo "    curl --location http://goo.gl/iUW1u2 -o ${CLANG_FORMAT_DIFF}"
+      echo "    curl --location https://tinyurl.com/y2kvokof -o ${REPO_ROOT}/clang-format-diff.py"
       echo "You can download clang-format by running:"
       echo "    brew install clang-format"
       echo "  Or"
       echo "    apt install clang-format"
       echo "  This might work too:"
       echo "    yum install git-clang-format"
-      echo "Then, move both files (i.e. ${CLANG_FORMAT_DIFF} and clang-format) to some directory within PATH=${PATH}"
-      echo "and make sure ${CLANG_FORMAT_DIFF} is executable."
+      echo "Then make sure clang-format is available and executable from \$PATH:"
+      echo "    clang-format --version"
       exit 128
     fi
     # Check argparse pre-req on interpreter, or it will fail
@@ -75,17 +75,15 @@ else
       exit 129
     fi
     # Unfortunately, some machines have a Python2 clang-format-diff.py
-    # installed but only a Python3 interpreter installed. Rather than trying
-    # different Python versions that might be installed, we can try migrating
-    # the code to Python3 if it looks like Python2
+    # installed but only a Python3 interpreter installed. Unfortunately,
+    # automatic 2to3 migration is insufficient, so suggest downloading latest.
     if grep -q "print '" "$CFD_PATH" && \
        ${PYTHON:-python3} --version | grep -q 'ython 3'; then
-      if [ ! -f "$REPO_ROOT/.py3/clang-format-diff.py" ]; then
-        echo "Migrating $CFD_PATH to Python3 in a hidden file"
-        mkdir -p "$REPO_ROOT/.py3"
-        ${PYTHON:-python3} -m lib2to3 -w -n -o "$REPO_ROOT/.py3" "$CFD_PATH" > /dev/null || exit 128
-      fi
-      CFD_PATH="$REPO_ROOT/.py3/clang-format-diff.py"
+      echo "You have clang-format-diff.py for Python 2 but are using a Python 3"
+      echo "interpreter (${PYTHON:-python3})."
+      echo "You can download clang-format-diff.py for Python 3 by running: "
+      echo "    curl --location https://tinyurl.com/y2kvokof -o ${REPO_ROOT}/clang-format-diff.py"
+      exit 130
     fi
     CLANG_FORMAT_DIFF="${PYTHON:-python3} $CFD_PATH"
     # This had better work after all those checks

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

@@ -1101,6 +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`
@@ -1113,6 +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`

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_helpers.h

@@ -0,0 +1,114 @@
+//  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 <cassert>
+
+#include "rocksdb/cache.h"
+#include "rocksdb/rocksdb_namespace.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+// Returns the cached value given a cache handle.
+template <typename T>
+T* GetFromCacheHandle(Cache* cache, Cache::Handle* handle) {
+  assert(cache);
+  assert(handle);
+
+  return static_cast<T*>(cache->Value(handle));
+}
+
+// Simple generic deleter for Cache (to be used with Cache::Insert).
+template <typename T>
+void DeleteCacheEntry(const Slice& /* key */, void* value) {
+  delete static_cast<T*>(value);
+}
+
+// Turns a T* into a Slice so it can be used as a key with Cache.
+template <typename T>
+Slice GetSlice(const T* t) {
+  return Slice(reinterpret_cast<const char*>(t), sizeof(T));
+}
+
+// Generic resource management object for cache handles that releases the handle
+// when destroyed. Has unique ownership of the handle, so copying it is not
+// allowed, while moving it transfers ownership.
+template <typename T>
+class CacheHandleGuard {
+ public:
+  CacheHandleGuard() = default;
+
+  CacheHandleGuard(Cache* cache, Cache::Handle* handle)
+      : cache_(cache),
+        handle_(handle),
+        value_(GetFromCacheHandle<T>(cache, handle)) {
+    assert(cache_ && handle_ && value_);
+  }
+
+  CacheHandleGuard(const CacheHandleGuard&) = delete;
+  CacheHandleGuard& operator=(const CacheHandleGuard&) = delete;
+
+  CacheHandleGuard(CacheHandleGuard&& rhs) noexcept
+      : cache_(rhs.cache_), handle_(rhs.handle_), value_(rhs.value_) {
+    assert((!cache_ && !handle_ && !value_) || (cache_ && handle_ && value_));
+
+    rhs.ResetFields();
+  }
+
+  CacheHandleGuard& operator=(CacheHandleGuard&& rhs) noexcept {
+    if (this == &rhs) {
+      return *this;
+    }
+
+    ReleaseHandle();
+
+    cache_ = rhs.cache_;
+    handle_ = rhs.handle_;
+    value_ = rhs.value_;
+
+    assert((!cache_ && !handle_ && !value_) || (cache_ && handle_ && value_));
+
+    rhs.ResetFields();
+
+    return *this;
+  }
+
+  ~CacheHandleGuard() { ReleaseHandle(); }
+
+  bool IsEmpty() const { return !handle_; }
+
+  Cache* GetCache() const { return cache_; }
+  Cache::Handle* GetCacheHandle() const { return handle_; }
+  T* GetValue() const { return value_; }
+
+  void Reset() {
+    ReleaseHandle();
+    ResetFields();
+  }
+
+ private:
+  void ReleaseHandle() {
+    if (IsEmpty()) {
+      return;
+    }
+
+    assert(cache_);
+    cache_->Release(handle_);
+  }
+
+  void ResetFields() {
+    cache_ = nullptr;
+    handle_ = nullptr;
+    value_ = nullptr;
+  }
+
+ private:
+  Cache* cache_ = nullptr;
+  Cache::Handle* handle_ = nullptr;
+  T* value_ = nullptr;
+};
+
+}  // namespace ROCKSDB_NAMESPACE

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.cc

@@ -11,7 +11,6 @@
 
 #include <assert.h>
 #include <stdio.h>
-#include <stdlib.h>
 #include <string>
 
 #include "util/mutexlock.h"

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_cache.cc

@@ -0,0 +1,99 @@
+//  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_cache.h"
+
+#include <cassert>
+#include <memory>
+
+#include "db/blob/blob_file_reader.h"
+#include "options/cf_options.h"
+#include "rocksdb/cache.h"
+#include "rocksdb/slice.h"
+#include "test_util/sync_point.h"
+#include "util/hash.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+BlobFileCache::BlobFileCache(Cache* cache,
+                             const ImmutableCFOptions* immutable_cf_options,
+                             const FileOptions* file_options,
+                             uint32_t column_family_id,
+                             HistogramImpl* blob_file_read_hist)
+    : cache_(cache),
+      mutex_(kNumberOfMutexStripes, GetSliceNPHash64),
+      immutable_cf_options_(immutable_cf_options),
+      file_options_(file_options),
+      column_family_id_(column_family_id),
+      blob_file_read_hist_(blob_file_read_hist) {
+  assert(cache_);
+  assert(immutable_cf_options_);
+  assert(file_options_);
+}
+
+Status BlobFileCache::GetBlobFileReader(
+    uint64_t blob_file_number,
+    CacheHandleGuard<BlobFileReader>* blob_file_reader) {
+  assert(blob_file_reader);
+  assert(blob_file_reader->IsEmpty());
+
+  const Slice key = GetSlice(&blob_file_number);
+
+  assert(cache_);
+
+  Cache::Handle* handle = cache_->Lookup(key);
+  if (handle) {
+    *blob_file_reader = CacheHandleGuard<BlobFileReader>(cache_, handle);
+    return Status::OK();
+  }
+
+  TEST_SYNC_POINT("BlobFileCache::GetBlobFileReader:DoubleCheck");
+
+  // Check again while holding mutex
+  MutexLock lock(mutex_.get(key));
+
+  handle = cache_->Lookup(key);
+  if (handle) {
+    *blob_file_reader = CacheHandleGuard<BlobFileReader>(cache_, handle);
+    return Status::OK();
+  }
+
+  assert(immutable_cf_options_);
+  Statistics* const statistics = immutable_cf_options_->statistics;
+
+  RecordTick(statistics, NO_FILE_OPENS);
+
+  std::unique_ptr<BlobFileReader> reader;
+
+  {
+    assert(file_options_);
+    const Status s = BlobFileReader::Create(
+        *immutable_cf_options_, *file_options_, column_family_id_,
+        blob_file_read_hist_, blob_file_number, &reader);
+    if (!s.ok()) {
+      RecordTick(statistics, NO_FILE_ERRORS);
+      return s;
+    }
+  }
+
+  {
+    constexpr size_t charge = 1;
+
+    const Status s = cache_->Insert(key, reader.get(), charge,
+                                    &DeleteCacheEntry<BlobFileReader>, &handle);
+    if (!s.ok()) {
+      RecordTick(statistics, NO_FILE_ERRORS);
+      return s;
+    }
+  }
+
+  reader.release();
+
+  *blob_file_reader = CacheHandleGuard<BlobFileReader>(cache_, handle);
+
+  return Status::OK();
+}
+
+}  // namespace ROCKSDB_NAMESPACE

db/blob/blob_file_cache.h

@@ -0,0 +1,49 @@
+//  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 "cache/cache_helpers.h"
+#include "rocksdb/rocksdb_namespace.h"
+#include "util/mutexlock.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+class Cache;
+struct ImmutableCFOptions;
+struct FileOptions;
+class HistogramImpl;
+class Status;
+class BlobFileReader;
+class Slice;
+
+class BlobFileCache {
+ public:
+  BlobFileCache(Cache* cache, const ImmutableCFOptions* immutable_cf_options,
+                const FileOptions* file_options, uint32_t column_family_id,
+                HistogramImpl* blob_file_read_hist);
+
+  BlobFileCache(const BlobFileCache&) = delete;
+  BlobFileCache& operator=(const BlobFileCache&) = delete;
+
+  Status GetBlobFileReader(uint64_t blob_file_number,
+                           CacheHandleGuard<BlobFileReader>* blob_file_reader);
+
+ private:
+  Cache* cache_;
+  // Note: mutex_ below is used to guard against multiple threads racing to open
+  // the same file.
+  Striped<port::Mutex, Slice> mutex_;
+  const ImmutableCFOptions* immutable_cf_options_;
+  const FileOptions* file_options_;
+  uint32_t column_family_id_;
+  HistogramImpl* blob_file_read_hist_;
+
+  static constexpr size_t kNumberOfMutexStripes = 1 << 7;
+};
+
+}  // namespace ROCKSDB_NAMESPACE

db/blob/blob_file_cache_test.cc

@@ -0,0 +1,268 @@
+//  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_cache.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/cache.h"
+#include "rocksdb/env.h"
+#include "rocksdb/file_system.h"
+#include "rocksdb/options.h"
+#include "rocksdb/statistics.h"
+#include "test_util/sync_point.h"
+#include "test_util/testharness.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+namespace {
+
+// Creates a test blob file with a single blob in it.
+void WriteBlobFile(uint32_t column_family_id,
+                   const ImmutableCFOptions& immutable_cf_options,
+                   uint64_t blob_file_number) {
+  assert(!immutable_cf_options.cf_paths.empty());
+
+  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);
+
+  constexpr bool has_ttl = false;
+  constexpr ExpirationRange expiration_range;
+
+  BlobLogHeader header(column_family_id, kNoCompression, has_ttl,
+                       expiration_range);
+
+  ASSERT_OK(blob_log_writer.WriteHeader(header));
+
+  constexpr char key[] = "key";
+  constexpr char blob[] = "blob";
+
+  std::string compressed_blob;
+  Slice blob_to_write;
+
+  uint64_t key_offset = 0;
+  uint64_t blob_offset = 0;
+
+  ASSERT_OK(blob_log_writer.AddRecord(key, blob, &key_offset, &blob_offset));
+
+  BlobLogFooter footer;
+  footer.blob_count = 1;
+  footer.expiration_range = expiration_range;
+
+  std::string checksum_method;
+  std::string checksum_value;
+
+  ASSERT_OK(
+      blob_log_writer.AppendFooter(footer, &checksum_method, &checksum_value));
+}
+
+}  // anonymous namespace
+
+class BlobFileCacheTest : public testing::Test {
+ protected:
+  BlobFileCacheTest() : mock_env_(Env::Default()) {}
+
+  MockEnv mock_env_;
+};
+
+TEST_F(BlobFileCacheTest, GetBlobFileReader) {
+  Options options;
+  options.env = &mock_env_;
+  options.statistics = CreateDBStatistics();
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_, "BlobFileCacheTest_GetBlobFileReader"),
+      0);
+  options.enable_blob_files = true;
+
+  constexpr uint32_t column_family_id = 1;
+  ImmutableCFOptions immutable_cf_options(options);
+  constexpr uint64_t blob_file_number = 123;
+
+  WriteBlobFile(column_family_id, immutable_cf_options, blob_file_number);
+
+  constexpr size_t capacity = 10;
+  std::shared_ptr<Cache> backing_cache = NewLRUCache(capacity);
+
+  FileOptions file_options;
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  BlobFileCache blob_file_cache(backing_cache.get(), &immutable_cf_options,
+                                &file_options, column_family_id,
+                                blob_file_read_hist);
+
+  // First try: reader should be opened and put in cache
+  CacheHandleGuard<BlobFileReader> first;
+
+  ASSERT_OK(blob_file_cache.GetBlobFileReader(blob_file_number, &first));
+  ASSERT_NE(first.GetValue(), nullptr);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_OPENS), 1);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_ERRORS), 0);
+
+  // Second try: reader should be served from cache
+  CacheHandleGuard<BlobFileReader> second;
+
+  ASSERT_OK(blob_file_cache.GetBlobFileReader(blob_file_number, &second));
+  ASSERT_NE(second.GetValue(), nullptr);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_OPENS), 1);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_ERRORS), 0);
+
+  ASSERT_EQ(first.GetValue(), second.GetValue());
+}
+
+TEST_F(BlobFileCacheTest, GetBlobFileReader_Race) {
+  Options options;
+  options.env = &mock_env_;
+  options.statistics = CreateDBStatistics();
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileCacheTest_GetBlobFileReader_Race"),
+      0);
+  options.enable_blob_files = true;
+
+  constexpr uint32_t column_family_id = 1;
+  ImmutableCFOptions immutable_cf_options(options);
+  constexpr uint64_t blob_file_number = 123;
+
+  WriteBlobFile(column_family_id, immutable_cf_options, blob_file_number);
+
+  constexpr size_t capacity = 10;
+  std::shared_ptr<Cache> backing_cache = NewLRUCache(capacity);
+
+  FileOptions file_options;
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  BlobFileCache blob_file_cache(backing_cache.get(), &immutable_cf_options,
+                                &file_options, column_family_id,
+                                blob_file_read_hist);
+
+  CacheHandleGuard<BlobFileReader> first;
+  CacheHandleGuard<BlobFileReader> second;
+
+  SyncPoint::GetInstance()->SetCallBack(
+      "BlobFileCache::GetBlobFileReader:DoubleCheck", [&](void* /* arg */) {
+        // Disabling sync points to prevent infinite recursion
+        SyncPoint::GetInstance()->DisableProcessing();
+
+        ASSERT_OK(blob_file_cache.GetBlobFileReader(blob_file_number, &second));
+        ASSERT_NE(second.GetValue(), nullptr);
+        ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_OPENS), 1);
+        ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_ERRORS), 0);
+      });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  ASSERT_OK(blob_file_cache.GetBlobFileReader(blob_file_number, &first));
+  ASSERT_NE(first.GetValue(), nullptr);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_OPENS), 1);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_ERRORS), 0);
+
+  ASSERT_EQ(first.GetValue(), second.GetValue());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+TEST_F(BlobFileCacheTest, GetBlobFileReader_IOError) {
+  Options options;
+  options.env = &mock_env_;
+  options.statistics = CreateDBStatistics();
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileCacheTest_GetBlobFileReader_IOError"),
+      0);
+  options.enable_blob_files = true;
+
+  constexpr size_t capacity = 10;
+  std::shared_ptr<Cache> backing_cache = NewLRUCache(capacity);
+
+  ImmutableCFOptions immutable_cf_options(options);
+  FileOptions file_options;
+  constexpr uint32_t column_family_id = 1;
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  BlobFileCache blob_file_cache(backing_cache.get(), &immutable_cf_options,
+                                &file_options, column_family_id,
+                                blob_file_read_hist);
+
+  // Note: there is no blob file with the below number
+  constexpr uint64_t blob_file_number = 123;
+
+  CacheHandleGuard<BlobFileReader> reader;
+
+  ASSERT_TRUE(
+      blob_file_cache.GetBlobFileReader(blob_file_number, &reader).IsIOError());
+  ASSERT_EQ(reader.GetValue(), nullptr);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_OPENS), 1);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_ERRORS), 1);
+}
+
+TEST_F(BlobFileCacheTest, GetBlobFileReader_CacheFull) {
+  Options options;
+  options.env = &mock_env_;
+  options.statistics = CreateDBStatistics();
+  options.cf_paths.emplace_back(
+      test::PerThreadDBPath(&mock_env_,
+                            "BlobFileCacheTest_GetBlobFileReader_CacheFull"),
+      0);
+  options.enable_blob_files = true;
+
+  constexpr uint32_t column_family_id = 1;
+  ImmutableCFOptions immutable_cf_options(options);
+  constexpr uint64_t blob_file_number = 123;
+
+  WriteBlobFile(column_family_id, immutable_cf_options, blob_file_number);
+
+  constexpr size_t capacity = 0;
+  constexpr int num_shard_bits = -1;  // determined automatically
+  constexpr bool strict_capacity_limit = true;
+  std::shared_ptr<Cache> backing_cache =
+      NewLRUCache(capacity, num_shard_bits, strict_capacity_limit);
+
+  FileOptions file_options;
+  constexpr HistogramImpl* blob_file_read_hist = nullptr;
+
+  BlobFileCache blob_file_cache(backing_cache.get(), &immutable_cf_options,
+                                &file_options, column_family_id,
+                                blob_file_read_hist);
+
+  // Insert into cache should fail since it has zero capacity and
+  // strict_capacity_limit is set
+  CacheHandleGuard<BlobFileReader> reader;
+
+  ASSERT_TRUE(blob_file_cache.GetBlobFileReader(blob_file_number, &reader)
+                  .IsIncomplete());
+  ASSERT_EQ(reader.GetValue(), nullptr);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_OPENS), 1);
+  ASSERT_EQ(options.statistics->getTickerCount(NO_FILE_ERRORS), 1);
+}
+
+}  // 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 +1,771 @@
+//  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/sync_point.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"
 
@@ -144,4 +143,3 @@ Status BlobLogRecord::CheckBlobCRC() const {
 }
 
 }  // 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>
 
@@ -106,6 +104,13 @@ 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) {
+    return key_size + kHeaderSize;
+  }
+
   uint64_t key_size = 0;
   uint64_t value_size = 0;
   uint64_t expiration = 0;
@@ -125,5 +130,19 @@ struct BlobLogRecord {
   Status CheckBlobCRC() const;
 };
 
+// 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

@@ -3,11 +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 <algorithm>
+#include "db/blob/blob_log_sequential_reader.h"
 
 #include "file/random_access_file_reader.h"
 #include "monitoring/statistics.h"
@@ -15,16 +12,21 @@
 
 namespace ROCKSDB_NAMESPACE {
 
-BlobLogReader::BlobLogReader(
+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 BlobLogReader::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 +41,13 @@ Status BlobLogReader::ReadSlice(uint64_t size, Slice* slice, char* buf) {
   return s;
 }
 
-Status BlobLogReader::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 +60,13 @@ Status BlobLogReader::ReadHeader(BlobLogHeader* header) {
   return header->DecodeFrom(buffer_);
 }
 
-Status BlobLogReader::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;
@@ -101,5 +112,21 @@ Status BlobLogReader::ReadRecord(BlobLogRecord* record, ReadLevel level,
   return s;
 }
 
+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

@@ -5,31 +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;
+class RandomAccessFileReader;
+class Env;
+class Statistics;
+class Status;
 
 /**
- * BlobLogReader 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 BlobLogReader {
+
+class BlobLogSequentialReader {
  public:
   enum ReadLevel {
     kReadHeader,
@@ -37,27 +33,28 @@ class BlobLogReader {
     kReadHeaderKeyBlob,
   };
 
-  // Create a reader that will return log records from "*file".
-  // "*file" must remain live while this BlobLogReader is in use.
-  BlobLogReader(std::unique_ptr<RandomAccessFileReader>&& file_reader, Env* env,
-                Statistics* statistics);
-  // No copying allowed
-  BlobLogReader(const BlobLogReader&) = delete;
-  BlobLogReader& operator=(const BlobLogReader&) = delete;
+  // Create a reader that will return log records from "*file_reader".
+  BlobLogSequentialReader(std::unique_ptr<RandomAccessFileReader>&& file_reader,
+                          Env* env, Statistics* statistics);
 
-  ~BlobLogReader() = 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_; }
@@ -72,9 +69,8 @@ class BlobLogReader {
   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 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"
 
@@ -30,6 +29,8 @@ BlobLogWriter::BlobLogWriter(std::unique_ptr<WritableFileWriter>&& dest,
       use_fsync_(use_fs),
       last_elem_type_(kEtNone) {}
 
+BlobLogWriter::~BlobLogWriter() = default;
+
 Status BlobLogWriter::Sync() {
   TEST_SYNC_POINT("BlobLogWriter::Sync");
 
@@ -56,7 +57,9 @@ Status BlobLogWriter::WriteHeader(BlobLogHeader& header) {
   return s;
 }
 
-Status BlobLogWriter::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);
 
@@ -66,10 +69,34 @@ Status BlobLogWriter::AppendFooter(BlobLogFooter& footer) {
   Status s = dest_->Append(Slice(str));
   if (s.ok()) {
     block_offset_ += str.size();
+
     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();
   }
 
@@ -139,4 +166,3 @@ Status BlobLogWriter::EmitPhysicalRecord(const std::string& headerbuf,
 }
 
 }  // 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>
@@ -41,7 +39,7 @@ class BlobLogWriter {
   BlobLogWriter(const BlobLogWriter&) = delete;
   BlobLogWriter& operator=(const BlobLogWriter&) = delete;
 
-  ~BlobLogWriter() = default;
+  ~BlobLogWriter();
 
   static void ConstructBlobHeader(std::string* buf, const Slice& key,
                                   const Slice& val, uint64_t expiration);
@@ -56,7 +54,8 @@ class BlobLogWriter {
                             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);
 
@@ -82,4 +81,3 @@ class BlobLogWriter {
 };
 
 }  // namespace ROCKSDB_NAMESPACE
-#endif  // ROCKSDB_LITE

db/blob/db_blob_basic_test.cc

@@ -0,0 +1,184 @@
+//  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_index.h"
+#include "db/db_test_util.h"
+#include "port/stack_trace.h"
+#include "test_util/sync_point.h"
+#include "utilities/fault_injection_env.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+class DBBlobBasicTest : public DBTestBase {
+ protected:
+  DBBlobBasicTest()
+      : DBTestBase("/db_blob_basic_test", /* env_do_fsync */ false) {}
+};
+
+TEST_F(DBBlobBasicTest, GetBlob) {
+  Options options = GetDefaultOptions();
+  options.enable_blob_files = true;
+  options.min_blob_size = 0;
+
+  Reopen(options);
+
+  constexpr char key[] = "key";
+  constexpr char blob_value[] = "blob_value";
+
+  ASSERT_OK(Put(key, blob_value));
+
+  ASSERT_OK(Flush());
+
+  ASSERT_EQ(Get(key), blob_value);
+
+  // Try again with no I/O allowed. The table and the necessary blocks should
+  // already be in their respective caches; however, the blob itself can only be
+  // read from the blob file, so the read should return Incomplete.
+  ReadOptions read_options;
+  read_options.read_tier = kBlockCacheTier;
+
+  PinnableSlice result;
+  ASSERT_TRUE(db_->Get(read_options, db_->DefaultColumnFamily(), key, &result)
+                  .IsIncomplete());
+}
+
+TEST_F(DBBlobBasicTest, GetBlob_CorruptIndex) {
+  Options options = GetDefaultOptions();
+  options.enable_blob_files = true;
+  options.min_blob_size = 0;
+
+  Reopen(options);
+
+  constexpr char key[] = "key";
+
+  // Fake a corrupt blob index.
+  const std::string blob_index("foobar");
+
+  WriteBatch batch;
+  ASSERT_OK(WriteBatchInternal::PutBlobIndex(&batch, 0, key, blob_index));
+  ASSERT_OK(db_->Write(WriteOptions(), &batch));
+
+  ASSERT_OK(Flush());
+
+  PinnableSlice result;
+  ASSERT_TRUE(db_->Get(ReadOptions(), db_->DefaultColumnFamily(), key, &result)
+                  .IsCorruption());
+}
+
+TEST_F(DBBlobBasicTest, GetBlob_InlinedTTLIndex) {
+  constexpr uint64_t min_blob_size = 10;
+
+  Options options = GetDefaultOptions();
+  options.enable_blob_files = true;
+  options.min_blob_size = min_blob_size;
+
+  Reopen(options);
+
+  constexpr char key[] = "key";
+  constexpr char blob[] = "short";
+  static_assert(sizeof(short) - 1 < min_blob_size,
+                "Blob too long to be inlined");
+
+  // Fake an inlined TTL blob index.
+  std::string blob_index;
+
+  constexpr uint64_t expiration = 1234567890;
+
+  BlobIndex::EncodeInlinedTTL(&blob_index, expiration, blob);
+
+  WriteBatch batch;
+  ASSERT_OK(WriteBatchInternal::PutBlobIndex(&batch, 0, key, blob_index));
+  ASSERT_OK(db_->Write(WriteOptions(), &batch));
+
+  ASSERT_OK(Flush());
+
+  PinnableSlice result;
+  ASSERT_TRUE(db_->Get(ReadOptions(), db_->DefaultColumnFamily(), key, &result)
+                  .IsCorruption());
+}
+
+TEST_F(DBBlobBasicTest, GetBlob_IndexWithInvalidFileNumber) {
+  Options options = GetDefaultOptions();
+  options.enable_blob_files = true;
+  options.min_blob_size = 0;
+
+  Reopen(options);
+
+  constexpr char key[] = "key";
+
+  // Fake a blob index referencing a non-existent blob file.
+  std::string blob_index;
+
+  constexpr uint64_t blob_file_number = 1000;
+  constexpr uint64_t offset = 1234;
+  constexpr uint64_t size = 5678;
+
+  BlobIndex::EncodeBlob(&blob_index, blob_file_number, offset, size,
+                        kNoCompression);
+
+  WriteBatch batch;
+  ASSERT_OK(WriteBatchInternal::PutBlobIndex(&batch, 0, key, blob_index));
+  ASSERT_OK(db_->Write(WriteOptions(), &batch));
+
+  ASSERT_OK(Flush());
+
+  PinnableSlice result;
+  ASSERT_TRUE(db_->Get(ReadOptions(), db_->DefaultColumnFamily(), key, &result)
+                  .IsCorruption());
+}
+
+class DBBlobBasicIOErrorTest : public DBBlobBasicTest,
+                               public testing::WithParamInterface<std::string> {
+ protected:
+  DBBlobBasicIOErrorTest() : sync_point_(GetParam()) {
+    fault_injection_env_.reset(new FaultInjectionTestEnv(env_));
+  }
+  ~DBBlobBasicIOErrorTest() { Close(); }
+
+  std::unique_ptr<FaultInjectionTestEnv> fault_injection_env_;
+  std::string sync_point_;
+};
+
+INSTANTIATE_TEST_CASE_P(DBBlobBasicTest, DBBlobBasicIOErrorTest,
+                        ::testing::ValuesIn(std::vector<std::string>{
+                            "BlobFileReader::OpenFile:NewRandomAccessFile",
+                            "BlobFileReader::GetBlob:ReadFromFile"}));
+
+TEST_P(DBBlobBasicIOErrorTest, GetBlob_IOError) {
+  Options options;
+  options.env = fault_injection_env_.get();
+  options.enable_blob_files = true;
+  options.min_blob_size = 0;
+
+  Reopen(options);
+
+  constexpr char key[] = "key";
+  constexpr char blob_value[] = "blob_value";
+
+  ASSERT_OK(Put(key, blob_value));
+
+  ASSERT_OK(Flush());
+
+  SyncPoint::GetInstance()->SetCallBack(sync_point_, [this](void* /* arg */) {
+    fault_injection_env_->SetFilesystemActive(false,
+                                              Status::IOError(sync_point_));
+  });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  PinnableSlice result;
+  ASSERT_TRUE(db_->Get(ReadOptions(), db_->DefaultColumnFamily(), key, &result)
+                  .IsIOError());
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+}
+
+}  // namespace ROCKSDB_NAMESPACE
+
+int main(int argc, char** argv) {
+  ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

db/blob/db_blob_index_test.cc

@@ -41,7 +41,8 @@ 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(); }
 
@@ -72,6 +73,9 @@ class DBBlobIndexTest : public DBTestBase {
     if (s.IsNotFound()) {
       return "NOT_FOUND";
     }
+    if (s.IsCorruption()) {
+      return "CORRUPTION";
+    }
     if (s.IsNotSupported()) {
       return "NOT_SUPPORTED";
     }
@@ -99,6 +103,7 @@ class DBBlobIndexTest : public DBTestBase {
 
   Options GetTestOptions() {
     Options options;
+    options.env = CurrentOptions().env;
     options.create_if_missing = true;
     options.num_levels = 2;
     options.disable_auto_compactions = true;
@@ -152,8 +157,13 @@ TEST_F(DBBlobIndexTest, Write) {
   }
 }
 
-// Get should be able to return blob index if is_blob_index is provided,
-// otherwise return Status::NotSupported status.
+// Note: the following test case pertains to the StackableDB-based BlobDB
+// implementation. Get should be able to return blob index if is_blob_index is
+// provided, otherwise it should return Status::NotSupported (when reading from
+// memtable) or Status::Corruption (when reading from SST). Reading from SST
+// returns Corruption because we can't differentiate between the application
+// accidentally opening the base DB of a stacked BlobDB and actual corruption
+// when using the integrated BlobDB.
 TEST_F(DBBlobIndexTest, Get) {
   for (auto tier : kAllTiers) {
     DestroyAndReopen(GetTestOptions());
@@ -170,15 +180,22 @@ TEST_F(DBBlobIndexTest, Get) {
     ASSERT_EQ("value", GetImpl("key", &is_blob_index));
     ASSERT_FALSE(is_blob_index);
     // Verify blob index
-    ASSERT_TRUE(Get("blob_key", &value).IsNotSupported());
-    ASSERT_EQ("NOT_SUPPORTED", GetImpl("blob_key"));
+    if (tier <= kImmutableMemtables) {
+      ASSERT_TRUE(Get("blob_key", &value).IsNotSupported());
+      ASSERT_EQ("NOT_SUPPORTED", GetImpl("blob_key"));
+    } else {
+      ASSERT_TRUE(Get("blob_key", &value).IsCorruption());
+      ASSERT_EQ("CORRUPTION", GetImpl("blob_key"));
+    }
     ASSERT_EQ("blob_index", GetImpl("blob_key", &is_blob_index));
     ASSERT_TRUE(is_blob_index);
   }
 }
 
-// Get should NOT return Status::NotSupported if blob index is updated with
-// a normal value.
+// Note: the following test case pertains to the StackableDB-based BlobDB
+// implementation. Get should NOT return Status::NotSupported/Status::Corruption
+// if blob index is updated with a normal value. See the test case above for
+// more details.
 TEST_F(DBBlobIndexTest, Updated) {
   for (auto tier : kAllTiers) {
     DestroyAndReopen(GetTestOptions());
@@ -205,7 +222,11 @@ TEST_F(DBBlobIndexTest, Updated) {
       ASSERT_EQ("blob_index", GetBlobIndex("key" + ToString(i), snapshot));
     }
     ASSERT_EQ("new_value", Get("key1"));
-    ASSERT_EQ("NOT_SUPPORTED", GetImpl("key2"));
+    if (tier <= kImmutableMemtables) {
+      ASSERT_EQ("NOT_SUPPORTED", GetImpl("key2"));
+    } else {
+      ASSERT_EQ("CORRUPTION", GetImpl("key2"));
+    }
     ASSERT_EQ("NOT_FOUND", Get("key3"));
     ASSERT_EQ("NOT_FOUND", Get("key4"));
     ASSERT_EQ("a,b,c", GetImpl("key5"));

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 ImmutableCFOptions& ioptions,
+    const std::string& dbname, VersionSet* versions,
+    const ImmutableDBOptions& db_options, 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,
@@ -83,19 +86,23 @@ 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,
-    const std::string& db_session_id) {
+    const std::string& db_session_id, const std::string* full_history_ts_low) {
   assert((column_family_id ==
           TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
          column_family_name.empty());
   // Reports the IOStats for flush for every following bytes.
   const size_t kReportFlushIOStatsEvery = 1048576;
-  uint64_t paranoid_hash = 0;
+  OutputValidator output_validator(
+      internal_comparator,
+      /*enable_order_check=*/
+      mutable_cf_options.check_flush_compaction_key_order,
+      /*enable_hash=*/paranoid_file_checks);
   Status s;
-  IOStatus io_s;
   meta->fd.file_size = 0;
   iter->SeekToFirst();
   std::unique_ptr<CompactionRangeDelAggregator> range_del_agg(
@@ -106,26 +113,29 @@ 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
+  Env* env = db_options.env;
+  assert(env);
+  FileSystem* fs = db_options.fs.get();
+  assert(fs);
   TableProperties tp;
   if (iter->Valid() || !range_del_agg->IsEmpty()) {
     TableBuilder* builder;
     std::unique_ptr<WritableFileWriter> file_writer;
-    // Currently we only enable dictionary compression during compaction to the
-    // bottommost level.
-    CompressionOptions compression_opts_for_flush(compression_opts);
-    compression_opts_for_flush.max_dict_bytes = 0;
-    compression_opts_for_flush.zstd_max_train_bytes = 0;
     {
       std::unique_ptr<FSWritableFile> file;
 #ifndef NDEBUG
       bool use_direct_writes = file_options.use_direct_writes;
       TEST_SYNC_POINT_CALLBACK("BuildTable:create_file", &use_direct_writes);
 #endif  // !NDEBUG
-      io_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;
@@ -133,21 +143,23 @@ Status BuildTable(
       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,
           int_tbl_prop_collector_factories, column_family_id,
           column_family_name, file_writer.get(), compression,
-          sample_for_compression, compression_opts_for_flush, level,
+          sample_for_compression, compression_opts, level,
           false /* skip_filters */, creation_time, oldest_key_time,
           0 /*target_file_size*/, file_creation_time, db_id, db_session_id);
     }
@@ -158,20 +170,35 @@ 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,
+        /*compaction=*/nullptr,
+        /*compaction_filter=*/nullptr, /*shutting_down=*/nullptr,
+        /*preserve_deletes_seqnum=*/0, /*manual_compaction_paused=*/nullptr,
+        db_options.info_log, full_history_ts_low);
+
     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();
-      if (paranoid_file_checks) {
-        // Generate a rolling 64-bit hash of the key and values
-        paranoid_hash = Hash64(key.data(), key.size(), paranoid_hash);
-        paranoid_hash = Hash64(value.data(), value.size(), paranoid_hash);
+      // 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);
@@ -183,29 +210,33 @@ Status BuildTable(
             ThreadStatus::FLUSH_BYTES_WRITTEN, IOSTATS(bytes_written));
       }
     }
-
-    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 (!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);
+      }
     }
 
-    // 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_s = builder->io_status();
     if (io_status->ok()) {
-      *io_status = io_s;
+      *io_status = builder->io_status();
     }
 
     if (s.ok() && !empty) {
@@ -221,10 +252,12 @@ 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);
     }
+    TEST_SYNC_POINT("BuildTable:BeforeCloseTableFile");
     if (s.ok() && io_status->ok() && !empty) {
       *io_status = file_writer->Close();
     }
@@ -232,12 +265,22 @@ Status BuildTable(
       // 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 (s.ok()) {
       s = *io_status;
     }
 
+    if (blob_file_builder) {
+      if (s.ok()) {
+        s = blob_file_builder->Finish();
+      }
+
+      blob_file_builder.reset();
+    }
+
     // TODO Also check the IO status when create the Iterator.
 
     if (s.ok() && !empty) {
@@ -247,8 +290,9 @@ 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
@@ -261,16 +305,16 @@ Status BuildTable(
           /*allow_unprepared_value*/ false));
       s = it->status();
       if (s.ok() && paranoid_file_checks) {
-        uint64_t check_hash = 0;
+        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
-          check_hash = Hash64(it->key().data(), it->key().size(), check_hash);
-          check_hash =
-              Hash64(it->value().data(), it->value().size(), check_hash);
+          file_validator.Add(it->key(), it->value()).PermitUncheckedError();
         }
         s = it->status();
-        if (s.ok() && check_hash != paranoid_hash) {
-          s = Status::Corruption("Paraniod checksums do not match");
+        if (s.ok() && !output_validator.CompareValidator(file_validator)) {
+          s = Status::Corruption("Paranoid checksums do not match");
         }
       }
     }
@@ -282,7 +326,21 @@ Status BuildTable(
   }
 
   if (!s.ok() || meta->fd.GetFileSize() == 0) {
-    fs->DeleteFile(fname, IOOptions(), nullptr);
+    TEST_SYNC_POINT("BuildTable:BeforeDeleteFile");
+
+    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();
+      }
+    }
   }
 
   if (meta->fd.GetFileSize() == 0) {
@@ -291,7 +349,8 @@ 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 ImmutableCFOptions& options,
+    const std::string& dbname, VersionSet* versions,
+    const ImmutableDBOptions& db_options, 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,
@@ -79,12 +82,14 @@ 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,
     Env::WriteLifeTimeHint write_hint = Env::WLTH_NOT_SET,
     const uint64_t file_creation_time = 0, const std::string& db_id = "",
-    const std::string& db_session_id = "");
+    const std::string& db_session_id = "",
+    const std::string* full_history_ts_low = nullptr);
 
 }  // namespace ROCKSDB_NAMESPACE

db/c.cc

@@ -60,7 +60,6 @@ using ROCKSDB_NAMESPACE::ColumnFamilyDescriptor;
 using ROCKSDB_NAMESPACE::ColumnFamilyHandle;
 using ROCKSDB_NAMESPACE::ColumnFamilyOptions;
 using ROCKSDB_NAMESPACE::CompactionFilter;
-using ROCKSDB_NAMESPACE::CompactionFilterContext;
 using ROCKSDB_NAMESPACE::CompactionFilterFactory;
 using ROCKSDB_NAMESPACE::CompactionOptionsFIFO;
 using ROCKSDB_NAMESPACE::CompactRangeOptions;
@@ -82,7 +81,6 @@ using ROCKSDB_NAMESPACE::LiveFileMetaData;
 using ROCKSDB_NAMESPACE::Logger;
 using ROCKSDB_NAMESPACE::MemoryUtil;
 using ROCKSDB_NAMESPACE::MergeOperator;
-using ROCKSDB_NAMESPACE::MergeOperators;
 using ROCKSDB_NAMESPACE::NewBloomFilterPolicy;
 using ROCKSDB_NAMESPACE::NewGenericRateLimiter;
 using ROCKSDB_NAMESPACE::NewLRUCache;
@@ -115,10 +113,8 @@ using ROCKSDB_NAMESPACE::WriteBatch;
 using ROCKSDB_NAMESPACE::WriteBatchWithIndex;
 using ROCKSDB_NAMESPACE::WriteOptions;
 
-using std::shared_ptr;
 using std::vector;
 using std::unordered_set;
-using std::map;
 
 extern "C" {
 
@@ -181,6 +177,9 @@ struct rocksdb_transaction_options_t {
 struct rocksdb_transaction_t {
   Transaction* rep;
 };
+struct rocksdb_backupable_db_options_t {
+  BackupableDBOptions rep;
+};
 struct rocksdb_checkpoint_t {
   Checkpoint* rep;
 };
@@ -504,13 +503,13 @@ 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;
@@ -549,6 +548,18 @@ rocksdb_backup_engine_t* rocksdb_backup_engine_open(
   return result;
 }
 
+rocksdb_backup_engine_t* rocksdb_backup_engine_open_opts(
+    const rocksdb_backupable_db_options_t* options, rocksdb_env_t* env,
+    char** errptr) {
+  BackupEngine* be;
+  if (SaveError(errptr, BackupEngine::Open(options->rep, env->rep, &be))) {
+    return nullptr;
+  }
+  rocksdb_backup_engine_t* result = new rocksdb_backup_engine_t;
+  result->rep = be;
+  return result;
+}
+
 void rocksdb_backup_engine_create_new_backup(rocksdb_backup_engine_t* be,
                                              rocksdb_t* db,
                                              char** errptr) {
@@ -645,6 +656,128 @@ void rocksdb_backup_engine_close(rocksdb_backup_engine_t* be) {
   delete be;
 }
 
+rocksdb_backupable_db_options_t* rocksdb_backupable_db_options_create(
+    const char* backup_dir) {
+  return new rocksdb_backupable_db_options_t{
+      BackupableDBOptions(std::string(backup_dir))};
+}
+
+void rocksdb_backupable_db_options_set_backup_dir(
+    rocksdb_backupable_db_options_t* options, const char* backup_dir) {
+  options->rep.backup_dir = std::string(backup_dir);
+}
+
+void rocksdb_backupable_db_options_set_env(
+    rocksdb_backupable_db_options_t* options, rocksdb_env_t* env) {
+  options->rep.backup_env = (env ? env->rep : nullptr);
+}
+
+void rocksdb_backupable_db_options_set_share_table_files(
+    rocksdb_backupable_db_options_t* options, unsigned char val) {
+  options->rep.share_table_files = val;
+}
+
+unsigned char rocksdb_backupable_db_options_get_share_table_files(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.share_table_files;
+}
+
+void rocksdb_backupable_db_options_set_sync(
+    rocksdb_backupable_db_options_t* options, unsigned char val) {
+  options->rep.sync = val;
+}
+
+unsigned char rocksdb_backupable_db_options_get_sync(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.sync;
+}
+
+void rocksdb_backupable_db_options_set_destroy_old_data(
+    rocksdb_backupable_db_options_t* options, unsigned char val) {
+  options->rep.destroy_old_data = val;
+}
+
+unsigned char rocksdb_backupable_db_options_get_destroy_old_data(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.destroy_old_data;
+}
+
+void rocksdb_backupable_db_options_set_backup_log_files(
+    rocksdb_backupable_db_options_t* options, unsigned char val) {
+  options->rep.backup_log_files = val;
+}
+
+unsigned char rocksdb_backupable_db_options_get_backup_log_files(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.backup_log_files;
+}
+
+void rocksdb_backupable_db_options_set_backup_rate_limit(
+    rocksdb_backupable_db_options_t* options, uint64_t limit) {
+  options->rep.backup_rate_limit = limit;
+}
+
+uint64_t rocksdb_backupable_db_options_get_backup_rate_limit(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.backup_rate_limit;
+}
+
+void rocksdb_backupable_db_options_set_restore_rate_limit(
+    rocksdb_backupable_db_options_t* options, uint64_t limit) {
+  options->rep.restore_rate_limit = limit;
+}
+
+uint64_t rocksdb_backupable_db_options_get_restore_rate_limit(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.restore_rate_limit;
+}
+
+void rocksdb_backupable_db_options_set_max_background_operations(
+    rocksdb_backupable_db_options_t* options, int val) {
+  options->rep.max_background_operations = val;
+}
+
+int rocksdb_backupable_db_options_get_max_background_operations(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.max_background_operations;
+}
+
+void rocksdb_backupable_db_options_set_callback_trigger_interval_size(
+    rocksdb_backupable_db_options_t* options, uint64_t size) {
+  options->rep.callback_trigger_interval_size = size;
+}
+
+uint64_t rocksdb_backupable_db_options_get_callback_trigger_interval_size(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.callback_trigger_interval_size;
+}
+
+void rocksdb_backupable_db_options_set_max_valid_backups_to_open(
+    rocksdb_backupable_db_options_t* options, int val) {
+  options->rep.max_valid_backups_to_open = val;
+}
+
+int rocksdb_backupable_db_options_get_max_valid_backups_to_open(
+    rocksdb_backupable_db_options_t* options) {
+  return options->rep.max_valid_backups_to_open;
+}
+
+void rocksdb_backupable_db_options_set_share_files_with_checksum_naming(
+    rocksdb_backupable_db_options_t* options, int val) {
+  options->rep.share_files_with_checksum_naming =
+      static_cast<BackupableDBOptions::ShareFilesNaming>(val);
+}
+
+int rocksdb_backupable_db_options_get_share_files_with_checksum_naming(
+    rocksdb_backupable_db_options_t* options) {
+  return static_cast<int>(options->rep.share_files_with_checksum_naming);
+}
+
+void rocksdb_backupable_db_options_destroy(
+    rocksdb_backupable_db_options_t* options) {
+  delete options;
+}
+
 rocksdb_checkpoint_t* rocksdb_checkpoint_object_create(rocksdb_t* db,
                                                        char** errptr) {
   Checkpoint* checkpoint;
@@ -707,12 +840,47 @@ 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(
@@ -722,8 +890,10 @@ 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;
   }
 
@@ -805,6 +975,18 @@ 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,
@@ -3612,11 +3794,20 @@ 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) {
@@ -3653,11 +3844,19 @@ 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;
@@ -3668,37 +3867,70 @@ 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;
 }
@@ -3712,33 +3944,61 @@ 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;
 }
@@ -3752,21 +4012,40 @@ 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;
 }
@@ -3780,6 +4059,10 @@ 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);
@@ -3794,6 +4077,10 @@ 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();
 }
@@ -3831,10 +4118,36 @@ void rocksdb_env_set_background_threads(rocksdb_env_t* env, int n) {
   env->rep->SetBackgroundThreads(n);
 }
 
+int rocksdb_env_get_background_threads(rocksdb_env_t* env) {
+  return env->rep->GetBackgroundThreads();
+}
+
+void rocksdb_env_set_bottom_priority_background_threads(rocksdb_env_t* env,
+                                                        int n) {
+  env->rep->SetBackgroundThreads(n, Env::BOTTOM);
+}
+
+int rocksdb_env_get_bottom_priority_background_threads(rocksdb_env_t* env) {
+  return env->rep->GetBackgroundThreads(Env::BOTTOM);
+}
+
 void rocksdb_env_set_high_priority_background_threads(rocksdb_env_t* env, int n) {
   env->rep->SetBackgroundThreads(n, Env::HIGH);
 }
 
+int rocksdb_env_get_high_priority_background_threads(rocksdb_env_t* env) {
+  return env->rep->GetBackgroundThreads(Env::HIGH);
+}
+
+void rocksdb_env_set_low_priority_background_threads(rocksdb_env_t* env,
+                                                     int n) {
+  env->rep->SetBackgroundThreads(n, Env::LOW);
+}
+
+int rocksdb_env_get_low_priority_background_threads(rocksdb_env_t* env) {
+  return env->rep->GetBackgroundThreads(Env::LOW);
+}
+
 void rocksdb_env_join_all_threads(rocksdb_env_t* env) {
   env->rep->WaitForJoin();
 }
@@ -4062,32 +4375,62 @@ void rocksdb_universal_compaction_options_set_size_ratio(
   uco->rep->size_ratio = ratio;
 }
 
+int rocksdb_universal_compaction_options_get_size_ratio(
+    rocksdb_universal_compaction_options_t* uco) {
+  return uco->rep->size_ratio;
+}
+
 void rocksdb_universal_compaction_options_set_min_merge_width(
   rocksdb_universal_compaction_options_t* uco, int w) {
   uco->rep->min_merge_width = w;
 }
 
+int rocksdb_universal_compaction_options_get_min_merge_width(
+    rocksdb_universal_compaction_options_t* uco) {
+  return uco->rep->min_merge_width;
+}
+
 void rocksdb_universal_compaction_options_set_max_merge_width(
   rocksdb_universal_compaction_options_t* uco, int w) {
   uco->rep->max_merge_width = w;
 }
 
+int rocksdb_universal_compaction_options_get_max_merge_width(
+    rocksdb_universal_compaction_options_t* uco) {
+  return uco->rep->max_merge_width;
+}
+
 void rocksdb_universal_compaction_options_set_max_size_amplification_percent(
   rocksdb_universal_compaction_options_t* uco, int p) {
   uco->rep->max_size_amplification_percent = p;
 }
 
+int rocksdb_universal_compaction_options_get_max_size_amplification_percent(
+    rocksdb_universal_compaction_options_t* uco) {
+  return uco->rep->max_size_amplification_percent;
+}
+
 void rocksdb_universal_compaction_options_set_compression_size_percent(
   rocksdb_universal_compaction_options_t* uco, int p) {
   uco->rep->compression_size_percent = p;
 }
 
+int rocksdb_universal_compaction_options_get_compression_size_percent(
+    rocksdb_universal_compaction_options_t* uco) {
+  return uco->rep->compression_size_percent;
+}
+
 void rocksdb_universal_compaction_options_set_stop_style(
   rocksdb_universal_compaction_options_t* uco, int style) {
   uco->rep->stop_style =
       static_cast<ROCKSDB_NAMESPACE::CompactionStopStyle>(style);
 }
 
+int rocksdb_universal_compaction_options_get_stop_style(
+    rocksdb_universal_compaction_options_t* uco) {
+  return static_cast<int>(uco->rep->stop_style);
+}
+
 void rocksdb_universal_compaction_options_destroy(
   rocksdb_universal_compaction_options_t* uco) {
   delete uco->rep;
@@ -4105,6 +4448,11 @@ void rocksdb_fifo_compaction_options_set_max_table_files_size(
   fifo_opts->rep.max_table_files_size = size;
 }
 
+uint64_t rocksdb_fifo_compaction_options_get_max_table_files_size(
+    rocksdb_fifo_compaction_options_t* fifo_opts) {
+  return fifo_opts->rep.max_table_files_size;
+}
+
 void rocksdb_fifo_compaction_options_destroy(
     rocksdb_fifo_compaction_options_t* fifo_opts) {
   delete fifo_opts;

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
@@ -2235,6 +2239,240 @@ int main(int argc, char** argv) {
     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("env");
+  {
+    rocksdb_env_t* e;
+    e = rocksdb_create_default_env();
+
+    rocksdb_env_set_background_threads(e, 10);
+    CheckCondition(10 == rocksdb_env_get_background_threads(e));
+
+    rocksdb_env_set_high_priority_background_threads(e, 20);
+    CheckCondition(20 == rocksdb_env_get_high_priority_background_threads(e));
+
+    rocksdb_env_set_low_priority_background_threads(e, 30);
+    CheckCondition(30 == rocksdb_env_get_low_priority_background_threads(e));
+
+    rocksdb_env_set_bottom_priority_background_threads(e, 40);
+    CheckCondition(40 == rocksdb_env_get_bottom_priority_background_threads(e));
+
+    rocksdb_env_destroy(e);
+  }
+
+  StartPhase("universal_compaction_options");
+  {
+    rocksdb_universal_compaction_options_t* uco;
+    uco = rocksdb_universal_compaction_options_create();
+
+    rocksdb_universal_compaction_options_set_size_ratio(uco, 5);
+    CheckCondition(5 ==
+                   rocksdb_universal_compaction_options_get_size_ratio(uco));
+
+    rocksdb_universal_compaction_options_set_min_merge_width(uco, 15);
+    CheckCondition(
+        15 == rocksdb_universal_compaction_options_get_min_merge_width(uco));
+
+    rocksdb_universal_compaction_options_set_max_merge_width(uco, 25);
+    CheckCondition(
+        25 == rocksdb_universal_compaction_options_get_max_merge_width(uco));
+
+    rocksdb_universal_compaction_options_set_max_size_amplification_percent(uco,
+                                                                            35);
+    CheckCondition(
+        35 ==
+        rocksdb_universal_compaction_options_get_max_size_amplification_percent(
+            uco));
+
+    rocksdb_universal_compaction_options_set_compression_size_percent(uco, 45);
+    CheckCondition(
+        45 ==
+        rocksdb_universal_compaction_options_get_compression_size_percent(uco));
+
+    rocksdb_universal_compaction_options_set_stop_style(uco, 1);
+    CheckCondition(1 ==
+                   rocksdb_universal_compaction_options_get_stop_style(uco));
+
+    rocksdb_universal_compaction_options_destroy(uco);
+  }
+
+  StartPhase("fifo_compaction_options");
+  {
+    rocksdb_fifo_compaction_options_t* fco;
+    fco = rocksdb_fifo_compaction_options_create();
+
+    rocksdb_fifo_compaction_options_set_max_table_files_size(fco, 100000);
+    CheckCondition(
+        100000 ==
+        rocksdb_fifo_compaction_options_get_max_table_files_size(fco));
+
+    rocksdb_fifo_compaction_options_destroy(fco);
+  }
+
+  StartPhase("backupable_db_option");
+  {
+    rocksdb_backupable_db_options_t* bdo;
+    bdo = rocksdb_backupable_db_options_create("path");
+
+    rocksdb_backupable_db_options_set_share_table_files(bdo, 1);
+    CheckCondition(1 ==
+                   rocksdb_backupable_db_options_get_share_table_files(bdo));
+
+    rocksdb_backupable_db_options_set_sync(bdo, 1);
+    CheckCondition(1 == rocksdb_backupable_db_options_get_sync(bdo));
+
+    rocksdb_backupable_db_options_set_destroy_old_data(bdo, 1);
+    CheckCondition(1 ==
+                   rocksdb_backupable_db_options_get_destroy_old_data(bdo));
+
+    rocksdb_backupable_db_options_set_backup_log_files(bdo, 1);
+    CheckCondition(1 ==
+                   rocksdb_backupable_db_options_get_backup_log_files(bdo));
+
+    rocksdb_backupable_db_options_set_backup_rate_limit(bdo, 123);
+    CheckCondition(123 ==
+                   rocksdb_backupable_db_options_get_backup_rate_limit(bdo));
+
+    rocksdb_backupable_db_options_set_restore_rate_limit(bdo, 37);
+    CheckCondition(37 ==
+                   rocksdb_backupable_db_options_get_restore_rate_limit(bdo));
+
+    rocksdb_backupable_db_options_set_max_background_operations(bdo, 20);
+    CheckCondition(
+        20 == rocksdb_backupable_db_options_get_max_background_operations(bdo));
+
+    rocksdb_backupable_db_options_set_callback_trigger_interval_size(bdo, 9000);
+    CheckCondition(
+        9000 ==
+        rocksdb_backupable_db_options_get_callback_trigger_interval_size(bdo));
+
+    rocksdb_backupable_db_options_set_max_valid_backups_to_open(bdo, 40);
+    CheckCondition(
+        40 == rocksdb_backupable_db_options_get_max_valid_backups_to_open(bdo));
+
+    rocksdb_backupable_db_options_set_share_files_with_checksum_naming(bdo, 2);
+    CheckCondition(
+        2 == rocksdb_backupable_db_options_get_share_files_with_checksum_naming(
+                 bdo));
+
+    rocksdb_backupable_db_options_destroy(bdo);
+  }
+
   StartPhase("iterate_upper_bound");
   {
     // Create new empty database

db/column_family.cc

@@ -12,9 +12,11 @@
 #include <algorithm>
 #include <cinttypes>
 #include <limits>
+#include <sstream>
 #include <string>
 #include <vector>
 
+#include "db/blob/blob_file_cache.h"
 #include "db/compaction/compaction_picker.h"
 #include "db/compaction/compaction_picker_fifo.h"
 #include "db/compaction/compaction_picker_level.h"
@@ -31,7 +33,7 @@
 #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"
@@ -148,6 +150,16 @@ 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();
 }
 
@@ -346,8 +358,8 @@ 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::kName);
+  bool is_block_based_table = (result.table_factory->IsInstanceOf(
+      TableFactory::kBlockBasedTableName()));
 
   const uint64_t kAdjustedTtl = 30 * 24 * 60 * 60;
   if (result.ttl == kDefaultTtl) {
@@ -488,7 +500,8 @@ 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),
@@ -546,7 +559,11 @@ 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));
+    blob_file_cache_.reset(
+        new BlobFileCache(_table_cache, ioptions(), soptions(), id_,
+                          internal_stats_->GetBlobFileReadHist()));
+
     if (ioptions_.compaction_style == kCompactionStyleLevel) {
       compaction_picker_.reset(
           new LevelCompactionPicker(ioptions_, &internal_comparator_));
@@ -1043,7 +1060,8 @@ 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(
@@ -1070,7 +1088,7 @@ bool ColumnFamilyData::RangeOverlapWithCompaction(
 
 Status ColumnFamilyData::RangesOverlapWithMemtables(
     const autovector<Range>& ranges, SuperVersion* super_version,
-    bool* overlap) {
+    bool allow_data_in_errors, bool* overlap) {
   assert(overlap != nullptr);
   *overlap = false;
   // Create an InternalIterator over all unflushed memtables
@@ -1089,10 +1107,12 @@ 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();
@@ -1101,12 +1121,12 @@ Status ColumnFamilyData::RangesOverlapWithMemtables(
     memtable_iter->Seek(range_start.Encode());
     status = memtable_iter->status();
     ParsedInternalKey seek_result;
-    if (status.ok()) {
-      if (memtable_iter->Valid() &&
-          !ParseInternalKey(memtable_iter->key(), &seek_result)) {
-        status = Status::Corruption("DB have corrupted keys");
-      }
+
+    if (status.ok() && memtable_iter->Valid()) {
+      status = ParseInternalKey(memtable_iter->key(), &seek_result,
+                                allow_data_in_errors);
     }
+
     if (status.ok()) {
       if (memtable_iter->Valid() &&
           ucmp->Compare(seek_result.user_key, ranges[i].limit) <= 0) {
@@ -1303,7 +1323,8 @@ Status ColumnFamilyData::ValidateOptions(
   }
 
   if (cf_options.ttl > 0 && cf_options.ttl != kDefaultTtl) {
-    if (cf_options.table_factory->Name() != BlockBasedTableFactory::kName) {
+    if (!cf_options.table_factory->IsInstanceOf(
+            TableFactory::kBlockBasedTableName())) {
       return Status::NotSupported(
           "TTL is only supported in Block-Based Table format. ");
     }
@@ -1311,7 +1332,8 @@ Status ColumnFamilyData::ValidateOptions(
 
   if (cf_options.periodic_compaction_seconds > 0 &&
       cf_options.periodic_compaction_seconds != kDefaultPeriodicCompSecs) {
-    if (cf_options.table_factory->Name() != BlockBasedTableFactory::kName) {
+    if (!cf_options.table_factory->IsInstanceOf(
+            TableFactory::kBlockBasedTableName())) {
       return Status::NotSupported(
           "Periodic Compaction is only supported in "
           "Block-Based Table format. ");
@@ -1403,12 +1425,13 @@ 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),
@@ -1416,7 +1439,8 @@ 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_;
@@ -1482,7 +1506,7 @@ 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);

db/column_family.h

@@ -44,6 +44,7 @@ class LogBuffer;
 class InstrumentedMutex;
 class InstrumentedMutexLock;
 struct SuperVersionContext;
+class BlobFileCache;
 
 extern const double kIncSlowdownRatio;
 // This file contains a list of data structures for managing column family
@@ -381,6 +382,7 @@ class ColumnFamilyData {
                          SequenceNumber earliest_seq);
 
   TableCache* table_cache() const { return table_cache_.get(); }
+  BlobFileCache* blob_file_cache() const { return blob_file_cache_.get(); }
 
   // See documentation in compaction_picker.h
   // REQUIRES: DB mutex held
@@ -404,7 +406,8 @@ class ColumnFamilyData {
   //
   // Thread-safe
   Status RangesOverlapWithMemtables(const autovector<Range>& ranges,
-                                    SuperVersion* super_version, bool* overlap);
+                                    SuperVersion* super_version,
+                                    bool allow_data_in_errors, bool* overlap);
 
   // A flag to tell a manual compaction is to compact all levels together
   // instead of a specific level.
@@ -518,7 +521,8 @@ 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;
 
@@ -542,6 +546,7 @@ class ColumnFamilyData {
   const bool is_delete_range_supported_;
 
   std::unique_ptr<TableCache> table_cache_;
+  std::unique_ptr<BlobFileCache> blob_file_cache_;
 
   std::unique_ptr<InternalStats> internal_stats_;
 
@@ -651,7 +656,8 @@ 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;
@@ -715,6 +721,7 @@ 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

@@ -35,10 +35,10 @@ namespace ROCKSDB_NAMESPACE {
 static const int kValueSize = 1000;
 
 // counts how many operations were performed
-class EnvCounter : public EnvWrapper {
+class EnvCounter : public SpecialEnv {
  public:
   explicit EnvCounter(Env* base)
-      : EnvWrapper(base), num_new_writable_file_(0) {}
+      : SpecialEnv(base), num_new_writable_file_(0) {}
   int GetNumberOfNewWritableFileCalls() {
     return num_new_writable_file_;
   }
@@ -68,6 +68,7 @@ class ColumnFamilyTestBase : public testing::Test {
 #endif  // !ROCKSDB_LITE
     EXPECT_NE(nullptr, base_env);
     env_ = new EnvCounter(base_env);
+    env_->skip_fsync_ = true;
     dbname_ = test::PerThreadDBPath("column_family_test");
     db_options_.create_if_missing = true;
     db_options_.fail_if_options_file_error = true;
@@ -79,10 +80,14 @@ 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();
     ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
     Destroy(column_families);
     delete env_;
@@ -168,7 +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();
@@ -279,9 +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++;
     }
@@ -307,7 +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] = "";
     }
@@ -328,7 +335,7 @@ class ColumnFamilyTestBase : public testing::Test {
         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
@@ -650,7 +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);
@@ -713,7 +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;
@@ -842,13 +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);
@@ -887,10 +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));
     }
   }
 
@@ -918,7 +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);
@@ -983,6 +992,7 @@ TEST_P(ColumnFamilyTest, FlushTest) {
     ASSERT_OK(Put(0, "foofoo", "bar"));
 
     for (auto* it : iterators) {
+      ASSERT_OK(it->status());
       delete it;
     }
   }
@@ -1080,8 +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);
@@ -2067,6 +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;
@@ -2321,7 +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.
@@ -2377,7 +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]));
@@ -2427,7 +2437,7 @@ TEST_P(ColumnFamilyTest, FlushAndDropRaceCondition) {
   // 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;
   PutRandomData(1, kKeysNum, 100);
@@ -3028,6 +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);
 
@@ -3080,6 +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);
 
@@ -3174,6 +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());
 
@@ -3204,8 +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",
@@ -3215,11 +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");
@@ -3257,7 +3271,7 @@ TEST_P(ColumnFamilyTest, DISABLED_LogTruncationTest) {
     FileType type;
     if (!(ParseFileName(filenames[i], &number, &type))) continue;
 
-    if (type != kLogFile) continue;
+    if (type != kWalFile) continue;
 
     logfs.push_back(filenames[i]);
   }
@@ -3302,7 +3316,7 @@ TEST_P(ColumnFamilyTest, DISABLED_LogTruncationTest) {
   Close();
 
   // cleanup
-  env_->DeleteDir(backup_logs);
+  ASSERT_OK(env_->DeleteDir(backup_logs));
 }
 
 TEST_P(ColumnFamilyTest, DefaultCfPathsTest) {
@@ -3364,9 +3378,11 @@ TEST_P(ColumnFamilyTest, MultipleCFPathsTest) {
     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/compacted_db_impl.cc

@@ -41,8 +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();
   }
@@ -74,10 +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;
@@ -149,7 +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

@@ -248,12 +248,6 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
   if (max_subcompactions_ == 0) {
     max_subcompactions_ = _mutable_db_options.max_subcompactions;
   }
-  if (!bottommost_level_) {
-    // Currently we only enable dictionary compression during compaction to the
-    // bottommost level.
-    output_compression_opts_.max_dict_bytes = 0;
-    output_compression_opts_.zstd_max_train_bytes = 0;
-  }
 
 #ifndef NDEBUG
   for (size_t i = 1; i < inputs_.size(); ++i) {
@@ -383,7 +377,13 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
         auto* f = files[level_ptrs->at(lvl)];
         if (user_cmp->Compare(user_key, f->largest.user_key()) <= 0) {
           // We've advanced far enough
-          if (user_cmp->Compare(user_key, f->smallest.user_key()) >= 0) {
+          // In the presence of user-defined timestamp, we may need to handle
+          // the case in which f->smallest.user_key() (including ts) has the
+          // same user key, but the ts part is smaller. If so,
+          // Compare(user_key, f->smallest.user_key()) returns -1.
+          // That's why we need CompareWithoutTimestamp().
+          if (user_cmp->CompareWithoutTimestamp(user_key,
+                                                f->smallest.user_key()) >= 0) {
             // Key falls in this file's range, so it may
             // exist beyond output level
             return false;

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"
@@ -36,20 +38,23 @@ 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::shared_ptr<Logger> info_log)
+    const std::atomic<int>* manual_compaction_paused,
+    const std::shared_ptr<Logger> info_log,
+    const std::string* full_history_ts_low)
     : 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 ? new RealCompaction(compaction) : nullptr),
           compaction_filter, shutting_down, preserve_deletes_seqnum,
-          manual_compaction_paused, info_log) {}
+          manual_compaction_paused, info_log, full_history_ts_low) {}
 
 CompactionIterator::CompactionIterator(
     InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
@@ -58,12 +63,14 @@ 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::shared_ptr<Logger> info_log)
+    const std::atomic<int>* manual_compaction_paused,
+    const std::shared_ptr<Logger> info_log,
+    const std::string* full_history_ts_low)
     : input_(input),
       cmp_(cmp),
       merge_helper_(merge_helper),
@@ -74,20 +81,27 @@ 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),
       manual_compaction_paused_(manual_compaction_paused),
       preserve_deletes_seqnum_(preserve_deletes_seqnum),
+      info_log_(info_log),
+      allow_data_in_errors_(allow_data_in_errors),
+      timestamp_size_(cmp_ ? cmp_->timestamp_size() : 0),
+      full_history_ts_low_(full_history_ts_low),
       current_user_key_sequence_(0),
       current_user_key_snapshot_(0),
       merge_out_iter_(merge_helper_),
       current_key_committed_(false),
-      info_log_(info_log) {
+      cmp_with_history_ts_low_(0) {
   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);
   }
@@ -108,6 +122,8 @@ CompactionIterator::CompactionIterator(
   for (size_t i = 1; i < snapshots_->size(); ++i) {
     assert(snapshots_->at(i - 1) < snapshots_->at(i));
   }
+  assert(timestamp_size_ == 0 || !full_history_ts_low_ ||
+         timestamp_size_ == full_history_ts_low_->size());
 #endif
   input_->SetPinnedItersMgr(&pinned_iters_mgr_);
   TEST_SYNC_POINT_CALLBACK("CompactionIterator:AfterInit", compaction_.get());
@@ -142,14 +158,13 @@ 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_, allow_data_in_errors_);
       // 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) {
-        ROCKS_LOG_FATAL(info_log_, "Invalid key (%s) in compaction",
-                        key_.ToString(true).c_str());
+      assert(s.ok());
+      if (!s.ok()) {
+        ROCKS_LOG_FATAL(info_log_, "Invalid key in compaction. %s",
+                        s.getState());
       }
 
       // Keep current_key_ in sync.
@@ -262,15 +277,15 @@ void CompactionIterator::NextFromInput() {
     value_ = input_->value();
     iter_stats_.num_input_records++;
 
-    if (!ParseInternalKey(key_, &ikey_)) {
+    Status pik_status = ParseInternalKey(key_, &ikey_, allow_data_in_errors_);
+    if (!pik_status.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;
+        status_ = pik_status;
+        return;
       }
       key_ = current_key_.SetInternalKey(key_);
       has_current_user_key_ = false;
@@ -282,7 +297,8 @@ void CompactionIterator::NextFromInput() {
     TEST_SYNC_POINT_CALLBACK("CompactionIterator:ProcessKV", &ikey_);
 
     // Update input statistics
-    if (ikey_.type == kTypeDeletion || ikey_.type == kTypeSingleDeletion) {
+    if (ikey_.type == kTypeDeletion || ikey_.type == kTypeSingleDeletion ||
+        ikey_.type == kTypeDeletionWithTimestamp) {
       iter_stats_.num_input_deletion_records++;
     }
     iter_stats_.total_input_raw_key_bytes += key_.size();
@@ -295,19 +311,56 @@ void CompactionIterator::NextFromInput() {
     // merge_helper_->compaction_filter_skip_until_.
     Slice skip_until;
 
+    int cmp_user_key_without_ts = 0;
+    int cmp_ts = 0;
+    if (has_current_user_key_) {
+      cmp_user_key_without_ts =
+          timestamp_size_
+              ? cmp_->CompareWithoutTimestamp(ikey_.user_key, current_user_key_)
+              : cmp_->Compare(ikey_.user_key, current_user_key_);
+      // if timestamp_size_ > 0, then curr_ts_ has been initialized by a
+      // previous key.
+      cmp_ts = timestamp_size_ ? cmp_->CompareTimestamp(
+                                     ExtractTimestampFromUserKey(
+                                         ikey_.user_key, timestamp_size_),
+                                     curr_ts_)
+                               : 0;
+    }
+
     // Check whether the user key changed. After this if statement current_key_
     // is a copy of the current input key (maybe converted to a delete by the
     // compaction filter). ikey_.user_key is pointing to the copy.
-    if (!has_current_user_key_ ||
-        !cmp_->Equal(ikey_.user_key, current_user_key_)) {
+    if (!has_current_user_key_ || cmp_user_key_without_ts != 0 || cmp_ts != 0) {
       // First occurrence of this user key
       // Copy key for output
       key_ = current_key_.SetInternalKey(key_, &ikey_);
+
+      // If timestamp_size_ > 0, then copy from ikey_ to curr_ts_ for the use
+      // in next iteration to compare with the timestamp of next key.
+      UpdateTimestampAndCompareWithFullHistoryLow();
+
+      // If
+      // (1) !has_current_user_key_, OR
+      // (2) timestamp is disabled, OR
+      // (3) all history will be preserved, OR
+      // (4) user key (excluding timestamp) is different from previous key, OR
+      // (5) timestamp is NO older than *full_history_ts_low_
+      // then current_user_key_ must be treated as a different user key.
+      // This means, if a user key (excluding ts) is the same as the previous
+      // user key, and its ts is older than *full_history_ts_low_, then we
+      // consider this key for GC, e.g. it may be dropped if certain conditions
+      // match.
+      if (!has_current_user_key_ || !timestamp_size_ || !full_history_ts_low_ ||
+          0 != cmp_user_key_without_ts || cmp_with_history_ts_low_ >= 0) {
+        // Initialize for future comparison for rule (A) and etc.
+        current_user_key_sequence_ = kMaxSequenceNumber;
+        current_user_key_snapshot_ = 0;
+        has_current_user_key_ = true;
+      }
       current_user_key_ = ikey_.user_key;
-      has_current_user_key_ = true;
+
       has_outputted_key_ = false;
-      current_user_key_sequence_ = kMaxSequenceNumber;
-      current_user_key_snapshot_ = 0;
+
       current_key_committed_ = KeyCommitted(ikey_.sequence);
 
       // Apply the compaction filter to the first committed version of the user
@@ -350,8 +403,7 @@ 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
@@ -423,7 +475,9 @@ 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, allow_data_in_errors_)
+              .ok() &&
           cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
         // Check whether the next key belongs to the same snapshot as the
         // SingleDelete.
@@ -527,9 +581,12 @@ void CompactionIterator::NextFromInput() {
                         last_sequence, current_user_key_sequence_);
       }
 
-      ++iter_stats_.num_record_drop_hidden;  // (A)
+      ++iter_stats_.num_record_drop_hidden;  // rule (A)
       input_->Next();
-    } else if (compaction_ != nullptr && ikey_.type == kTypeDeletion &&
+    } else if (compaction_ != nullptr &&
+               (ikey_.type == kTypeDeletion ||
+                (ikey_.type == kTypeDeletionWithTimestamp &&
+                 cmp_with_history_ts_low_ < 0)) &&
                IN_EARLIEST_SNAPSHOT(ikey_.sequence) &&
                ikeyNotNeededForIncrementalSnapshot() &&
                compaction_->KeyNotExistsBeyondOutputLevel(ikey_.user_key,
@@ -553,30 +610,49 @@ void CompactionIterator::NextFromInput() {
       // given that:
       // (1) The deletion is earlier than earliest_write_conflict_snapshot, and
       // (2) No value exist earlier than the deletion.
+      //
+      // Note also that a deletion marker of type kTypeDeletionWithTimestamp
+      // will be treated as a different user key unless the timestamp is older
+      // than *full_history_ts_low_.
       ++iter_stats_.num_record_drop_obsolete;
       if (!bottommost_level_) {
         ++iter_stats_.num_optimized_del_drop_obsolete;
       }
       input_->Next();
-    } else if ((ikey_.type == kTypeDeletion) && bottommost_level_ &&
-               ikeyNotNeededForIncrementalSnapshot()) {
+    } else if ((ikey_.type == kTypeDeletion ||
+                (ikey_.type == kTypeDeletionWithTimestamp &&
+                 cmp_with_history_ts_low_ < 0)) &&
+               bottommost_level_ && ikeyNotNeededForIncrementalSnapshot()) {
       // 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) &&
-             cmp_->Equal(ikey_.user_key, next_ikey.user_key) &&
+      // Skip over all versions of this key that happen to occur in the same
+      // snapshot range as the delete.
+      //
+      // Note that a deletion marker of type kTypeDeletionWithTimestamp will be
+      // considered to have a different user key unless the timestamp is older
+      // than *full_history_ts_low_.
+      while (!IsPausingManualCompaction() && !IsShuttingDown() &&
+             input_->Valid() &&
+             (ParseInternalKey(input_->key(), &next_ikey, allow_data_in_errors_)
+                  .ok()) &&
+             0 == cmp_->CompareWithoutTimestamp(ikey_.user_key,
+                                                next_ikey.user_key) &&
              (prev_snapshot == 0 ||
               DEFINITELY_NOT_IN_SNAPSHOT(next_ikey.sequence, prev_snapshot))) {
         input_->Next();
       }
       // 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) &&
-          cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
+      if (input_->Valid() &&
+          (ParseInternalKey(input_->key(), &next_ikey, allow_data_in_errors_)
+               .ok()) &&
+          0 == cmp_->CompareWithoutTimestamp(ikey_.user_key,
+                                             next_ikey.user_key)) {
         valid_ = true;
         at_next_ = true;
       }
@@ -592,8 +668,9 @@ void CompactionIterator::NextFromInput() {
       // have hit (A)
       // We encapsulate the merge related state machine in a different
       // object to minimize change to the existing flow.
-      Status s = merge_helper_->MergeUntil(input_, range_del_agg_,
-                                           prev_snapshot, bottommost_level_);
+      Status s =
+          merge_helper_->MergeUntil(input_, range_del_agg_, prev_snapshot,
+                                    bottommost_level_, allow_data_in_errors_);
       merge_out_iter_.SeekToFirst();
 
       if (!s.ok() && !s.IsMergeInProgress()) {
@@ -604,14 +681,13 @@ 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_);
+        pik_status = ParseInternalKey(key_, &ikey_, allow_data_in_errors_);
         // 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) {
-          ROCKS_LOG_FATAL(info_log_, "Invalid key (%s) in compaction",
-                          key_.ToString(true).c_str());
+        assert(pik_status.ok());
+        if (!pik_status.ok()) {
+          ROCKS_LOG_FATAL(info_log_, "Invalid key in compaction. %s",
+                          pik_status.getState());
         }
         // Keep current_key_ in sync.
         current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
@@ -659,20 +735,37 @@ 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_;
+        }
       }
     }
 
@@ -698,7 +791,18 @@ void CompactionIterator::PrepareOutput() {
                         ikey_.type);
       }
       ikey_.sequence = 0;
-      current_key_.UpdateInternalKey(0, ikey_.type);
+      if (!timestamp_size_) {
+        current_key_.UpdateInternalKey(0, ikey_.type);
+      } else if (full_history_ts_low_ && cmp_with_history_ts_low_ < 0) {
+        // We can also zero out timestamp for better compression.
+        // For the same user key (excluding timestamp), the timestamp-based
+        // history can be collapsed to save some space if the timestamp is
+        // older than *full_history_ts_low_.
+        const std::string kTsMin(timestamp_size_, static_cast<char>(0));
+        const Slice ts_slice = kTsMin;
+        ikey_.SetTimestamp(ts_slice);
+        current_key_.UpdateInternalKey(0, ikey_.type, &ts_slice);
+      }
     }
   }
 }

db/compaction/compaction_iterator.h

@@ -21,72 +21,103 @@
 
 namespace ROCKSDB_NAMESPACE {
 
+class BlobFileBuilder;
+
 class CompactionIterator {
  public:
   // A wrapper around Compaction. Has a much smaller interface, only what
   // CompactionIterator uses. Tests can override it.
   class CompactionProxy {
    public:
-    explicit CompactionProxy(const Compaction* compaction)
-        : compaction_(compaction) {}
-
     virtual ~CompactionProxy() = default;
-    virtual int level(size_t /*compaction_input_level*/ = 0) const {
-      return compaction_->level();
-    }
+
+    virtual int level() const = 0;
+
     virtual bool KeyNotExistsBeyondOutputLevel(
-        const Slice& user_key, std::vector<size_t>* level_ptrs) const {
+        const Slice& user_key, std::vector<size_t>* level_ptrs) const = 0;
+
+    virtual bool bottommost_level() const = 0;
+
+    virtual int number_levels() const = 0;
+
+    virtual Slice GetLargestUserKey() const = 0;
+
+    virtual bool allow_ingest_behind() const = 0;
+
+    virtual bool preserve_deletes() const = 0;
+  };
+
+  class RealCompaction : public CompactionProxy {
+   public:
+    explicit RealCompaction(const Compaction* compaction)
+        : compaction_(compaction) {
+      assert(compaction_);
+      assert(compaction_->immutable_cf_options());
+    }
+
+    int level() const override { return compaction_->level(); }
+
+    bool KeyNotExistsBeyondOutputLevel(
+        const Slice& user_key, std::vector<size_t>* level_ptrs) const override {
       return compaction_->KeyNotExistsBeyondOutputLevel(user_key, level_ptrs);
     }
-    virtual bool bottommost_level() const {
+
+    bool bottommost_level() const override {
       return compaction_->bottommost_level();
     }
-    virtual int number_levels() const { return compaction_->number_levels(); }
-    virtual Slice GetLargestUserKey() const {
+
+    int number_levels() const override { return compaction_->number_levels(); }
+
+    Slice GetLargestUserKey() const override {
       return compaction_->GetLargestUserKey();
     }
-    virtual bool allow_ingest_behind() const {
+
+    bool allow_ingest_behind() const override {
       return compaction_->immutable_cf_options()->allow_ingest_behind;
     }
-    virtual bool preserve_deletes() const {
+
+    bool preserve_deletes() const override {
       return compaction_->immutable_cf_options()->preserve_deletes;
     }
 
-   protected:
-    CompactionProxy() = default;
-
    private:
     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,
+                     const std::string* full_history_ts_low = 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,
+                     const std::string* full_history_ts_low = nullptr);
 
   ~CompactionIterator();
 
@@ -146,6 +177,20 @@ class CompactionIterator {
 
   bool IsInEarliestSnapshot(SequenceNumber sequence);
 
+  // Extract user-defined timestamp from user key if possible and compare it
+  // with *full_history_ts_low_ if applicable.
+  inline void UpdateTimestampAndCompareWithFullHistoryLow() {
+    if (!timestamp_size_) {
+      return;
+    }
+    Slice ts = ExtractTimestampFromUserKey(ikey_.user_key, timestamp_size_);
+    curr_ts_.assign(ts.data(), ts.size());
+    if (full_history_ts_low_) {
+      cmp_with_history_ts_low_ =
+          cmp_->CompareTimestamp(ts, *full_history_ts_low_);
+    }
+  }
+
   InternalIterator* input_;
   const Comparator* cmp_;
   MergeHelper* merge_helper_;
@@ -163,10 +208,11 @@ 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;
@@ -174,6 +220,20 @@ class CompactionIterator {
   SequenceNumber earliest_snapshot_;
   SequenceNumber latest_snapshot_;
 
+  std::shared_ptr<Logger> info_log_;
+
+  bool allow_data_in_errors_;
+
+  // Comes from comparator.
+  const size_t timestamp_size_;
+
+  // Lower bound timestamp to retain full history in terms of user-defined
+  // timestamp. If a key's timestamp is older than full_history_ts_low_, then
+  // the key *may* be eligible for garbage collection (GC). The skipping logic
+  // is in `NextFromInput()` and `PrepareOutput()`.
+  // If nullptr, NO GC will be performed and all history will be preserved.
+  const std::string* const full_history_ts_low_;
+
   // State
   //
   // Points to a copy of the current compaction iterator output (current_key_)
@@ -192,11 +252,13 @@ class CompactionIterator {
   // Stores whether ikey_.user_key is valid. If set to false, the user key is
   // not compared against the current key in the underlying iterator.
   bool has_current_user_key_ = false;
-  bool at_next_ = false;  // If false, the iterator
-  // Holds a copy of the current compaction iterator output (or current key in
-  // the underlying iterator during NextFromInput()).
+  // If false, the iterator holds a copy of the current compaction iterator
+  // output (or current key in the underlying iterator during NextFromInput()).
+  bool at_next_ = false;
+
   IterKey current_key_;
   Slice current_user_key_;
+  std::string curr_ts_;
   SequenceNumber current_user_key_sequence_;
   SequenceNumber current_user_key_snapshot_;
 
@@ -211,6 +273,7 @@ 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
@@ -225,7 +288,9 @@ class CompactionIterator {
   // Used to avoid purging uncommitted values. The application can specify
   // uncommitted values by providing a SnapshotChecker object.
   bool current_key_committed_;
-  std::shared_ptr<Logger> info_log_;
+
+  // Saved result of ucmp->CompareTimestamp(current_ts_, *full_history_ts_low_)
+  int cmp_with_history_ts_low_;
 
   bool IsShuttingDown() {
     // This is a best-effort facility, so memory_order_relaxed is sufficient.
@@ -235,7 +300,7 @@ 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

@@ -156,26 +156,31 @@ class LoggingForwardVectorIterator : public InternalIterator {
 
 class FakeCompaction : public CompactionIterator::CompactionProxy {
  public:
-  FakeCompaction() = default;
+  int level() const override { return 0; }
 
-  int level(size_t /*compaction_input_level*/) const override { return 0; }
   bool KeyNotExistsBeyondOutputLevel(
       const Slice& /*user_key*/,
       std::vector<size_t>* /*level_ptrs*/) const override {
     return is_bottommost_level || key_not_exists_beyond_output_level;
   }
+
   bool bottommost_level() const override { return is_bottommost_level; }
+
   int number_levels() const override { return 1; }
+
   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
@@ -214,6 +219,9 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
   CompactionIteratorTest()
       : cmp_(BytewiseComparator()), icmp_(cmp_), snapshots_({}) {}
 
+  explicit CompactionIteratorTest(const Comparator* ucmp)
+      : cmp_(ucmp), icmp_(cmp_), snapshots_({}) {}
+
   void InitIterators(
       const std::vector<std::string>& ks, const std::vector<std::string>& vs,
       const std::vector<std::string>& range_del_ks,
@@ -222,7 +230,9 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
       SequenceNumber last_committed_sequence = kMaxSequenceNumber,
       MergeOperator* merge_op = nullptr, CompactionFilter* filter = nullptr,
       bool bottommost_level = false,
-      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
+      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber,
+      bool key_not_exists_beyond_output_level = false,
+      const std::string* full_history_ts_low = nullptr) {
     std::unique_ptr<InternalIterator> unfragmented_range_del_iter(
         new test::VectorIterator(range_del_ks, range_del_vs));
     auto tombstone_list = std::make_shared<FragmentedRangeTombstoneList>(
@@ -234,9 +244,12 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
     range_del_agg_->AddTombstones(std::move(range_del_iter));
 
     std::unique_ptr<CompactionIterator::CompactionProxy> compaction;
-    if (filter || bottommost_level) {
+    if (filter || bottommost_level || key_not_exists_beyond_output_level) {
       compaction_proxy_ = new FakeCompaction();
       compaction_proxy_->is_bottommost_level = bottommost_level;
+      compaction_proxy_->is_allow_ingest_behind = AllowIngestBehind();
+      compaction_proxy_->key_not_exists_beyond_output_level =
+          key_not_exists_beyond_output_level;
       compaction.reset(compaction_proxy_);
     }
     bool use_snapshot_checker = UseSnapshotChecker() || GetParam();
@@ -249,13 +262,22 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
                         0 /*latest_snapshot*/, snapshot_checker_.get(),
                         0 /*level*/, nullptr /*statistics*/, &shutting_down_));
 
+    if (c_iter_) {
+      // Since iter_ is still used in ~CompactionIterator(), we call
+      // ~CompactionIterator() first.
+      c_iter_.reset();
+    }
     iter_.reset(new LoggingForwardVectorIterator(ks, vs));
     iter_->SeekToFirst();
     c_iter_.reset(new CompactionIterator(
         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 */,
+        true /*allow_data_in_errors*/, std::move(compaction), filter,
+        &shutting_down_, /*preserve_deletes_seqnum=*/0,
+        /*manual_compaction_paused=*/nullptr, /*info_log=*/nullptr,
+        full_history_ts_low));
   }
 
   void AddSnapshot(SequenceNumber snapshot,
@@ -266,6 +288,8 @@ 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,
@@ -275,10 +299,13 @@ class CompactionIteratorTest : public testing::TestWithParam<bool> {
       MergeOperator* merge_operator = nullptr,
       CompactionFilter* compaction_filter = nullptr,
       bool bottommost_level = false,
-      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
+      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber,
+      bool key_not_exists_beyond_output_level = false,
+      const std::string* full_history_ts_low = nullptr) {
     InitIterators(input_keys, input_values, {}, {}, kMaxSequenceNumber,
                   last_committed_seq, merge_operator, compaction_filter,
-                  bottommost_level, earliest_write_conflict_snapshot);
+                  bottommost_level, earliest_write_conflict_snapshot,
+                  key_not_exists_beyond_output_level, full_history_ts_low);
     c_iter_->SeekToFirst();
     for (size_t i = 0; i < expected_keys.size(); i++) {
       std::string info = "i = " + ToString(i);
@@ -288,9 +315,15 @@ 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());
   }
 
+  void ClearSnapshots() {
+    snapshots_.clear();
+    snapshot_map_.clear();
+  }
+
   const Comparator* cmp_;
   const InternalKeyComparator icmp_;
   std::vector<SequenceNumber> snapshots_;
@@ -312,6 +345,7 @@ 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 +367,7 @@ 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 +384,7 @@ 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 +406,7 @@ 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 +500,7 @@ 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 +694,7 @@ 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 +736,18 @@ 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 +1019,227 @@ 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));
+
+class CompactionIteratorTsGcTest : public CompactionIteratorTest {
+ public:
+  CompactionIteratorTsGcTest()
+      : CompactionIteratorTest(test::ComparatorWithU64Ts()) {}
+};
+
+TEST_P(CompactionIteratorTsGcTest, NoKeyEligibleForGC) {
+  constexpr char user_key[][2] = {{'a', '\0'}, {'b', '\0'}};
+  const std::vector<std::string> input_keys = {
+      test::KeyStr(/*ts=*/103, user_key[0], /*seq=*/4, kTypeValue),
+      test::KeyStr(/*ts=*/102, user_key[0], /*seq=*/3,
+                   kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/104, user_key[1], /*seq=*/5, kTypeValue)};
+  const std::vector<std::string> input_values = {"a3", "", "b2"};
+  std::string full_history_ts_low;
+  // All keys' timestamps are newer than or equal to 102, thus none of them
+  // will be eligible for GC.
+  PutFixed64(&full_history_ts_low, 102);
+  const std::vector<std::string>& expected_keys = input_keys;
+  const std::vector<std::string>& expected_values = input_values;
+  const std::vector<std::pair<bool, bool>> params = {
+      {false, false}, {false, true}, {true, true}};
+  for (const std::pair<bool, bool>& param : params) {
+    const bool bottommost_level = param.first;
+    const bool key_not_exists_beyond_output_level = param.second;
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            bottommost_level,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            key_not_exists_beyond_output_level, &full_history_ts_low);
+  }
+}
+
+TEST_P(CompactionIteratorTsGcTest, AllKeysOlderThanThreshold) {
+  constexpr char user_key[][2] = {{'a', '\0'}, {'b', '\0'}};
+  const std::vector<std::string> input_keys = {
+      test::KeyStr(/*ts=*/103, user_key[0], /*seq=*/4,
+                   kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/102, user_key[0], /*seq=*/3, kTypeValue),
+      test::KeyStr(/*ts=*/101, user_key[0], /*seq=*/2, kTypeValue),
+      test::KeyStr(/*ts=*/104, user_key[1], /*seq=*/5, kTypeValue)};
+  const std::vector<std::string> input_values = {"", "a2", "a1", "b5"};
+  std::string full_history_ts_low;
+  PutFixed64(&full_history_ts_low, std::numeric_limits<uint64_t>::max());
+  {
+    // With a snapshot at seq 3, both the deletion marker and the key at 3 must
+    // be preserved.
+    AddSnapshot(3);
+    const std::vector<std::string> expected_keys = {
+        input_keys[0], input_keys[1], input_keys[3]};
+    const std::vector<std::string> expected_values = {"", "a2", "b5"};
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/false,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/false, &full_history_ts_low);
+    ClearSnapshots();
+  }
+  {
+    // No snapshot, the deletion marker should be preserved because the user
+    // key may appear beyond output level.
+    const std::vector<std::string> expected_keys = {input_keys[0],
+                                                    input_keys[3]};
+    const std::vector<std::string> expected_values = {"", "b5"};
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/false,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/false, &full_history_ts_low);
+  }
+  {
+    // No snapshot, the deletion marker can be dropped because the user key
+    // does not appear in higher levels.
+    const std::vector<std::string> expected_keys = {input_keys[3]};
+    const std::vector<std::string> expected_values = {"b5"};
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/false,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/true, &full_history_ts_low);
+  }
+}
+
+TEST_P(CompactionIteratorTsGcTest, NewHidesOldSameSnapshot) {
+  constexpr char user_key[] = "a";
+  const std::vector<std::string> input_keys = {
+      test::KeyStr(/*ts=*/103, user_key, /*seq=*/4, kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/102, user_key, /*seq=*/3, kTypeValue),
+      test::KeyStr(/*ts=*/101, user_key, /*seq=*/2, kTypeValue),
+      test::KeyStr(/*ts=*/100, user_key, /*seq=*/1, kTypeValue)};
+  const std::vector<std::string> input_values = {"", "a2", "a1", "a0"};
+  {
+    std::string full_history_ts_low;
+    // Keys whose timestamps larger than or equal to 102 will be preserved.
+    PutFixed64(&full_history_ts_low, 102);
+    const std::vector<std::string> expected_keys = {input_keys[0],
+                                                    input_keys[1]};
+    const std::vector<std::string> expected_values = {"", "a2"};
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/false,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/false, &full_history_ts_low);
+  }
+}
+
+TEST_P(CompactionIteratorTsGcTest, DropTombstones) {
+  constexpr char user_key[] = "a";
+  const std::vector<std::string> input_keys = {
+      test::KeyStr(/*ts=*/103, user_key, /*seq=*/4, kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/102, user_key, /*seq=*/3, kTypeValue),
+      test::KeyStr(/*ts=*/101, user_key, /*seq=*/2, kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/100, user_key, /*seq=*/1, kTypeValue)};
+  const std::vector<std::string> input_values = {"", "a2", "", "a0"};
+  const std::vector<std::string> expected_keys = {input_keys[0], input_keys[1]};
+  const std::vector<std::string> expected_values = {"", "a2"};
+
+  // Take a snapshot at seq 2.
+  AddSnapshot(2);
+
+  {
+    // Non-bottommost level, but key does not exist beyond output level.
+    std::string full_history_ts_low;
+    PutFixed64(&full_history_ts_low, 102);
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_sequence=*/kMaxSequenceNumber,
+            /*merge_op=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/false,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/true, &full_history_ts_low);
+  }
+  {
+    // Bottommost level
+    std::string full_history_ts_low;
+    PutFixed64(&full_history_ts_low, 102);
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/true,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/false, &full_history_ts_low);
+  }
+}
+
+TEST_P(CompactionIteratorTsGcTest, RewriteTs) {
+  constexpr char user_key[] = "a";
+  const std::vector<std::string> input_keys = {
+      test::KeyStr(/*ts=*/103, user_key, /*seq=*/4, kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/102, user_key, /*seq=*/3, kTypeValue),
+      test::KeyStr(/*ts=*/101, user_key, /*seq=*/2, kTypeDeletionWithTimestamp),
+      test::KeyStr(/*ts=*/100, user_key, /*seq=*/1, kTypeValue)};
+  const std::vector<std::string> input_values = {"", "a2", "", "a0"};
+  const std::vector<std::string> expected_keys = {
+      input_keys[0], input_keys[1], input_keys[2],
+      test::KeyStr(/*ts=*/0, user_key, /*seq=*/0, kTypeValue)};
+  const std::vector<std::string> expected_values = {"", "a2", "", "a0"};
+
+  AddSnapshot(1);
+  AddSnapshot(2);
+
+  {
+    // Bottommost level and need to rewrite both ts and seq.
+    std::string full_history_ts_low;
+    PutFixed64(&full_history_ts_low, 102);
+    RunTest(input_keys, input_values, expected_keys, expected_values,
+            /*last_committed_seq=*/kMaxSequenceNumber,
+            /*merge_operator=*/nullptr, /*compaction_filter=*/nullptr,
+            /*bottommost_level=*/true,
+            /*earliest_write_conflict_snapshot=*/kMaxSequenceNumber,
+            /*key_not_exists_beyond_output_level=*/true, &full_history_ts_low);
+  }
+}
+
+INSTANTIATE_TEST_CASE_P(CompactionIteratorTsGcTestInstance,
+                        CompactionIteratorTsGcTest,
+                        testing::Values(true, false));
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/compaction/compaction_job.cc

@@ -20,6 +20,8 @@
 #include <utility>
 #include <vector>
 
+#include "db/blob/blob_file_addition.h"
+#include "db/blob/blob_file_builder.h"
 #include "db/builder.h"
 #include "db/db_impl/db_impl.h"
 #include "db/db_iter.h"
@@ -32,6 +34,7 @@
 #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"
@@ -124,16 +127,23 @@ struct CompactionJob::SubcompactionState {
 
   // 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;
-    uint64_t paranoid_hash;
     std::shared_ptr<const TableProperties> table_properties;
   };
 
   // State kept for output being generated
   std::vector<Output> outputs;
+  std::vector<BlobFileAddition> blob_file_additions;
   std::unique_ptr<WritableFileWriter> outfile;
   std::unique_ptr<TableBuilder> builder;
+
   Output* current_output() {
     if (outputs.empty()) {
       // This subcompaction's output could be empty if compaction was aborted
@@ -147,13 +157,13 @@ 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
@@ -162,63 +172,23 @@ 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);
-    io_status = std::move(o.io_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;
-  }
-
-  // Because member std::unique_ptrs do not have these.
-  SubcompactionState(const SubcompactionState&) = delete;
-
-  SubcompactionState& operator=(const SubcompactionState&) = delete;
-
   // Adds the key and value to the builder
   // If paranoid is true, adds the key-value to the paranoid hash
-  void AddToBuilder(const Slice& key, const Slice& value, bool paranoid) {
+  Status AddToBuilder(const Slice& key, const Slice& value) {
     auto curr = current_output();
     assert(builder != nullptr);
     assert(curr != nullptr);
-    if (paranoid) {
-      // Generate a rolling 64-bit hash of the key and values
-      curr->paranoid_hash = Hash64(key.data(), key.size(), curr->paranoid_hash);
-      curr->paranoid_hash =
-          Hash64(value.data(), value.size(), curr->paranoid_hash);
+    Status s = curr->validator.Add(key, value);
+    if (!s.ok()) {
+      return s;
     }
     builder->Add(key, value);
+    return Status::OK();
   }
 
   // Returns true iff we should stop building the current output
@@ -264,21 +234,13 @@ struct CompactionJob::CompactionState {
   std::vector<CompactionJob::SubcompactionState> sub_compact_states;
   Status status;
 
-  uint64_t total_bytes;
-  uint64_t num_output_records;
+  size_t num_output_files = 0;
+  uint64_t total_bytes = 0;
+  size_t num_blob_output_files = 0;
+  uint64_t total_blob_bytes = 0;
+  uint64_t num_output_records = 0;
 
-  explicit CompactionState(Compaction* c)
-      : compaction(c),
-        total_bytes(0),
-        num_output_records(0) {}
-
-  size_t NumOutputFiles() {
-    size_t total = 0;
-    for (auto& s : sub_compact_states) {
-      total += s.outputs.size();
-    }
-    return total;
-  }
+  explicit CompactionState(Compaction* c) : compaction(c) {}
 
   Slice SmallestUserKey() {
     for (const auto& sub_compact_state : sub_compact_states) {
@@ -305,14 +267,30 @@ struct CompactionJob::CompactionState {
 };
 
 void CompactionJob::AggregateStatistics() {
+  assert(compact_);
+
   for (SubcompactionState& sc : compact_->sub_compact_states) {
-    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);
+    auto& outputs = sc.outputs;
+
+    if (!outputs.empty() && !outputs.back().meta.fd.file_size) {
+      // An error occurred, so ignore the last output.
+      outputs.pop_back();
     }
+
+    compact_->num_output_files += outputs.size();
+    compact_->total_bytes += sc.total_bytes;
+
+    const auto& blobs = sc.blob_file_additions;
+
+    compact_->num_blob_output_files += blobs.size();
+
+    for (const auto& blob : blobs) {
+      compact_->total_blob_bytes += blob.GetTotalBlobBytes();
+    }
+
+    compact_->num_output_records += sc.num_output_records;
+
+    compaction_job_stats_->Add(sc.compaction_job_stats);
   }
 }
 
@@ -321,15 +299,17 @@ CompactionJob::CompactionJob(
     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,
+    FSDirectory* db_directory, FSDirectory* output_directory,
+    FSDirectory* blob_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,
-    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, std::string full_history_ts_low)
     : job_id_(job_id),
       compact_(new CompactionState(compaction)),
       compaction_job_stats_(compaction_job_stats),
@@ -340,7 +320,8 @@ 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),
@@ -350,6 +331,7 @@ CompactionJob::CompactionJob(
       log_buffer_(log_buffer),
       db_directory_(db_directory),
       output_directory_(output_directory),
+      blob_output_directory_(blob_output_directory),
       stats_(stats),
       db_mutex_(db_mutex),
       db_error_handler_(db_error_handler),
@@ -362,7 +344,9 @@ CompactionJob::CompactionJob(
       paranoid_file_checks_(paranoid_file_checks),
       measure_io_stats_(measure_io_stats),
       write_hint_(Env::WLTH_NOT_SET),
-      thread_pri_(thread_pri) {
+      thread_pri_(thread_pri),
+      full_history_ts_low_(std::move(full_history_ts_low)) {
+  assert(compaction_job_stats_ != nullptr);
   assert(log_buffer_ != nullptr);
   const auto* cfd = compact_->compaction->column_family_data();
   ThreadStatusUtil::SetColumnFamily(cfd, cfd->ioptions()->env,
@@ -414,10 +398,9 @@ 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() {
@@ -449,7 +432,11 @@ 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);
   }
 }
 
@@ -633,18 +620,34 @@ Status CompactionJob::Run() {
   // Check if any thread encountered an error during execution
   Status status;
   IOStatus io_s;
+  bool wrote_new_blob_files = false;
+
   for (const auto& state : compact_->sub_compact_states) {
     if (!state.status.ok()) {
       status = state.status;
       io_s = state.io_status;
       break;
     }
+
+    if (!state.blob_file_additions.empty()) {
+      wrote_new_blob_files = true;
+    }
   }
+
   if (io_status_.ok()) {
     io_status_ = io_s;
   }
-  if (status.ok() && output_directory_) {
-    io_s = output_directory_->Fsync(IOOptions(), nullptr);
+  if (status.ok()) {
+    constexpr IODebugContext* dbg = nullptr;
+
+    if (output_directory_) {
+      io_s = output_directory_->Fsync(IOOptions(), dbg);
+    }
+
+    if (io_s.ok() && wrote_new_blob_files && blob_output_directory_ &&
+        blob_output_directory_ != output_directory_) {
+      io_s = blob_output_directory_->Fsync(IOOptions(), dbg);
+    }
   }
   if (io_status_.ok()) {
     io_status_ = io_s;
@@ -676,8 +679,9 @@ 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(),
+            read_options, file_options_, cfd->internal_comparator(),
             files_output[file_idx]->meta, /*range_del_agg=*/nullptr,
             prefix_extractor,
             /*table_reader_ptr=*/nullptr,
@@ -693,15 +697,21 @@ Status CompactionJob::Run() {
         auto s = iter->status();
 
         if (s.ok() && paranoid_file_checks_) {
-          uint64_t hash = 0;
+          OutputValidator validator(cfd->internal_comparator(),
+                                    /*_enable_order_check=*/true,
+                                    /*_enable_hash=*/true);
           for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
-            // Generate a rolling 64-bit hash of the key and values, using the
-            hash = Hash64(iter->key().data(), iter->key().size(), hash);
-            hash = Hash64(iter->value().data(), iter->value().size(), hash);
+            s = validator.Add(iter->key(), iter->value());
+            if (!s.ok()) {
+              break;
+            }
           }
-          s = iter->status();
-          if (s.ok() && hash != files_output[file_idx]->paranoid_hash) {
-            s = Status::Corruption("Paraniod checksums do not match");
+          if (s.ok()) {
+            s = iter->status();
+          }
+          if (s.ok() &&
+              !validator.CompareValidator(files_output[file_idx]->validator)) {
+            s = Status::Corruption("Paranoid checksums do not match");
           }
         }
 
@@ -743,6 +753,7 @@ Status CompactionJob::Run() {
   // Finish up all book-keeping to unify the subcompaction results
   AggregateStatistics();
   UpdateCompactionStats();
+
   RecordCompactionIOStats();
   LogFlush(db_options_.info_log);
   TEST_SYNC_POINT("CompactionJob::Run():End");
@@ -752,11 +763,16 @@ Status CompactionJob::Run() {
 }
 
 Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
+  assert(compact_);
+
   AutoThreadOperationStageUpdater stage_updater(
       ThreadStatus::STAGE_COMPACTION_INSTALL);
   db_mutex_->AssertHeld();
   Status status = compact_->status;
+
   ColumnFamilyData* cfd = compact_->compaction->column_family_data();
+  assert(cfd);
+
   cfd->internal_stats()->AddCompactionStats(
       compact_->compaction->output_level(), thread_pri_, compaction_stats_);
 
@@ -766,6 +782,7 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
   if (!versions_->io_status().ok()) {
     io_status_ = versions_->io_status();
   }
+
   VersionStorageInfo::LevelSummaryStorage tmp;
   auto vstorage = cfd->current()->storage_info();
   const auto& stats = compaction_stats_;
@@ -790,6 +807,8 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
         stats.bytes_written / static_cast<double>(stats.micros);
   }
 
+  const std::string& column_family_name = cfd->GetName();
+
   ROCKS_LOG_BUFFER(
       log_buffer_,
       "[%s] compacted to: %s, MB/sec: %.1f rd, %.1f wr, level %d, "
@@ -797,8 +816,9 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
       "MB in(%.1f, %.1f) out(%.1f), read-write-amplify(%.1f) "
       "write-amplify(%.1f) %s, records in: %" PRIu64
       ", records dropped: %" PRIu64 " output_compression: %s\n",
-      cfd->GetName().c_str(), vstorage->LevelSummary(&tmp), bytes_read_per_sec,
-      bytes_written_per_sec, compact_->compaction->output_level(),
+      column_family_name.c_str(), vstorage->LevelSummary(&tmp),
+      bytes_read_per_sec, bytes_written_per_sec,
+      compact_->compaction->output_level(),
       stats.num_input_files_in_non_output_levels,
       stats.num_input_files_in_output_level, stats.num_output_files,
       stats.bytes_read_non_output_levels / 1048576.0,
@@ -809,6 +829,15 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
       CompressionTypeToString(compact_->compaction->output_compression())
           .c_str());
 
+  const auto& blob_files = vstorage->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);
+  }
+
   UpdateCompactionJobStats(stats);
 
   auto stream = event_logger_->LogToBuffer(log_buffer_);
@@ -817,21 +846,26 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
          << "compaction_time_micros" << stats.micros
          << "compaction_time_cpu_micros" << stats.cpu_micros << "output_level"
          << compact_->compaction->output_level() << "num_output_files"
-         << compact_->NumOutputFiles() << "total_output_size"
-         << compact_->total_bytes << "num_input_records"
-         << stats.num_input_records << "num_output_records"
-         << compact_->num_output_records << "num_subcompactions"
-         << compact_->sub_compact_states.size() << "output_compression"
-         << CompressionTypeToString(compact_->compaction->output_compression());
+         << compact_->num_output_files << "total_output_size"
+         << compact_->total_bytes;
 
-  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;
+  if (compact_->num_blob_output_files > 0) {
+    stream << "num_blob_output_files" << compact_->num_blob_output_files
+           << "total_blob_output_size" << compact_->total_blob_bytes;
   }
 
-  if (measure_io_stats_ && compaction_job_stats_ != nullptr) {
+  stream << "num_input_records" << stats.num_input_records
+         << "num_output_records" << compact_->num_output_records
+         << "num_subcompactions" << compact_->sub_compact_states.size()
+         << "output_compression"
+         << CompressionTypeToString(compact_->compaction->output_compression());
+
+  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_) {
     stream << "file_write_nanos" << compaction_job_stats_->file_write_nanos;
     stream << "file_range_sync_nanos"
            << compaction_job_stats_->file_range_sync_nanos;
@@ -847,12 +881,18 @@ Status CompactionJob::Install(const MutableCFOptions& mutable_cf_options) {
   }
   stream.EndArray();
 
+  if (!blob_files.empty()) {
+    stream << "blob_file_head" << blob_files.begin()->first;
+    stream << "blob_file_tail" << blob_files.rbegin()->first;
+  }
+
   CleanupCompaction();
   return status;
 }
 
 void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
-  assert(sub_compact != nullptr);
+  assert(sub_compact);
+  assert(sub_compact->compaction);
 
   uint64_t prev_cpu_micros = env_->NowCPUNanos() / 1000;
 
@@ -877,11 +917,20 @@ 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);
@@ -914,11 +963,27 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
       snapshot_checker_, compact_->compaction->level(),
       db_options_.statistics.get());
 
+  const MutableCFOptions* mutable_cf_options =
+      sub_compact->compaction->mutable_cf_options();
+  assert(mutable_cf_options);
+
+  std::vector<std::string> blob_file_paths;
+
+  std::unique_ptr<BlobFileBuilder> blob_file_builder(
+      mutable_cf_options->enable_blob_files
+          ? new BlobFileBuilder(
+                versions_, env_, fs_.get(),
+                sub_compact->compaction->immutable_cf_options(),
+                mutable_cf_options, &file_options_, job_id_, cfd->GetID(),
+                cfd->GetName(), Env::IOPriority::IO_LOW, write_hint_,
+                &blob_file_paths, &sub_compact->blob_file_additions)
+          : nullptr);
+
   TEST_SYNC_POINT("CompactionJob::Run():Inprogress");
   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;
@@ -931,14 +996,17 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
   }
 
   Status status;
+  const std::string* const full_history_ts_low =
+      full_history_ts_low_.empty() ? nullptr : &full_history_ts_low_;
   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_),
       /*expect_valid_internal_key=*/true, &range_del_agg,
+      blob_file_builder.get(), db_options_.allow_data_in_errors,
       sub_compact->compaction, compaction_filter, shutting_down_,
-      preserve_deletes_seqnum_, manual_compaction_paused_,
-      db_options_.info_log));
+      preserve_deletes_seqnum_, manual_compaction_paused_, db_options_.info_log,
+      full_history_ts_low));
   auto c_iter = sub_compact->c_iter.get();
   c_iter->SeekToFirst();
   if (c_iter->Valid() && sub_compact->compaction->output_level() != 0) {
@@ -982,7 +1050,10 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
         break;
       }
     }
-    sub_compact->AddToBuilder(key, value, paranoid_file_checks_);
+    status = sub_compact->AddToBuilder(key, value);
+    if (!status.ok()) {
+      break;
+    }
 
     sub_compact->current_output_file_size =
         sub_compact->builder->EstimatedFileSize();
@@ -1000,19 +1071,17 @@ 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_);
@@ -1033,7 +1102,6 @@ void CompactionJob::ProcessKeyValueCompaction(SubcompactionState* sub_compact) {
         // (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;
       }
     }
@@ -1043,9 +1111,9 @@ 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);
     }
@@ -1079,7 +1147,7 @@ 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()) {
@@ -1101,12 +1169,20 @@ 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);
   }
 
+  if (blob_file_builder) {
+    if (status.ok()) {
+      status = blob_file_builder->Finish();
+    }
+
+    blob_file_builder.reset();
+  }
+
   sub_compact->compaction_job_stats.cpu_micros =
       env_->NowCPUNanos() / 1000 - prev_cpu_micros;
 
@@ -1127,6 +1203,16 @@ 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();
@@ -1187,6 +1273,8 @@ 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;
@@ -1288,8 +1376,8 @@ Status CompactionJob::FinishCompactionOutputFile(
       auto kv = tombstone.Serialize();
       assert(lower_bound == nullptr ||
              ucmp->Compare(*lower_bound, kv.second) < 0);
-      sub_compact->AddToBuilder(kv.first.Encode(), kv.second,
-                                paranoid_file_checks_);
+      // 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 &&
           ucmp->Compare(smallest_candidate.user_key(), *lower_bound) <= 0) {
@@ -1354,7 +1442,6 @@ Status CompactionJob::FinishCompactionOutputFile(
              ExtractInternalKeyFooter(meta->smallest.Encode()) !=
                  PackSequenceAndType(0, kTypeRangeDeletion));
     }
-    meta->marked_for_compaction = sub_compact->builder->NeedCompact();
   }
   const uint64_t current_entries = sub_compact->builder->NumEntries();
   if (s.ok()) {
@@ -1369,6 +1456,7 @@ Status CompactionJob::FinishCompactionOutputFile(
   const uint64_t current_bytes = sub_compact->builder->FileSize();
   if (s.ok()) {
     meta->fd.file_size = current_bytes;
+    meta->marked_for_compaction = sub_compact->builder->NeedCompact();
   }
   sub_compact->current_output()->finished = true;
   sub_compact->total_bytes += current_bytes;
@@ -1386,12 +1474,17 @@ Status CompactionJob::FinishCompactionOutputFile(
     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 (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();
 
@@ -1442,14 +1535,18 @@ 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)?
@@ -1458,7 +1555,9 @@ 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
@@ -1470,9 +1569,13 @@ Status CompactionJob::FinishCompactionOutputFile(
 
 Status CompactionJob::InstallCompactionResults(
     const MutableCFOptions& mutable_cf_options) {
+  assert(compact_);
+
   db_mutex_->AssertHeld();
 
   auto* compaction = compact_->compaction;
+  assert(compaction);
+
   // paranoia: verify that the files that we started with
   // still exist in the current version and in the same original level.
   // This ensures that a concurrent compaction did not erroneously
@@ -1488,31 +1591,52 @@ Status CompactionJob::InstallCompactionResults(
 
   {
     Compaction::InputLevelSummaryBuffer inputs_summary;
-    ROCKS_LOG_INFO(
-        db_options_.info_log, "[%s] [JOB %d] Compacted %s => %" PRIu64 " bytes",
-        compaction->column_family_data()->GetName().c_str(), job_id_,
-        compaction->InputLevelSummary(&inputs_summary), compact_->total_bytes);
+    ROCKS_LOG_INFO(db_options_.info_log,
+                   "[%s] [JOB %d] Compacted %s => %" PRIu64 " bytes",
+                   compaction->column_family_data()->GetName().c_str(), job_id_,
+                   compaction->InputLevelSummary(&inputs_summary),
+                   compact_->total_bytes + compact_->total_blob_bytes);
   }
 
+  VersionEdit* const edit = compaction->edit();
+  assert(edit);
+
   // Add compaction inputs
-  compaction->AddInputDeletions(compact_->compaction->edit());
+  compaction->AddInputDeletions(edit);
 
   for (const auto& sub_compact : compact_->sub_compact_states) {
     for (const auto& out : sub_compact.outputs) {
-      compaction->edit()->AddFile(compaction->output_level(), out.meta);
+      edit->AddFile(compaction->output_level(), out.meta);
+    }
+
+    for (const auto& blob : sub_compact.blob_file_additions) {
+      edit->AddBlobFile(blob);
     }
   }
+
   return versions_->LogAndApply(compaction->column_family_data(),
-                                mutable_cf_options, compaction->edit(),
-                                db_mutex_, db_directory_);
+                                mutable_cf_options, edit, db_mutex_,
+                                db_directory_);
 }
 
 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);
@@ -1542,10 +1666,14 @@ Status CompactionJob::OpenCompactionOutputFile(
                            &syncpoint_arg);
 #endif
   Status s;
-  IOStatus io_s = NewWritableFile(fs_, fname, &writable_file, file_options_);
+  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(
@@ -1558,7 +1686,8 @@ 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;
   }
 
@@ -1580,14 +1709,17 @@ 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;
-    out.paranoid_hash = 0;
-    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);
@@ -1596,10 +1728,10 @@ 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
@@ -1640,6 +1772,9 @@ 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;
@@ -1657,6 +1792,8 @@ void CopyPrefix(const Slice& src, size_t prefix_length, std::string* dst) {
 #endif  // !ROCKSDB_LITE
 
 void CompactionJob::UpdateCompactionStats() {
+  assert(compact_);
+
   Compaction* compaction = compact_->compaction;
   compaction_stats_.num_input_files_in_non_output_levels = 0;
   compaction_stats_.num_input_files_in_output_level = 0;
@@ -1674,27 +1811,15 @@ void CompactionJob::UpdateCompactionStats() {
     }
   }
 
-  uint64_t num_output_records = 0;
+  compaction_stats_.num_output_files =
+      static_cast<int>(compact_->num_output_files) +
+      static_cast<int>(compact_->num_blob_output_files);
+  compaction_stats_.bytes_written =
+      compact_->total_bytes + compact_->total_blob_bytes;
 
-  for (const auto& sub_compact : compact_->sub_compact_states) {
-    size_t num_output_files = sub_compact.outputs.size();
-    if (sub_compact.builder != nullptr) {
-      // An error occurred so ignore the last output.
-      assert(num_output_files > 0);
-      --num_output_files;
-    }
-    compaction_stats_.num_output_files += static_cast<int>(num_output_files);
-
-    num_output_records += sub_compact.num_output_records;
-
-    for (const auto& out : sub_compact.outputs) {
-      compaction_stats_.bytes_written += out.meta.fd.file_size;
-    }
-  }
-
-  if (compaction_stats_.num_input_records > num_output_records) {
+  if (compaction_stats_.num_input_records > compact_->num_output_records) {
     compaction_stats_.num_dropped_records =
-        compaction_stats_.num_input_records - num_output_records;
+        compaction_stats_.num_input_records - compact_->num_output_records;
   }
 }
 
@@ -1716,32 +1841,29 @@ 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 (stats.num_output_files > 0) {
+    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 +62,24 @@ 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,
+      FSDirectory* blob_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 = "",
+      std::string full_history_ts_low = "");
 
   ~CompactionJob();
 
@@ -160,16 +159,18 @@ 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_;
   FSDirectory* output_directory_;
+  FSDirectory* blob_output_directory_;
   Statistics* stats_;
   InstrumentedMutex* db_mutex_;
   ErrorHandler* db_error_handler_;
@@ -201,6 +202,7 @@ class CompactionJob {
   Env::WriteLifeTimeHint write_hint_;
   Env::Priority thread_pri_;
   IOStatus io_status_;
+  std::string full_history_ts_low_;
 };
 
 }  // namespace ROCKSDB_NAMESPACE

db/compaction/compaction_job_stats_test.cc

@@ -24,7 +24,6 @@
 #include "db/write_batch_internal.h"
 #include "env/mock_env.h"
 #include "file/filename.h"
-#include "logging/logging.h"
 #include "memtable/hash_linklist_rep.h"
 #include "monitoring/statistics.h"
 #include "monitoring/thread_status_util.h"
@@ -459,6 +458,7 @@ 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);
 
@@ -571,7 +571,7 @@ 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();
 
@@ -595,6 +595,7 @@ 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;
@@ -796,7 +797,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
     }
 
     ASSERT_OK(Flush(1));
-    static_cast_with_check<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);
@@ -894,7 +895,7 @@ 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
@@ -981,26 +982,21 @@ 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);
 
@@ -1011,7 +1007,7 @@ TEST_P(CompactionJobStatsTest, UniversalCompactionTest) {
         &rnd, start_key, start_key + key_base - 1,
         kKeySize, kValueSize, key_interval,
         compression_ratio, 1);
-    static_cast_with_check<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

@@ -67,31 +67,37 @@ void VerifyInitializationOfCompactionJobStats(
 
 }  // namespace
 
-// TODO(icanadi) Make it simpler once we mock out VersionSet
-class CompactionJobTest : public testing::Test {
- public:
-  CompactionJobTest()
+class CompactionJobTestBase : public testing::Test {
+ protected:
+  CompactionJobTestBase(std::string dbname, const Comparator* ucmp,
+                        std::function<std::string(uint64_t)> encode_u64_ts)
       : env_(Env::Default()),
         fs_(std::make_shared<LegacyFileSystemWrapper>(env_)),
-        dbname_(test::PerThreadDBPath("compaction_job_test")),
+        dbname_(std::move(dbname)),
+        ucmp_(ucmp),
         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()),
-        error_handler_(nullptr, db_options_, &mutex_) {
+        error_handler_(nullptr, db_options_, &mutex_),
+        encode_u64_ts_(std::move(encode_u64_ts)) {}
+
+  void SetUp() override {
     EXPECT_OK(env_->CreateDirIfMissing(dbname_));
     db_options_.env = env_;
     db_options_.fs = fs_;
     db_options_.db_paths.emplace_back(dbname_,
                                       std::numeric_limits<uint64_t>::max());
+    cf_options_.comparator = ucmp_;
+    cf_options_.table_factory = mock_table_factory_;
   }
 
   std::string GenerateFileName(uint64_t file_number) {
@@ -102,9 +108,10 @@ class CompactionJobTest : public testing::Test {
     return TableFileName(db_paths, meta.fd.GetNumber(), meta.fd.GetPathId());
   }
 
-  static std::string KeyStr(const std::string& user_key,
-                            const SequenceNumber seq_num, const ValueType t) {
-    return InternalKey(user_key, seq_num, t).Encode().ToString();
+  std::string KeyStr(const std::string& user_key, const SequenceNumber seq_num,
+                     const ValueType t, uint64_t ts = 0) {
+    std::string user_key_with_ts = user_key + encode_u64_ts_(ts);
+    return InternalKey(user_key_with_ts, seq_num, t).Encode().ToString();
   }
 
   static std::string BlobStr(uint64_t blob_file_number, uint64_t offset,
@@ -130,7 +137,7 @@ 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;
@@ -144,7 +151,8 @@ class CompactionJobTest : public testing::Test {
       std::string skey;
       std::string value;
       std::tie(skey, value) = kv;
-      bool parsed = ParseInternalKey(skey, &key);
+      const Status pik_status =
+          ParseInternalKey(skey, &key, true /* log_err_key */);
 
       smallest_seqno = std::min(smallest_seqno, key.sequence);
       largest_seqno = std::max(largest_seqno, key.sequence);
@@ -162,7 +170,7 @@ class CompactionJobTest : public testing::Test {
 
       first_key = false;
 
-      if (parsed && key.type == kTypeBlobIndex) {
+      if (pik_status.ok() && key.type == kTypeBlobIndex) {
         BlobIndex blob_index;
         const Status s = blob_index.DecodeFrom(value);
         if (!s.ok()) {
@@ -192,8 +200,9 @@ 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();
   }
 
@@ -204,11 +213,11 @@ class CompactionJobTest : public testing::Test {
   }
 
   // returns expected result after compaction
-  stl_wrappers::KVMap CreateTwoFiles(bool gen_corrupted_keys) {
-    auto expected_results = mock::MakeMockFile();
-    const int kKeysPerFile = 10000;
-    const int kCorruptKeysPerFile = 200;
-    const int kMatchingKeys = kKeysPerFile / 2;
+  mock::KVVector CreateTwoFiles(bool gen_corrupted_keys) {
+    stl_wrappers::KVMap expected_results;
+    constexpr int kKeysPerFile = 10000;
+    constexpr int kCorruptKeysPerFile = 200;
+    constexpr int kMatchingKeys = kKeysPerFile / 2;
     SequenceNumber sequence_number = 0;
 
     auto corrupt_id = [&](int id) {
@@ -231,38 +240,38 @@ 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, ucmp_);
 
       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(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_);
 
     VersionEdit new_db;
-    if (db_options_.write_dbid_to_manifest) {
-      DBImpl* impl = new DBImpl(DBOptions(), dbname_);
-      std::string db_id;
-      impl->GetDbIdentityFromIdentityFile(&db_id);
-      new_db.SetDBId(db_id);
-    }
     new_db.SetLogNumber(0);
     new_db.SetNextFile(2);
     new_db.SetLastSequence(0);
@@ -284,19 +293,20 @@ class CompactionJobTest : public testing::Test {
     // Make "CURRENT" file that points to the new manifest file.
     s = SetCurrentFile(fs_.get(), dbname_, 1, nullptr);
 
-    std::vector<ColumnFamilyDescriptor> column_families;
-    cf_options_.table_factory = mock_table_factory_;
+    ASSERT_OK(s);
+
     cf_options_.merge_operator = merge_op_;
     cf_options_.compaction_filter = compaction_filter_.get();
+    std::vector<ColumnFamilyDescriptor> column_families;
     column_families.emplace_back(kDefaultColumnFamilyName, cf_options_);
 
-    EXPECT_OK(versions_->Recover(column_families, false));
+    ASSERT_OK(versions_->Recover(column_families, false));
     cfd_ = versions_->GetColumnFamilySet()->GetDefault();
   }
 
   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,
@@ -328,22 +338,27 @@ class CompactionJobTest : public testing::Test {
     EventLogger event_logger(db_options_.info_log.get());
     // TODO(yiwu) add a mock snapshot checker and add test for it.
     SnapshotChecker* snapshot_checker = nullptr;
+    ASSERT_TRUE(full_history_ts_low_.empty() ||
+                ucmp_->timestamp_size() == full_history_ts_low_.size());
     CompactionJob compaction_job(
         0, &compaction, db_options_, env_options_, versions_.get(),
         &shutting_down_, preserve_deletes_seqnum_, &log_buffer, nullptr,
-        nullptr, nullptr, &mutex_, &error_handler_, snapshots,
+        nullptr, 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 */,
+        /*manual_compaction_paused=*/nullptr, /*db_id=*/"",
+        /*db_session_id=*/"", full_history_ts_low_);
     VerifyInitializationOfCompactionJobStats(compaction_job_stats_);
 
     compaction_job.Prepare();
     mutex_.Unlock();
-    Status s;
-    s = compaction_job.Run();
+    Status 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) {
@@ -368,6 +383,7 @@ class CompactionJobTest : public testing::Test {
   Env* env_;
   std::shared_ptr<FileSystem> fs_;
   std::string dbname_;
+  const Comparator* const ucmp_;
   EnvOptions env_options_;
   ImmutableDBOptions db_options_;
   ColumnFamilyOptions cf_options_;
@@ -386,6 +402,17 @@ class CompactionJobTest : public testing::Test {
   std::unique_ptr<CompactionFilter> compaction_filter_;
   std::shared_ptr<MergeOperator> merge_op_;
   ErrorHandler error_handler_;
+  std::string full_history_ts_low_;
+  const std::function<std::string(uint64_t)> encode_u64_ts_;
+};
+
+// TODO(icanadi) Make it simpler once we mock out VersionSet
+class CompactionJobTest : public CompactionJobTestBase {
+ public:
+  CompactionJobTest()
+      : CompactionJobTestBase(test::PerThreadDBPath("compaction_job_test"),
+                              BytewiseComparator(),
+                              [](uint64_t /*ts*/) { return ""; }) {}
 };
 
 TEST_F(CompactionJobTest, Simple) {
@@ -639,7 +666,7 @@ 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);
 }
 
@@ -1066,6 +1093,118 @@ TEST_F(CompactionJobTest, OldestBlobFileNumber) {
                 /* expected_oldest_blob_file_number */ 19);
 }
 
+class CompactionJobTimestampTest : public CompactionJobTestBase {
+ public:
+  CompactionJobTimestampTest()
+      : CompactionJobTestBase(test::PerThreadDBPath("compaction_job_ts_test"),
+                              test::ComparatorWithU64Ts(), test::EncodeInt) {}
+};
+
+TEST_F(CompactionJobTimestampTest, GCDisabled) {
+  NewDB();
+
+  auto file1 =
+      mock::MakeMockFile({{KeyStr("a", 10, ValueType::kTypeValue, 100), "a10"},
+                          {KeyStr("a", 9, ValueType::kTypeValue, 99), "a9"},
+                          {KeyStr("b", 8, ValueType::kTypeValue, 98), "b8"}});
+  AddMockFile(file1);
+
+  auto file2 = mock::MakeMockFile(
+      {{KeyStr("b", 7, ValueType::kTypeDeletionWithTimestamp, 97), ""},
+       {KeyStr("c", 6, ValueType::kTypeDeletionWithTimestamp, 96), ""},
+       {KeyStr("c", 5, ValueType::kTypeValue, 95), "c5"}});
+  AddMockFile(file2);
+
+  SetLastSequence(10);
+
+  auto expected_results = mock::MakeMockFile(
+      {{KeyStr("a", 10, ValueType::kTypeValue, 100), "a10"},
+       {KeyStr("a", 9, ValueType::kTypeValue, 99), "a9"},
+       {KeyStr("b", 8, ValueType::kTypeValue, 98), "b8"},
+       {KeyStr("b", 7, ValueType::kTypeDeletionWithTimestamp, 97), ""},
+       {KeyStr("c", 6, ValueType::kTypeDeletionWithTimestamp, 96), ""},
+       {KeyStr("c", 5, ValueType::kTypeValue, 95), "c5"}});
+  const auto& files = cfd_->current()->storage_info()->LevelFiles(0);
+  RunCompaction({files}, expected_results);
+}
+
+TEST_F(CompactionJobTimestampTest, NoKeyExpired) {
+  NewDB();
+
+  auto file1 =
+      mock::MakeMockFile({{KeyStr("a", 6, ValueType::kTypeValue, 100), "a6"},
+                          {KeyStr("b", 7, ValueType::kTypeValue, 101), "b7"},
+                          {KeyStr("c", 5, ValueType::kTypeValue, 99), "c5"}});
+  AddMockFile(file1);
+
+  auto file2 =
+      mock::MakeMockFile({{KeyStr("a", 4, ValueType::kTypeValue, 98), "a4"},
+                          {KeyStr("c", 3, ValueType::kTypeValue, 97), "c3"}});
+  AddMockFile(file2);
+
+  SetLastSequence(101);
+
+  auto expected_results =
+      mock::MakeMockFile({{KeyStr("a", 6, ValueType::kTypeValue, 100), "a6"},
+                          {KeyStr("a", 4, ValueType::kTypeValue, 98), "a4"},
+                          {KeyStr("b", 7, ValueType::kTypeValue, 101), "b7"},
+                          {KeyStr("c", 5, ValueType::kTypeValue, 99), "c5"},
+                          {KeyStr("c", 3, ValueType::kTypeValue, 97), "c3"}});
+  const auto& files = cfd_->current()->storage_info()->LevelFiles(0);
+
+  full_history_ts_low_ = encode_u64_ts_(0);
+  RunCompaction({files}, expected_results);
+}
+
+TEST_F(CompactionJobTimestampTest, AllKeysExpired) {
+  NewDB();
+
+  auto file1 = mock::MakeMockFile(
+      {{KeyStr("a", 5, ValueType::kTypeDeletionWithTimestamp, 100), ""},
+       {KeyStr("b", 6, ValueType::kTypeValue, 99), "b6"}});
+  AddMockFile(file1);
+
+  auto file2 = mock::MakeMockFile(
+      {{KeyStr("a", 4, ValueType::kTypeValue, 98), "a4"},
+       {KeyStr("b", 3, ValueType::kTypeDeletionWithTimestamp, 97), ""},
+       {KeyStr("b", 2, ValueType::kTypeValue, 96), "b2"}});
+  AddMockFile(file2);
+
+  SetLastSequence(6);
+
+  auto expected_results =
+      mock::MakeMockFile({{KeyStr("b", 0, ValueType::kTypeValue, 0), "b6"}});
+  const auto& files = cfd_->current()->storage_info()->LevelFiles(0);
+
+  full_history_ts_low_ = encode_u64_ts_(std::numeric_limits<uint64_t>::max());
+  RunCompaction({files}, expected_results);
+}
+
+TEST_F(CompactionJobTimestampTest, SomeKeysExpired) {
+  NewDB();
+
+  auto file1 =
+      mock::MakeMockFile({{KeyStr("a", 5, ValueType::kTypeValue, 50), "a5"},
+                          {KeyStr("b", 6, ValueType::kTypeValue, 49), "b6"}});
+  AddMockFile(file1);
+
+  auto file2 = mock::MakeMockFile(
+      {{KeyStr("a", 3, ValueType::kTypeValue, 48), "a3"},
+       {KeyStr("a", 2, ValueType::kTypeValue, 46), "a2"},
+       {KeyStr("b", 4, ValueType::kTypeDeletionWithTimestamp, 47), ""}});
+  AddMockFile(file2);
+
+  SetLastSequence(6);
+
+  auto expected_results =
+      mock::MakeMockFile({{KeyStr("a", 5, ValueType::kTypeValue, 50), "a5"},
+                          {KeyStr("b", 6, ValueType::kTypeValue, 49), "b6"}});
+  const auto& files = cfd_->current()->storage_info()->LevelFiles(0);
+
+  full_history_ts_low_ = encode_u64_ts_(49);
+  RunCompaction({files}, expected_results);
+}
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/compaction/compaction_picker.cc

@@ -139,11 +139,9 @@ CompressionOptions GetCompressionOptions(const MutableCFOptions& cf_options,
   if (!enable_compression) {
     return cf_options.compression_opts;
   }
-  // If bottommost_compression is set and we are compacting to the
-  // bottommost level then we should use the specified compression options
-  // for the bottmomost_compression.
-  if (cf_options.bottommost_compression != kDisableCompressionOption &&
-      level >= (vstorage->num_non_empty_levels() - 1) &&
+  // If bottommost_compression_opts is enabled and we are compacting to the
+  // bottommost level then we should use the specified compression options.
+  if (level >= (vstorage->num_non_empty_levels() - 1) &&
       cf_options.bottommost_compression_opts.enabled) {
     return cf_options.bottommost_compression_opts;
   }
@@ -639,6 +637,7 @@ Compaction* CompactionPicker::CompactRange(
         compact_range_options.max_subcompactions, /* grandparents */ {},
         /* is manual */ true);
     RegisterCompaction(c);
+    vstorage->ComputeCompactionScore(ioptions_, mutable_cf_options);
     return c;
   }
 
@@ -1044,6 +1043,8 @@ void CompactionPicker::RegisterCompaction(Compaction* c) {
     level0_compactions_in_progress_.insert(c);
   }
   compactions_in_progress_.insert(c);
+  TEST_SYNC_POINT_CALLBACK("CompactionPicker::RegisterCompaction:Registered",
+                           c);
 }
 
 void CompactionPicker::UnregisterCompaction(Compaction* c) {

db/compaction/compaction_picker_test.cc

@@ -12,7 +12,6 @@
 #include "db/compaction/compaction_picker_level.h"
 #include "db/compaction/compaction_picker_universal.h"
 
-#include "logging/logging.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"
 #include "util/string_util.h"
@@ -2016,6 +2015,171 @@ TEST_F(CompactionPickerTest, UniversalMarkedCompactionStartOutputOverlap) {
     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

@@ -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
@@ -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;
 
@@ -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];
@@ -1019,9 +1025,9 @@ Compaction* UniversalCompactionBuilder::PickCompactionToOldest(
       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/comparator_db_test.cc

@@ -17,7 +17,6 @@
 #include "util/string_util.h"
 #include "utilities/merge_operators.h"
 
-using std::unique_ptr;
 
 namespace ROCKSDB_NAMESPACE {
 namespace {

db/corruption_test.cc

@@ -9,7 +9,6 @@
 
 #ifndef ROCKSDB_LITE
 
-#include <errno.h>
 #include <fcntl.h>
 #include <sys/stat.h>
 #include <sys/types.h>
@@ -33,12 +32,13 @@
 #include "table/mock_table.h"
 #include "test_util/testharness.h"
 #include "test_util/testutil.h"
+#include "util/cast_util.h"
 #include "util/random.h"
 #include "util/string_util.h"
 
 namespace ROCKSDB_NAMESPACE {
 
-static const int kValueSize = 1000;
+static constexpr int kValueSize = 1000;
 
 class CorruptionTest : public testing::Test {
  public:
@@ -56,7 +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;
@@ -68,8 +69,16 @@ class CorruptionTest : public testing::Test {
   }
 
   ~CorruptionTest() override {
+    SyncPoint::GetInstance()->DisableProcessing();
+    SyncPoint::GetInstance()->LoadDependency({});
+    SyncPoint::GetInstance()->ClearAllCallBacks();
     delete db_;
-    DestroyDB(dbname_, Options());
+    db_ = nullptr;
+    if (getenv("KEEP_DB")) {
+      fprintf(stdout, "db is still at %s\n", dbname_.c_str());
+    } else {
+      EXPECT_OK(DestroyDB(dbname_, Options()));
+    }
   }
 
   void CloseDb() {
@@ -104,22 +113,24 @@ 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 = static_cast_with_check<DBImpl>(db_);
-        dbi->TEST_FlushMemTable();
+        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;
@@ -134,6 +145,7 @@ 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) ||
@@ -150,6 +162,7 @@ class CorruptionTest : public testing::Test {
         correct++;
       }
     }
+    iter->status().PermitUncheckedError();
     delete iter;
 
     fprintf(stderr,
@@ -178,7 +191,7 @@ class CorruptionTest : public testing::Test {
     }
     ASSERT_TRUE(!fname.empty()) << filetype;
 
-    test::CorruptFile(fname, offset, bytes_to_corrupt);
+    ASSERT_OK(test::CorruptFile(&env_, fname, offset, bytes_to_corrupt));
   }
 
   // corrupts exactly one file at level `level`. if no file found at level,
@@ -188,7 +201,8 @@ class CorruptionTest : public testing::Test {
     db_->GetLiveFilesMetaData(&metadata);
     for (const auto& m : metadata) {
       if (m.level == level) {
-        test::CorruptFile(dbname_ + "/" + m.name, offset, bytes_to_corrupt);
+        ASSERT_OK(test::CorruptFile(&env_, dbname_ + "/" + m.name, offset,
+                                    bytes_to_corrupt));
         return;
       }
     }
@@ -243,8 +257,8 @@ TEST_F(CorruptionTest, Recovery) {
   // is not available for WAL though.
   CloseDb();
 #endif
-  Corrupt(kLogFile, 19, 1);      // WriteBatch tag for first record
-  Corrupt(kLogFile, log::kBlockSize + 1000, 1);  // Somewhere in second block
+  Corrupt(kWalFile, 19, 1);  // WriteBatch tag for first record
+  Corrupt(kWalFile, log::kBlockSize + 1000, 1);  // Somewhere in second block
   ASSERT_TRUE(!TryReopen().ok());
   options_.paranoid_checks = false;
   Reopen(&options_);
@@ -269,7 +283,7 @@ 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;
@@ -285,9 +299,9 @@ TEST_F(CorruptionTest, NewFileErrorDuringWrite) {
 TEST_F(CorruptionTest, TableFile) {
   Build(100);
   DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
-  dbi->TEST_CompactRange(1, nullptr, nullptr);
+  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);
@@ -308,9 +322,9 @@ TEST_F(CorruptionTest, VerifyChecksumReadahead) {
 
   Build(10000);
   DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
-  dbi->TEST_CompactRange(1, nullptr, nullptr);
+  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();
@@ -355,7 +369,7 @@ TEST_F(CorruptionTest, TableFileIndexData) {
   // build 2 tables, flush at 5000
   Build(10000, 5000);
   DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
-  dbi->TEST_FlushMemTable();
+  ASSERT_OK(dbi->TEST_FlushMemTable());
 
   // corrupt an index block of an entire file
   Corrupt(kTableFile, -2000, 500);
@@ -402,8 +416,8 @@ TEST_F(CorruptionTest, SequenceNumberRecovery) {
 TEST_F(CorruptionTest, CorruptedDescriptor) {
   ASSERT_OK(db_->Put(WriteOptions(), "foo", "hello"));
   DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
+  ASSERT_OK(dbi->TEST_FlushMemTable());
+  ASSERT_OK(dbi->TEST_CompactRange(0, nullptr, nullptr));
 
   Corrupt(kDescriptorFile, 0, 1000);
   Status s = TryReopen();
@@ -421,9 +435,9 @@ TEST_F(CorruptionTest, CompactionInputError) {
   Reopen(&options);
   Build(10);
   DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
-  dbi->TEST_FlushMemTable();
-  dbi->TEST_CompactRange(0, nullptr, nullptr);
-  dbi->TEST_CompactRange(1, nullptr, nullptr);
+  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);
@@ -446,12 +460,12 @@ TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
 
   // 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));
     }
   }
 
@@ -459,8 +473,8 @@ TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
 
   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);
@@ -485,7 +499,7 @@ TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
 TEST_F(CorruptionTest, UnrelatedKeys) {
   Build(10);
   DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
-  dbi->TEST_FlushMemTable();
+  ASSERT_OK(dbi->TEST_FlushMemTable());
   Corrupt(kTableFile, 100, 1);
   ASSERT_NOK(dbi->VerifyChecksum());
 
@@ -494,7 +508,7 @@ 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);
 }
@@ -523,7 +537,8 @@ TEST_F(CorruptionTest, RangeDeletionCorrupted) {
       ImmutableCFOptions(options_), kRangeDelBlock, &range_del_handle));
 
   ASSERT_OK(TryReopen());
-  test::CorruptFile(filename, static_cast<int>(range_del_handle.offset()), 1);
+  ASSERT_OK(test::CorruptFile(&env_, filename,
+                              static_cast<int>(range_del_handle.offset()), 1));
   ASSERT_TRUE(TryReopen().IsCorruption());
 }
 
@@ -538,7 +553,7 @@ TEST_F(CorruptionTest, FileSystemStateCorrupted) {
     DBImpl* dbi = static_cast_with_check<DBImpl>(db_);
     std::vector<LiveFileMetaData> metadata;
     dbi->GetLiveFilesMetaData(&metadata);
-    ASSERT_GT(metadata.size(), size_t(0));
+    ASSERT_GT(metadata.size(), 0);
     std::string filename = dbname_ + metadata[0].name;
 
     delete db_;
@@ -547,37 +562,42 @@ 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());
+      ASSERT_TRUE(x.IsCorruption());
     }
 
-    DestroyDB(dbname_, options_);
+    ASSERT_OK(DestroyDB(dbname_, options_));
   }
 }
 
-static const auto& corruption_modes = {mock::MockTableFactory::kCorruptNone,
-                                       mock::MockTableFactory::kCorruptKey,
-                                       mock::MockTableFactory::kCorruptValue};
+static const auto& corruption_modes = {
+    mock::MockTableFactory::kCorruptNone, mock::MockTableFactory::kCorruptKey,
+    mock::MockTableFactory::kCorruptValue,
+    mock::MockTableFactory::kCorruptReorderKey};
 
-TEST_F(CorruptionTest, ParaniodFileChecksOnFlush) {
+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) {
@@ -588,18 +608,21 @@ TEST_F(CorruptionTest, ParaniodFileChecksOnFlush) {
   }
 }
 
-TEST_F(CorruptionTest, ParaniodFileChecksOnCompact) {
+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_);
@@ -614,6 +637,237 @@ TEST_F(CorruptionTest, ParaniodFileChecksOnCompact) {
   }
 }
 
+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();
+}
+
+TEST_F(CorruptionTest, VerifyWholeTableChecksum) {
+  CloseDb();
+  Options options;
+  options.env = &env_;
+  ASSERT_OK(DestroyDB(dbname_, options));
+  options.create_if_missing = true;
+  options.file_checksum_gen_factory =
+      ROCKSDB_NAMESPACE::GetFileChecksumGenCrc32cFactory();
+  Reopen(&options);
+
+  Build(10, 5);
+
+  ASSERT_OK(db_->VerifyFileChecksums(ReadOptions()));
+  CloseDb();
+
+  // Corrupt the first byte of each table file, this must be data block.
+  Corrupt(kTableFile, 0, 1);
+
+  ASSERT_OK(TryReopen(&options));
+
+  SyncPoint::GetInstance()->DisableProcessing();
+  SyncPoint::GetInstance()->ClearAllCallBacks();
+  int count{0};
+  SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::VerifySstFileChecksum:mismatch", [&](void* arg) {
+        auto* s = reinterpret_cast<Status*>(arg);
+        assert(s);
+        ++count;
+        ASSERT_NOK(*s);
+      });
+  SyncPoint::GetInstance()->EnableProcessing();
+  ASSERT_TRUE(db_->VerifyFileChecksums(ReadOptions()).IsCorruption());
+  ASSERT_EQ(1, count);
+}
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/cuckoo_table_db_test.cc

@@ -63,6 +63,15 @@ class CuckooTableDBTest : public testing::Test {
     ASSERT_OK(DB::Open(opts, dbname_, &db_));
   }
 
+  void DestroyAndReopen(Options* options) {
+    assert(options);
+    ASSERT_OK(db_->Close());
+    delete db_;
+    db_ = nullptr;
+    ASSERT_OK(DestroyDB(dbname_, *options));
+    Reopen(options);
+  }
+
   Status Put(const Slice& k, const Slice& v) {
     return db_->Put(WriteOptions(), k, v);
   }
@@ -205,7 +214,7 @@ static std::string Uint64Key(uint64_t i) {
 TEST_F(CuckooTableDBTest, Uint64Comparator) {
   Options options = CurrentOptions();
   options.comparator = test::Uint64Comparator();
-  Reopen(&options);
+  DestroyAndReopen(&options);
 
   ASSERT_OK(Put(Uint64Key(1), "v1"));
   ASSERT_OK(Put(Uint64Key(2), "v2"));

db/db_block_cache_test.cc

@@ -34,7 +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;
@@ -49,7 +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;
   }
 
@@ -157,7 +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);
 
@@ -181,7 +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);
 
@@ -237,7 +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);
 
@@ -298,7 +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"));
@@ -354,7 +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"));
@@ -392,7 +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"));
@@ -473,7 +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;
@@ -572,7 +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.
@@ -661,7 +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"));
@@ -836,8 +837,9 @@ TEST_F(DBBlockCacheTest, CacheCompressionDict) {
   Random rnd(301);
   for (auto compression_type : compression_types) {
     Options options = CurrentOptions();
-    options.compression = compression_type;
-    options.compression_opts.max_dict_bytes = 4096;
+    options.bottommost_compression = compression_type;
+    options.bottommost_compression_opts.max_dict_bytes = 4096;
+    options.bottommost_compression_opts.enabled = true;
     options.create_if_missing = true;
     options.num_levels = 2;
     options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
@@ -845,7 +847,7 @@ 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);
@@ -887,6 +889,166 @@ 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;
+    }
+  }
+  if (unpartitioned_pinning_ == PinningTier::kNone) {
+    ++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));
+
+  // 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);
@@ -509,24 +514,24 @@ INSTANTIATE_TEST_CASE_P(
     ::testing::Values(
         std::make_tuple(BFP::kDeprecatedBlock, false,
                         test::kDefaultFormatVersion),
-        std::make_tuple(BFP::kAuto, true, test::kDefaultFormatVersion),
-        std::make_tuple(BFP::kAuto, false, test::kDefaultFormatVersion)));
+        std::make_tuple(BFP::kAutoBloom, true, test::kDefaultFormatVersion),
+        std::make_tuple(BFP::kAutoBloom, false, test::kDefaultFormatVersion)));
 
 INSTANTIATE_TEST_CASE_P(
     FormatDef, DBBloomFilterTestWithParam,
     ::testing::Values(
         std::make_tuple(BFP::kDeprecatedBlock, false,
                         test::kDefaultFormatVersion),
-        std::make_tuple(BFP::kAuto, true, test::kDefaultFormatVersion),
-        std::make_tuple(BFP::kAuto, false, test::kDefaultFormatVersion)));
+        std::make_tuple(BFP::kAutoBloom, true, test::kDefaultFormatVersion),
+        std::make_tuple(BFP::kAutoBloom, false, test::kDefaultFormatVersion)));
 
 INSTANTIATE_TEST_CASE_P(
     FormatLatest, DBBloomFilterTestWithParam,
     ::testing::Values(
         std::make_tuple(BFP::kDeprecatedBlock, false,
                         test::kLatestFormatVersion),
-        std::make_tuple(BFP::kAuto, true, test::kLatestFormatVersion),
-        std::make_tuple(BFP::kAuto, false, test::kLatestFormatVersion)));
+        std::make_tuple(BFP::kAutoBloom, true, test::kLatestFormatVersion),
+        std::make_tuple(BFP::kAutoBloom, false, test::kLatestFormatVersion)));
 #endif  // ROCKSDB_VALGRIND_RUN
 
 TEST_F(DBBloomFilterTest, BloomFilterRate) {
@@ -1039,7 +1044,7 @@ class DBBloomFilterTestVaryPrefixAndFormatVer
 
  public:
   DBBloomFilterTestVaryPrefixAndFormatVer()
-      : DBTestBase("/db_bloom_filter_tests") {}
+      : DBTestBase("/db_bloom_filter_tests", /*env_do_fsync=*/true) {}
 
   ~DBBloomFilterTestVaryPrefixAndFormatVer() override {}
 
@@ -1725,6 +1730,7 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
     int using_full_builder = bfp_impl != BFP::kDeprecatedBlock;
     Options options;
     options.create_if_missing = true;
+    options.env = CurrentOptions().env;
     options.prefix_extractor.reset(NewCappedPrefixTransform(4));
     options.disable_auto_compactions = true;
     options.statistics = CreateDBStatistics();
@@ -1855,6 +1861,7 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterMultipleSST) {
   for (auto bfp_impl : BFP::kAllFixedImpls) {
     int using_full_builder = bfp_impl != BFP::kDeprecatedBlock;
     Options options;
+    options.env = CurrentOptions().env;
     options.create_if_missing = true;
     options.prefix_extractor.reset(NewFixedPrefixTransform(1));
     options.disable_auto_compactions = true;
@@ -2047,6 +2054,7 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterNewColumnFamily) {
 TEST_F(DBBloomFilterTest, DynamicBloomFilterOptions) {
   for (auto bfp_impl : BFP::kAllFixedImpls) {
     Options options;
+    options.env = CurrentOptions().env;
     options.create_if_missing = true;
     options.prefix_extractor.reset(NewFixedPrefixTransform(1));
     options.disable_auto_compactions = true;

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, true /* log_err_key */));
       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, true /* log_err_key */));
       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, true /* log_err_key */));
       total++;
       if (ikey.sequence != 0) {
         count++;

db/db_compaction_test.cc

@@ -26,14 +26,16 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBCompactionTest : public DBTestBase {
  public:
-  DBCompactionTest() : DBTestBase("/db_compaction_test") {}
+  DBCompactionTest()
+      : DBTestBase("/db_compaction_test", /*env_do_fsync=*/true) {}
 };
 
 class DBCompactionTestWithParam
     : public DBTestBase,
       public testing::WithParamInterface<std::tuple<uint32_t, bool>> {
  public:
-  DBCompactionTestWithParam() : DBTestBase("/db_compaction_test") {
+  DBCompactionTestWithParam()
+      : DBTestBase("/db_compaction_test", /*env_do_fsync=*/true) {
     max_subcompactions_ = std::get<0>(GetParam());
     exclusive_manual_compaction_ = std::get<1>(GetParam());
   }
@@ -46,12 +48,34 @@ class DBCompactionTestWithParam
   bool exclusive_manual_compaction_;
 };
 
+class DBCompactionTestWithBottommostParam
+    : public DBTestBase,
+      public testing::WithParamInterface<BottommostLevelCompaction> {
+ public:
+  DBCompactionTestWithBottommostParam()
+      : DBTestBase("/db_compaction_test", /*env_do_fsync=*/true) {
+    bottommost_level_compaction_ = GetParam();
+  }
+
+  BottommostLevelCompaction bottommost_level_compaction_;
+};
+
 class DBCompactionDirectIOTest : public DBCompactionTest,
                                  public ::testing::WithParamInterface<bool> {
  public:
   DBCompactionDirectIOTest() : DBCompactionTest() {}
 };
 
+// Param = true : target level is non-empty
+// Param = false: level between target level and source level
+//  is not empty.
+class ChangeLevelConflictsWithAuto
+    : public DBCompactionTest,
+      public ::testing::WithParamInterface<bool> {
+ public:
+  ChangeLevelConflictsWithAuto() : DBCompactionTest() {}
+};
+
 namespace {
 
 class FlushedFileCollector : public EventListener {
@@ -3246,7 +3270,7 @@ TEST_P(DBCompactionTestWithParam, IntraL0Compaction) {
   table_options.block_cache = NewLRUCache(64 << 20);  // 64MB
   table_options.cache_index_and_filter_blocks = true;
   table_options.pin_l0_filter_and_index_blocks_in_cache = true;
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
 
   DestroyAndReopen(options);
 
@@ -3254,8 +3278,13 @@ TEST_P(DBCompactionTestWithParam, IntraL0Compaction) {
   Random rnd(301);
   std::string value(rnd.RandomString(kValueSize));
 
+  // The L0->L1 must be picked before we begin flushing files to trigger
+  // intra-L0 compaction, and must not finish until after an intra-L0
+  // compaction has been picked.
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency(
-      {{"LevelCompactionPicker::PickCompactionBySize:0",
+      {{"LevelCompactionPicker::PickCompaction:Return",
+        "DBCompactionTest::IntraL0Compaction:L0ToL1Ready"},
+       {"LevelCompactionPicker::PickCompactionBySize:0",
         "CompactionJob::Run():Start"}});
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
 
@@ -3273,6 +3302,7 @@ TEST_P(DBCompactionTestWithParam, IntraL0Compaction) {
   for (int i = 0; i < 10; ++i) {
     ASSERT_OK(Put(Key(0), ""));  // prevents trivial move
     if (i == 5) {
+      TEST_SYNC_POINT("DBCompactionTest::IntraL0Compaction:L0ToL1Ready");
       ASSERT_OK(Put(Key(i + 1), value + value));
     } else {
       ASSERT_OK(Put(Key(i + 1), value));
@@ -3318,8 +3348,14 @@ TEST_P(DBCompactionTestWithParam, IntraL0CompactionDoesNotObsoleteDeletions) {
   Random rnd(301);
   std::string value(rnd.RandomString(kValueSize));
 
+  // The L0->L1 must be picked before we begin flushing files to trigger
+  // intra-L0 compaction, and must not finish until after an intra-L0
+  // compaction has been picked.
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency(
-      {{"LevelCompactionPicker::PickCompactionBySize:0",
+      {{"LevelCompactionPicker::PickCompaction:Return",
+        "DBCompactionTest::IntraL0CompactionDoesNotObsoleteDeletions:"
+        "L0ToL1Ready"},
+       {"LevelCompactionPicker::PickCompactionBySize:0",
         "CompactionJob::Run():Start"}});
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
 
@@ -3343,6 +3379,11 @@ TEST_P(DBCompactionTestWithParam, IntraL0CompactionDoesNotObsoleteDeletions) {
     } else {
       ASSERT_OK(Delete(Key(0)));
     }
+    if (i == 5) {
+      TEST_SYNC_POINT(
+          "DBCompactionTest::IntraL0CompactionDoesNotObsoleteDeletions:"
+          "L0ToL1Ready");
+    }
     ASSERT_OK(Put(Key(i + 1), value));
     ASSERT_OK(Flush());
   }
@@ -3588,6 +3629,76 @@ TEST_F(DBCompactionTest, CompactBottomLevelFilesWithDeletions) {
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
 }
 
+TEST_F(DBCompactionTest, NoCompactBottomLevelFilesWithDeletions) {
+  // bottom-level files may contain deletions due to snapshots protecting the
+  // deleted keys. Once the snapshot is released, we should see files with many
+  // such deletions undergo single-file compactions. But when disabling auto
+  // compactions, it shouldn't be triggered which may causing too many
+  // background jobs.
+  const int kNumKeysPerFile = 1024;
+  const int kNumLevelFiles = 4;
+  const int kValueSize = 128;
+  Options options = CurrentOptions();
+  options.compression = kNoCompression;
+  options.disable_auto_compactions = true;
+  options.level0_file_num_compaction_trigger = kNumLevelFiles;
+  // inflate it a bit to account for key/metadata overhead
+  options.target_file_size_base = 120 * kNumKeysPerFile * kValueSize / 100;
+  Reopen(options);
+
+  Random rnd(301);
+  const Snapshot* snapshot = nullptr;
+  for (int i = 0; i < kNumLevelFiles; ++i) {
+    for (int j = 0; j < kNumKeysPerFile; ++j) {
+      ASSERT_OK(
+          Put(Key(i * kNumKeysPerFile + j), rnd.RandomString(kValueSize)));
+    }
+    if (i == kNumLevelFiles - 1) {
+      snapshot = db_->GetSnapshot();
+      // delete every other key after grabbing a snapshot, so these deletions
+      // and the keys they cover can't be dropped until after the snapshot is
+      // released.
+      for (int j = 0; j < kNumLevelFiles * kNumKeysPerFile; j += 2) {
+        ASSERT_OK(Delete(Key(j)));
+      }
+    }
+    Flush();
+    if (i < kNumLevelFiles - 1) {
+      ASSERT_EQ(i + 1, NumTableFilesAtLevel(0));
+    }
+  }
+  dbfull()->TEST_CompactRange(0, nullptr, nullptr, nullptr);
+  ASSERT_EQ(kNumLevelFiles, NumTableFilesAtLevel(1));
+
+  std::vector<LiveFileMetaData> pre_release_metadata, post_release_metadata;
+  db_->GetLiveFilesMetaData(&pre_release_metadata);
+  // just need to bump seqnum so ReleaseSnapshot knows the newest key in the SST
+  // files does not need to be preserved in case of a future snapshot.
+  ASSERT_OK(Put(Key(0), "val"));
+
+  // release snapshot and no compaction should be triggered.
+  std::atomic<int> num_compactions{0};
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
+      "DBImpl::BackgroundCompaction:Start",
+      [&](void* /*arg*/) { num_compactions.fetch_add(1); });
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
+  db_->ReleaseSnapshot(snapshot);
+  dbfull()->TEST_WaitForCompact();
+  ASSERT_EQ(0, num_compactions);
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
+
+  db_->GetLiveFilesMetaData(&post_release_metadata);
+  ASSERT_EQ(pre_release_metadata.size(), post_release_metadata.size());
+  for (size_t i = 0; i < pre_release_metadata.size(); ++i) {
+    const auto& pre_file = pre_release_metadata[i];
+    const auto& post_file = post_release_metadata[i];
+    ASSERT_EQ(1, pre_file.level);
+    ASSERT_EQ(1, post_file.level);
+    // each file is same as before with deletion markers/deleted keys.
+    ASSERT_EQ(post_file.size, pre_file.size);
+  }
+}
+
 TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
   const int kNumKeysPerFile = 32;
   const int kNumLevelFiles = 2;
@@ -3597,10 +3708,11 @@ TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
   options.compression = kNoCompression;
   options.ttl = 24 * 60 * 60;  // 24 hours
   options.max_open_files = -1;
-  env_->time_elapse_only_sleep_ = false;
+  env_->SetMockSleep();
   options.env = env_;
 
-  env_->addon_time_.store(0);
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
   DestroyAndReopen(options);
 
   Random rnd(301);
@@ -3627,7 +3739,7 @@ TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
   MoveFilesToLevel(1);
   ASSERT_EQ("0,2,0,2", FilesPerLevel());
 
-  env_->addon_time_.fetch_add(36 * 60 * 60);  // 36 hours
+  env_->MockSleepForSeconds(36 * 60 * 60);  // 36 hours
   ASSERT_EQ("0,2,0,2", FilesPerLevel());
 
   // Just do a simple write + flush so that the Ttl expired files get
@@ -3647,7 +3759,8 @@ TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
 
   // Test dynamically changing ttl.
 
-  env_->addon_time_.store(0);
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
   DestroyAndReopen(options);
 
   for (int i = 0; i < kNumLevelFiles; ++i) {
@@ -3675,7 +3788,7 @@ TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
 
   // Move time forward by 12 hours, and make sure that compaction still doesn't
   // trigger as ttl is set to 24 hours.
-  env_->addon_time_.fetch_add(12 * 60 * 60);
+  env_->MockSleepForSeconds(12 * 60 * 60);
   ASSERT_OK(Put("a", "1"));
   Flush();
   dbfull()->TEST_WaitForCompact();
@@ -3698,6 +3811,7 @@ TEST_F(DBCompactionTest, LevelCompactExpiredTtlFiles) {
 }
 
 TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
+  env_->SetMockSleep();
   const int kValueSize = 100;
 
   for (bool if_restart : {false, true}) {
@@ -3728,10 +3842,10 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
             }
           });
 
-      env_->time_elapse_only_sleep_ = false;
       options.env = env_;
 
-      env_->addon_time_.store(0);
+      // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
       DestroyAndReopen(options);
 
       int ttl_compactions = 0;
@@ -3759,7 +3873,7 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
                                       &level_to_files);
       uint64_t oldest_time = level_to_files[0][0].oldest_ancester_time;
       // Add 1 hour and do another flush.
-      env_->addon_time_.fetch_add(1 * 60 * 60);
+      env_->MockSleepForSeconds(1 * 60 * 60);
       for (int i = 101; i <= 200; ++i) {
         ASSERT_OK(Put(Key(i), rnd.RandomString(kValueSize)));
       }
@@ -3767,13 +3881,13 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
       MoveFilesToLevel(6);
       ASSERT_EQ("0,0,0,0,0,0,2", FilesPerLevel());
 
-      env_->addon_time_.fetch_add(1 * 60 * 60);
+      env_->MockSleepForSeconds(1 * 60 * 60);
       // Add two L4 files with key ranges: [1 .. 50], [51 .. 150].
       for (int i = 1; i <= 50; ++i) {
         ASSERT_OK(Put(Key(i), rnd.RandomString(kValueSize)));
       }
       Flush();
-      env_->addon_time_.fetch_add(1 * 60 * 60);
+      env_->MockSleepForSeconds(1 * 60 * 60);
       for (int i = 51; i <= 150; ++i) {
         ASSERT_OK(Put(Key(i), rnd.RandomString(kValueSize)));
       }
@@ -3781,7 +3895,7 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
       MoveFilesToLevel(4);
       ASSERT_EQ("0,0,0,0,2,0,2", FilesPerLevel());
 
-      env_->addon_time_.fetch_add(1 * 60 * 60);
+      env_->MockSleepForSeconds(1 * 60 * 60);
       // Add one L1 file with key range: [26, 75].
       for (int i = 26; i <= 75; ++i) {
         ASSERT_OK(Put(Key(i), rnd.RandomString(kValueSize)));
@@ -3811,7 +3925,7 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
       // 4. A TTL compaction happens between L5 and L6 files. Ouptut in L6.
 
       // Add 25 hours and do a write
-      env_->addon_time_.fetch_add(25 * 60 * 60);
+      env_->MockSleepForSeconds(25 * 60 * 60);
 
       ASSERT_OK(Put(Key(1), "1"));
       if (if_restart) {
@@ -3827,7 +3941,7 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
                                       &level_to_files);
       ASSERT_EQ(oldest_time, level_to_files[6][0].oldest_ancester_time);
 
-      env_->addon_time_.fetch_add(25 * 60 * 60);
+      env_->MockSleepForSeconds(25 * 60 * 60);
       ASSERT_OK(Put(Key(2), "1"));
       if (if_restart) {
         Reopen(options);
@@ -3844,6 +3958,7 @@ TEST_F(DBCompactionTest, LevelTtlCascadingCompactions) {
 }
 
 TEST_F(DBCompactionTest, LevelPeriodicCompaction) {
+  env_->SetMockSleep();
   const int kNumKeysPerFile = 32;
   const int kNumLevelFiles = 2;
   const int kValueSize = 100;
@@ -3875,10 +3990,10 @@ TEST_F(DBCompactionTest, LevelPeriodicCompaction) {
             }
           });
 
-      env_->time_elapse_only_sleep_ = false;
       options.env = env_;
 
-      env_->addon_time_.store(0);
+      // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
       DestroyAndReopen(options);
 
       int periodic_compactions = 0;
@@ -3906,7 +4021,7 @@ TEST_F(DBCompactionTest, LevelPeriodicCompaction) {
       ASSERT_EQ(0, periodic_compactions);
 
       // Add 50 hours and do a write
-      env_->addon_time_.fetch_add(50 * 60 * 60);
+      env_->MockSleepForSeconds(50 * 60 * 60);
       ASSERT_OK(Put("a", "1"));
       Flush();
       dbfull()->TEST_WaitForCompact();
@@ -3919,7 +4034,7 @@ TEST_F(DBCompactionTest, LevelPeriodicCompaction) {
       ASSERT_EQ("0,3", FilesPerLevel());
 
       // Add another 50 hours and do another write
-      env_->addon_time_.fetch_add(50 * 60 * 60);
+      env_->MockSleepForSeconds(50 * 60 * 60);
       ASSERT_OK(Put("b", "2"));
       if (if_restart) {
         Reopen(options);
@@ -3933,7 +4048,7 @@ TEST_F(DBCompactionTest, LevelPeriodicCompaction) {
       ASSERT_EQ(5, periodic_compactions);
 
       // Add another 50 hours and do another write
-      env_->addon_time_.fetch_add(50 * 60 * 60);
+      env_->MockSleepForSeconds(50 * 60 * 60);
       ASSERT_OK(Put("c", "3"));
       Flush();
       dbfull()->TEST_WaitForCompact();
@@ -3957,10 +4072,11 @@ TEST_F(DBCompactionTest, LevelPeriodicCompactionWithOldDB) {
   const int kValueSize = 100;
 
   Options options = CurrentOptions();
-  env_->time_elapse_only_sleep_ = false;
+  env_->SetMockSleep();
   options.env = env_;
 
-  env_->addon_time_.store(0);
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
   DestroyAndReopen(options);
 
   int periodic_compactions = 0;
@@ -4008,7 +4124,7 @@ TEST_F(DBCompactionTest, LevelPeriodicCompactionWithOldDB) {
 
   set_file_creation_time_to_zero = false;
   // Forward the clock by 2 days.
-  env_->addon_time_.fetch_add(2 * 24 * 60 * 60);
+  env_->MockSleepForSeconds(2 * 24 * 60 * 60);
   options.periodic_compaction_seconds = 1 * 24 * 60 * 60;  // 1 day
 
   Reopen(options);
@@ -4029,10 +4145,11 @@ TEST_F(DBCompactionTest, LevelPeriodicAndTtlCompaction) {
   options.ttl = 10 * 60 * 60;  // 10 hours
   options.periodic_compaction_seconds = 48 * 60 * 60;  // 2 days
   options.max_open_files = -1;   // needed for both periodic and ttl compactions
-  env_->time_elapse_only_sleep_ = false;
+  env_->SetMockSleep();
   options.env = env_;
 
-  env_->addon_time_.store(0);
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
   DestroyAndReopen(options);
 
   int periodic_compactions = 0;
@@ -4066,7 +4183,7 @@ TEST_F(DBCompactionTest, LevelPeriodicAndTtlCompaction) {
   ASSERT_EQ(0, ttl_compactions);
 
   // Add some time greater than periodic_compaction_time.
-  env_->addon_time_.fetch_add(50 * 60 * 60);
+  env_->MockSleepForSeconds(50 * 60 * 60);
   ASSERT_OK(Put("a", "1"));
   Flush();
   dbfull()->TEST_WaitForCompact();
@@ -4076,7 +4193,7 @@ TEST_F(DBCompactionTest, LevelPeriodicAndTtlCompaction) {
   ASSERT_EQ(0, ttl_compactions);
 
   // Add a little more time than ttl
-  env_->addon_time_.fetch_add(11 * 60 * 60);
+  env_->MockSleepForSeconds(11 * 60 * 60);
   ASSERT_OK(Put("b", "1"));
   Flush();
   dbfull()->TEST_WaitForCompact();
@@ -4088,7 +4205,7 @@ TEST_F(DBCompactionTest, LevelPeriodicAndTtlCompaction) {
   ASSERT_EQ(3, ttl_compactions);
 
   // Add some time greater than periodic_compaction_time.
-  env_->addon_time_.fetch_add(50 * 60 * 60);
+  env_->MockSleepForSeconds(50 * 60 * 60);
   ASSERT_OK(Put("c", "1"));
   Flush();
   dbfull()->TEST_WaitForCompact();
@@ -4121,9 +4238,10 @@ TEST_F(DBCompactionTest, LevelPeriodicCompactionWithCompactionFilters) {
 
   Options options = CurrentOptions();
   TestCompactionFilter test_compaction_filter;
-  env_->time_elapse_only_sleep_ = false;
+  env_->SetMockSleep();
   options.env = env_;
-  env_->addon_time_.store(0);
+
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
 
   enum CompactionFilterType {
     kUseCompactionFilter,
@@ -4174,7 +4292,7 @@ TEST_F(DBCompactionTest, LevelPeriodicCompactionWithCompactionFilters) {
     ASSERT_EQ(0, periodic_compactions);
 
     // Add 31 days and do a write
-    env_->addon_time_.fetch_add(31 * 24 * 60 * 60);
+    env_->MockSleepForSeconds(31 * 24 * 60 * 60);
     ASSERT_OK(Put("a", "1"));
     Flush();
     dbfull()->TEST_WaitForCompact();
@@ -4808,7 +4926,8 @@ INSTANTIATE_TEST_CASE_P(DBCompactionDirectIOTest, DBCompactionDirectIOTest,
 class CompactionPriTest : public DBTestBase,
                           public testing::WithParamInterface<uint32_t> {
  public:
-  CompactionPriTest() : DBTestBase("/compaction_pri_test") {
+  CompactionPriTest()
+      : DBTestBase("/compaction_pri_test", /*env_do_fsync=*/true) {
     compaction_pri_ = GetParam();
   }
 
@@ -4904,7 +5023,7 @@ TEST_F(DBCompactionTest, ManualCompactionFailsInReadOnlyMode) {
   // is in read-only mode. Verify it now at least returns, despite failing.
   const int kNumL0Files = 4;
   std::unique_ptr<FaultInjectionTestEnv> mock_env(
-      new FaultInjectionTestEnv(Env::Default()));
+      new FaultInjectionTestEnv(env_));
   Options opts = CurrentOptions();
   opts.disable_auto_compactions = true;
   opts.env = mock_env.get();
@@ -5131,7 +5250,7 @@ TEST_F(DBCompactionTest, ConsistencyFailTest2) {
   options.level0_file_num_compaction_trigger = 2;
   BlockBasedTableOptions bbto;
   bbto.block_size = 400;  // small block size
-  options.table_factory.reset(new BlockBasedTableFactory(bbto));
+  options.table_factory.reset(NewBlockBasedTableFactory(bbto));
   DestroyAndReopen(options);
 
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
@@ -5197,8 +5316,14 @@ TEST_P(DBCompactionTestWithParam,
   std::atomic<int> pick_intra_l0_count(0);
   std::string value(rnd.RandomString(kValueSize));
 
+  // The L0->L1 must be picked before we begin ingesting files to trigger
+  // intra-L0 compaction, and must not finish until after an intra-L0
+  // compaction has been picked.
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency(
-      {{"DBCompactionTestWithParam::FlushAfterIntraL0:1",
+      {{"LevelCompactionPicker::PickCompaction:Return",
+        "DBCompactionTestWithParam::"
+        "FlushAfterIntraL0CompactionCheckConsistencyFail:L0ToL1Ready"},
+       {"LevelCompactionPicker::PickCompactionBySize:0",
         "CompactionJob::Run():Start"}});
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
       "FindIntraL0Compaction",
@@ -5226,14 +5351,16 @@ TEST_P(DBCompactionTestWithParam,
   ASSERT_OK(Put(Key(0), "a"));
 
   ASSERT_EQ(5, NumTableFilesAtLevel(0));
+  TEST_SYNC_POINT(
+      "DBCompactionTestWithParam::"
+      "FlushAfterIntraL0CompactionCheckConsistencyFail:L0ToL1Ready");
 
   // Ingest 5 L0 sst. And this files would trigger PickIntraL0Compaction.
   for (int i = 5; i < 10; i++) {
+    ASSERT_EQ(i, NumTableFilesAtLevel(0));
     IngestOneKeyValue(dbfull(), Key(i), value, options);
-    ASSERT_EQ(i + 1, NumTableFilesAtLevel(0));
   }
 
-  TEST_SYNC_POINT("DBCompactionTestWithParam::FlushAfterIntraL0:1");
   // Put one key, to make biggest log sequence number in this memtable is bigger
   // than sst which would be ingested in next step.
   ASSERT_OK(Put(Key(2), "b"));
@@ -5275,8 +5402,14 @@ TEST_P(DBCompactionTestWithParam,
   ASSERT_EQ(0, NumTableFilesAtLevel(0));
 
   std::atomic<int> pick_intra_l0_count(0);
+  // The L0->L1 must be picked before we begin ingesting files to trigger
+  // intra-L0 compaction, and must not finish until after an intra-L0
+  // compaction has been picked.
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency(
-      {{"DBCompactionTestWithParam::IntraL0CompactionAfterFlush:1",
+      {{"LevelCompactionPicker::PickCompaction:Return",
+        "DBCompactionTestWithParam::"
+        "IntraL0CompactionAfterFlushCheckConsistencyFail:L0ToL1Ready"},
+       {"LevelCompactionPicker::PickCompactionBySize:0",
         "CompactionJob::Run():Start"}});
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
       "FindIntraL0Compaction",
@@ -5307,17 +5440,18 @@ TEST_P(DBCompactionTestWithParam,
   }
 
   ASSERT_EQ(6, NumTableFilesAtLevel(0));
+  TEST_SYNC_POINT(
+      "DBCompactionTestWithParam::"
+      "IntraL0CompactionAfterFlushCheckConsistencyFail:L0ToL1Ready");
   // ingest file to trigger IntraL0Compaction
   for (int i = 6; i < 10; ++i) {
     ASSERT_EQ(i, NumTableFilesAtLevel(0));
     IngestOneKeyValue(dbfull(), Key(i), value2, options);
   }
-  ASSERT_EQ(10, NumTableFilesAtLevel(0));
 
   // Wake up flush job
   sleeping_tasks.WakeUp();
   sleeping_tasks.WaitUntilDone();
-  TEST_SYNC_POINT("DBCompactionTestWithParam::IntraL0CompactionAfterFlush:1");
   dbfull()->TEST_WaitForCompact();
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
 
@@ -5333,6 +5467,47 @@ TEST_P(DBCompactionTestWithParam,
   }
 }
 
+TEST_P(DBCompactionTestWithBottommostParam, SequenceKeysManualCompaction) {
+  constexpr int kSstNum = 10;
+  Options options = CurrentOptions();
+  options.disable_auto_compactions = true;
+  DestroyAndReopen(options);
+
+  // Generate some sst files on level 0 with sequence keys (no overlap)
+  for (int i = 0; i < kSstNum; i++) {
+    for (int j = 1; j < UCHAR_MAX; j++) {
+      auto key = std::string(kSstNum, '\0');
+      key[kSstNum - i] += static_cast<char>(j);
+      Put(key, std::string(i % 1000, 'A'));
+    }
+    ASSERT_OK(Flush());
+  }
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
+
+  ASSERT_EQ(ToString(kSstNum), FilesPerLevel(0));
+
+  auto cro = CompactRangeOptions();
+  cro.bottommost_level_compaction = bottommost_level_compaction_;
+  db_->CompactRange(cro, nullptr, nullptr);
+  if (bottommost_level_compaction_ == BottommostLevelCompaction::kForce ||
+      bottommost_level_compaction_ ==
+          BottommostLevelCompaction::kForceOptimized) {
+    // Real compaction to compact all sst files from level 0 to 1 file on level
+    // 1
+    ASSERT_EQ("0,1", FilesPerLevel(0));
+  } else {
+    // Just trivial move from level 0 -> 1
+    ASSERT_EQ("0," + ToString(kSstNum), FilesPerLevel(0));
+  }
+}
+
+INSTANTIATE_TEST_CASE_P(
+    DBCompactionTestWithBottommostParam, DBCompactionTestWithBottommostParam,
+    ::testing::Values(BottommostLevelCompaction::kSkip,
+                      BottommostLevelCompaction::kIfHaveCompactionFilter,
+                      BottommostLevelCompaction::kForce,
+                      BottommostLevelCompaction::kForceOptimized));
+
 TEST_F(DBCompactionTest, UpdateLevelSubCompactionTest) {
   Options options = CurrentOptions();
   options.max_subcompactions = 10;
@@ -5435,7 +5610,418 @@ TEST_F(DBCompactionTest, UpdateUniversalSubCompactionTest) {
   ASSERT_TRUE(has_compaction);
 }
 
-#endif // !defined(ROCKSDB_LITE)
+TEST_P(ChangeLevelConflictsWithAuto, TestConflict) {
+  // A `CompactRange()` may race with an automatic compaction, we'll need
+  // to make sure it doesn't corrupte the data.
+  Options options = CurrentOptions();
+  options.level0_file_num_compaction_trigger = 2;
+  Reopen(options);
+
+  ASSERT_OK(Put("foo", "v1"));
+  ASSERT_OK(Put("bar", "v1"));
+  ASSERT_OK(Flush());
+  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
+
+  {
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 2;
+    ASSERT_OK(dbfull()->CompactRange(cro, nullptr, nullptr));
+  }
+  ASSERT_EQ("0,0,1", FilesPerLevel(0));
+
+  // Run a qury to refitting to level 1 while another thread writing to
+  // the same level.
+  SyncPoint::GetInstance()->LoadDependency({
+      // The first two dependencies ensure the foreground creates an L0 file
+      // between the background compaction's L0->L1 and its L1->L2.
+      {
+          "DBImpl::CompactRange:BeforeRefit:1",
+          "AutoCompactionFinished1",
+      },
+      {
+          "AutoCompactionFinished2",
+          "DBImpl::CompactRange:BeforeRefit:2",
+      },
+  });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  std::thread auto_comp([&] {
+    TEST_SYNC_POINT("AutoCompactionFinished1");
+    ASSERT_OK(Put("bar", "v2"));
+    ASSERT_OK(Put("foo", "v2"));
+    ASSERT_OK(Flush());
+    ASSERT_OK(Put("bar", "v3"));
+    ASSERT_OK(Put("foo", "v3"));
+    ASSERT_OK(Flush());
+    dbfull()->TEST_WaitForCompact();
+    TEST_SYNC_POINT("AutoCompactionFinished2");
+  });
+
+  {
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = GetParam() ? 1 : 0;
+    // This should return non-OK, but it's more important for the test to
+    // make sure that the DB is not corrupted.
+    dbfull()->CompactRange(cro, nullptr, nullptr);
+  }
+  auto_comp.join();
+  // Refitting didn't happen.
+  SyncPoint::GetInstance()->DisableProcessing();
+
+  // Write something to DB just make sure that consistency check didn't
+  // fail and make the DB readable.
+}
+
+INSTANTIATE_TEST_CASE_P(ChangeLevelConflictsWithAuto,
+                        ChangeLevelConflictsWithAuto, testing::Bool());
+
+TEST_F(DBCompactionTest, ChangeLevelCompactRangeConflictsWithManual) {
+  // A `CompactRange()` with `change_level == true` needs to execute its final
+  // step, `ReFitLevel()`, in isolation. Previously there was a bug where
+  // refitting could target the same level as an ongoing manual compaction,
+  // leading to overlapping files in that level.
+  //
+  // This test ensures that case is not possible by verifying any manual
+  // compaction issued during the `ReFitLevel()` phase fails with
+  // `Status::Incomplete`.
+  Options options = CurrentOptions();
+  options.memtable_factory.reset(
+      new SpecialSkipListFactory(KNumKeysByGenerateNewFile - 1));
+  options.level0_file_num_compaction_trigger = 2;
+  options.num_levels = 3;
+  Reopen(options);
+
+  // Setup an LSM with three levels populated.
+  Random rnd(301);
+  int key_idx = 0;
+  GenerateNewFile(&rnd, &key_idx);
+  {
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 2;
+    ASSERT_OK(dbfull()->CompactRange(cro, nullptr, nullptr));
+  }
+  ASSERT_EQ("0,0,2", FilesPerLevel(0));
+
+  GenerateNewFile(&rnd, &key_idx);
+  GenerateNewFile(&rnd, &key_idx);
+  ASSERT_EQ("1,1,2", FilesPerLevel(0));
+
+  // The background thread will refit L2->L1 while the
+  // foreground thread will try to simultaneously compact L0->L1.
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->LoadDependency({
+      // The first two dependencies ensure the foreground creates an L0 file
+      // between the background compaction's L0->L1 and its L1->L2.
+      {
+          "DBImpl::RunManualCompaction()::1",
+          "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:"
+          "PutFG",
+      },
+      {
+          "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:"
+          "FlushedFG",
+          "DBImpl::RunManualCompaction()::2",
+      },
+      // The next two dependencies ensure the foreground invokes
+      // `CompactRange()` while the background is refitting. The
+      // foreground's `CompactRange()` is guaranteed to attempt an L0->L1
+      // as we set it up with an empty memtable and a new L0 file.
+      {
+          "DBImpl::CompactRange:PreRefitLevel",
+          "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:"
+          "CompactFG",
+      },
+      {
+          "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:"
+          "CompactedFG",
+          "DBImpl::CompactRange:PostRefitLevel",
+      },
+  });
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
+
+  ROCKSDB_NAMESPACE::port::Thread refit_level_thread([&] {
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 1;
+    ASSERT_OK(dbfull()->CompactRange(cro, nullptr, nullptr));
+  });
+
+  TEST_SYNC_POINT(
+      "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:PutFG");
+  // Make sure we have something new to compact in the foreground.
+  // Note key 1 is carefully chosen as it ensures the file we create here
+  // overlaps with one of the files being refitted L2->L1 in the background.
+  // If we chose key 0, the file created here would not overlap.
+  ASSERT_OK(Put(Key(1), "val"));
+  ASSERT_OK(Flush());
+  TEST_SYNC_POINT(
+      "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:FlushedFG");
+
+  TEST_SYNC_POINT(
+      "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:CompactFG");
+  ASSERT_TRUE(dbfull()
+                  ->CompactRange(CompactRangeOptions(), nullptr, nullptr)
+                  .IsIncomplete());
+  TEST_SYNC_POINT(
+      "DBCompactionTest::ChangeLevelCompactRangeConflictsWithManual:"
+      "CompactedFG");
+  refit_level_thread.join();
+}
+
+TEST_F(DBCompactionTest, ChangeLevelErrorPathTest) {
+  // This test is added to ensure that RefitLevel() error paths are clearing
+  // internal flags and to test that subsequent valid RefitLevel() calls
+  // succeeds
+  Options options = CurrentOptions();
+  options.memtable_factory.reset(
+      new SpecialSkipListFactory(KNumKeysByGenerateNewFile - 1));
+  options.level0_file_num_compaction_trigger = 2;
+  options.num_levels = 3;
+  Reopen(options);
+
+  ASSERT_EQ("", FilesPerLevel(0));
+
+  // Setup an LSM with three levels populated.
+  Random rnd(301);
+  int key_idx = 0;
+  GenerateNewFile(&rnd, &key_idx);
+  ASSERT_EQ("1", FilesPerLevel(0));
+  {
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 2;
+    ASSERT_OK(dbfull()->CompactRange(cro, nullptr, nullptr));
+  }
+  ASSERT_EQ("0,0,2", FilesPerLevel(0));
+
+  auto start_idx = key_idx;
+  GenerateNewFile(&rnd, &key_idx);
+  GenerateNewFile(&rnd, &key_idx);
+  auto end_idx = key_idx - 1;
+  ASSERT_EQ("1,1,2", FilesPerLevel(0));
+
+  // Next two CompactRange() calls are used to test exercise error paths within
+  // RefitLevel() before triggering a valid RefitLevel() call
+
+  // Trigger a refit to L1 first
+  {
+    std::string begin_string = Key(start_idx);
+    std::string end_string = Key(end_idx);
+    Slice begin(begin_string);
+    Slice end(end_string);
+
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 1;
+    ASSERT_OK(dbfull()->CompactRange(cro, &begin, &end));
+  }
+  ASSERT_EQ("0,3,2", FilesPerLevel(0));
+
+  // Try a refit from L2->L1 - this should fail and exercise error paths in
+  // RefitLevel()
+  {
+    // Select key range that matches the bottom most level (L2)
+    std::string begin_string = Key(0);
+    std::string end_string = Key(start_idx - 1);
+    Slice begin(begin_string);
+    Slice end(end_string);
+
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 1;
+    ASSERT_NOK(dbfull()->CompactRange(cro, &begin, &end));
+  }
+  ASSERT_EQ("0,3,2", FilesPerLevel(0));
+
+  // Try a valid Refit request to ensure, the path is still working
+  {
+    CompactRangeOptions cro;
+    cro.change_level = true;
+    cro.target_level = 1;
+    ASSERT_OK(dbfull()->CompactRange(cro, nullptr, nullptr));
+  }
+  ASSERT_EQ("0,5", FilesPerLevel(0));
+}
+
+TEST_F(DBCompactionTest, CompactionWithBlob) {
+  Options options;
+  options.env = env_;
+  options.disable_auto_compactions = true;
+
+  Reopen(options);
+
+  constexpr char first_key[] = "first_key";
+  constexpr char second_key[] = "second_key";
+  constexpr char first_value[] = "first_value";
+  constexpr char second_value[] = "second_value";
+  constexpr char third_value[] = "third_value";
+
+  ASSERT_OK(Put(first_key, first_value));
+  ASSERT_OK(Put(second_key, first_value));
+  ASSERT_OK(Flush());
+
+  ASSERT_OK(Put(first_key, second_value));
+  ASSERT_OK(Put(second_key, second_value));
+  ASSERT_OK(Flush());
+
+  ASSERT_OK(Put(first_key, third_value));
+  ASSERT_OK(Put(second_key, third_value));
+  ASSERT_OK(Flush());
+
+  options.enable_blob_files = true;
+
+  Reopen(options);
+
+  constexpr Slice* begin = nullptr;
+  constexpr Slice* end = nullptr;
+
+  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), begin, end));
+
+  ASSERT_EQ(Get(first_key), third_value);
+  ASSERT_EQ(Get(second_key), third_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& l1_files = storage_info->LevelFiles(1);
+  ASSERT_EQ(l1_files.size(), 1);
+
+  const FileMetaData* const table_file = l1_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(), first_key);
+  ASSERT_EQ(table_file->largest.user_key(), second_key);
+  ASSERT_EQ(table_file->oldest_blob_file_number,
+            blob_file->GetBlobFileNumber());
+
+  ASSERT_EQ(blob_file->GetTotalBlobCount(), 2);
+
+  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_GE(compaction_stats.size(), 2);
+  ASSERT_EQ(compaction_stats[1].bytes_written, expected_bytes);
+  ASSERT_EQ(compaction_stats[1].num_output_files, 2);
+}
+
+class DBCompactionTestBlobError
+    : public DBCompactionTest,
+      public testing::WithParamInterface<std::string> {
+ public:
+  DBCompactionTestBlobError()
+      : fault_injection_env_(env_), sync_point_(GetParam()) {}
+  ~DBCompactionTestBlobError() { Close(); }
+
+  FaultInjectionTestEnv fault_injection_env_;
+  std::string sync_point_;
+};
+
+INSTANTIATE_TEST_CASE_P(DBCompactionTestBlobError, DBCompactionTestBlobError,
+                        ::testing::ValuesIn(std::vector<std::string>{
+                            "BlobFileBuilder::WriteBlobToFile:AddRecord",
+                            "BlobFileBuilder::WriteBlobToFile:AppendFooter"}));
+
+TEST_P(DBCompactionTestBlobError, CompactionError) {
+  Options options;
+  options.disable_auto_compactions = true;
+  options.env = env_;
+
+  Reopen(options);
+
+  constexpr char first_key[] = "first_key";
+  constexpr char second_key[] = "second_key";
+  constexpr char first_value[] = "first_value";
+  constexpr char second_value[] = "second_value";
+  constexpr char third_value[] = "third_value";
+
+  ASSERT_OK(Put(first_key, first_value));
+  ASSERT_OK(Put(second_key, first_value));
+  ASSERT_OK(Flush());
+
+  ASSERT_OK(Put(first_key, second_value));
+  ASSERT_OK(Put(second_key, second_value));
+  ASSERT_OK(Flush());
+
+  ASSERT_OK(Put(first_key, third_value));
+  ASSERT_OK(Put(second_key, third_value));
+  ASSERT_OK(Flush());
+
+  options.enable_blob_files = true;
+  options.env = &fault_injection_env_;
+
+  Reopen(options);
+
+  SyncPoint::GetInstance()->SetCallBack(sync_point_, [this](void* /* arg */) {
+    fault_injection_env_.SetFilesystemActive(false,
+                                             Status::IOError(sync_point_));
+  });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  constexpr Slice* begin = nullptr;
+  constexpr Slice* end = nullptr;
+
+  ASSERT_TRUE(db_->CompactRange(CompactRangeOptions(), begin, end).IsIOError());
+
+  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& l1_files = storage_info->LevelFiles(1);
+  ASSERT_TRUE(l1_files.empty());
+
+  const auto& blob_files = storage_info->GetBlobFiles();
+  ASSERT_TRUE(blob_files.empty());
+
+  const InternalStats* const internal_stats = cfd->internal_stats();
+  assert(internal_stats);
+
+  const auto& compaction_stats = internal_stats->TEST_GetCompactionStats();
+  ASSERT_GE(compaction_stats.size(), 2);
+
+  if (sync_point_ == "BlobFileBuilder::WriteBlobToFile:AddRecord") {
+    ASSERT_EQ(compaction_stats[1].bytes_written, 0);
+    ASSERT_EQ(compaction_stats[1].num_output_files, 0);
+  } else {
+    // SST file writing succeeded; blob file writing failed (during Finish)
+    ASSERT_GT(compaction_stats[1].bytes_written, 0);
+    ASSERT_EQ(compaction_stats[1].num_output_files, 1);
+  }
+}
+
+#endif  // !defined(ROCKSDB_LITE)
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/db_dynamic_level_test.cc

@@ -20,7 +20,8 @@
 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) {
@@ -140,6 +141,7 @@ TEST_F(DBTestDynamicLevel, DynamicLevelMaxBytesBase2) {
   options.max_background_compactions = 2;
   options.num_levels = 5;
   options.max_compaction_bytes = 0;  // Force not expanding in compactions
+  options.db_host_id = "";  // Setting this messes up the file size calculation
   BlockBasedTableOptions table_options;
   table_options.block_size = 1024;
   options.table_factory.reset(NewBlockBasedTableFactory(table_options));

db/db_encryption_test.cc

@@ -16,7 +16,15 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBEncryptionTest : public DBTestBase {
  public:
-  DBEncryptionTest() : DBTestBase("/db_encryption_test") {}
+  DBEncryptionTest()
+      : DBTestBase("/db_encryption_test", /*env_do_fsync=*/true) {}
+  Env* GetTargetEnv() {
+    if (encrypted_env_ != nullptr) {
+      return (static_cast<EnvWrapper*>(encrypted_env_))->target();
+    } else {
+      return env_;
+    }
+  }
 };
 
 #ifndef ROCKSDB_LITE
@@ -33,20 +41,20 @@ TEST_F(DBEncryptionTest, CheckEncrypted) {
   auto status = env_->GetChildren(dbname_, &fileNames);
   ASSERT_OK(status);
 
-  auto defaultEnv = Env::Default();
+  Env* target = GetTargetEnv();
   int hits = 0;
   for (auto it = fileNames.begin() ; it != fileNames.end(); ++it) {
-    if ((*it == "..") || (*it == ".")) {
+    if ((*it == "..") || (*it == ".") || (*it == "LOCK")) {
       continue;
     }
     auto filePath = dbname_ + "/" + *it;
     std::unique_ptr<SequentialFile> seqFile;
     auto envOptions = EnvOptions(CurrentOptions());
-    status = defaultEnv->NewSequentialFile(filePath, &seqFile, envOptions);
+    status = target->NewSequentialFile(filePath, &seqFile, envOptions);
     ASSERT_OK(status);
 
     uint64_t fileSize;
-    status = defaultEnv->GetFileSize(filePath, &fileSize);
+    status = target->GetFileSize(filePath, &fileSize);
     ASSERT_OK(status);
 
     std::string scratch;
@@ -84,7 +92,7 @@ TEST_F(DBEncryptionTest, CheckEncrypted) {
 }
 
 TEST_F(DBEncryptionTest, ReadEmptyFile) {
-  auto defaultEnv = Env::Default();
+  auto defaultEnv = GetTargetEnv();
 
   // create empty file for reading it back in later
   auto envOptions = EnvOptions(CurrentOptions());

db/db_filesnapshot.cc

@@ -116,7 +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

@@ -11,6 +11,7 @@
 
 #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/sync_point.h"
@@ -22,7 +23,7 @@ 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
@@ -87,7 +88,7 @@ TEST_F(DBFlushTest, SyncFail) {
   SyncPoint::GetInstance()->EnableProcessing();
 
   CreateAndReopenWithCF({"pikachu"}, options);
-  Put("key", "value");
+  ASSERT_OK(Put("key", "value"));
   auto* cfd =
       static_cast_with_check<ColumnFamilyHandleImpl>(db_->DefaultColumnFamily())
           ->cfd();
@@ -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,
@@ -443,6 +446,180 @@ TEST_F(DBFlushTest, FireOnFlushCompletedAfterCommittedResult) {
 }
 #endif  // !ROCKSDB_LITE
 
+TEST_F(DBFlushTest, FlushWithBlob) {
+  constexpr uint64_t min_blob_size = 10;
+
+  Options options;
+  options.env = CurrentOptions().env;
+  options.enable_blob_files = true;
+  options.min_blob_size = min_blob_size;
+  options.disable_auto_compactions = true;
+  options.env = env_;
+
+  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);
+  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_), sync_point_(GetParam()) {}
+  ~DBFlushTestBlobError() { Close(); }
+
+  FaultInjectionTestEnv fault_injection_env_;
+  std::string sync_point_;
+};
+
+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(sync_point_, [this](void* /* arg */) {
+    fault_injection_env_.SetFilesystemActive(false,
+                                             Status::IOError(sync_point_));
+  });
+  SyncPoint::GetInstance()->SetCallBack(
+      "BuildTable:BeforeDeleteFile", [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());
+
+  if (sync_point_ == "BlobFileBuilder::WriteBlobToFile:AddRecord") {
+    ASSERT_EQ(compaction_stats[0].bytes_written, 0);
+    ASSERT_EQ(compaction_stats[0].num_output_files, 0);
+  } else {
+    // SST file writing succeeded; blob file writing failed (during Finish)
+    ASSERT_GT(compaction_stats[0].bytes_written, 0);
+    ASSERT_EQ(compaction_stats[0].num_output_files, 1);
+  }
+
+  const uint64_t* const cf_stats_value = internal_stats->TEST_GetCFStatsValue();
+  ASSERT_EQ(cf_stats_value[InternalStats::BYTES_FLUSHED],
+            compaction_stats[0].bytes_written);
+#endif  // ROCKSDB_LITE
+}
+
 TEST_P(DBAtomicFlushTest, ManualAtomicFlush) {
   Options options = CurrentOptions();
   options.create_if_missing = true;
@@ -549,7 +726,8 @@ 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;
@@ -426,8 +431,28 @@ class DBImpl : public DB {
       const ExportImportFilesMetaData& metadata,
       ColumnFamilyHandle** handle) override;
 
+  using DB::VerifyFileChecksums;
+  Status VerifyFileChecksums(const ReadOptions& read_options) override;
+
   using DB::VerifyChecksum;
   virtual Status VerifyChecksum(const ReadOptions& /*read_options*/) override;
+  // Verify the checksums of files in db. Currently only tables are checked.
+  //
+  // read_options: controls file I/O behavior, e.g. read ahead size while
+  //               reading all the live table files.
+  //
+  // use_file_checksum: if false, verify the block checksums of all live table
+  //                    in db. Otherwise, obtain the file checksums and compare
+  //                    with the MANIFEST. Currently, file checksums are
+  //                    recomputed by reading all table files.
+  //
+  // Returns: OK if there is no file whose file or block checksum mismatches.
+  Status VerifyChecksumInternal(const ReadOptions& read_options,
+                                bool use_file_checksum);
+
+  Status VerifySstFileChecksum(const FileMetaData& fmeta,
+                               const std::string& fpath,
+                               const ReadOptions& read_options);
 
   using DB::StartTrace;
   virtual Status StartTrace(
@@ -445,6 +470,13 @@ 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,
@@ -594,8 +626,10 @@ 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);
 
@@ -721,10 +755,14 @@ 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
@@ -974,9 +1012,7 @@ 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(); }
@@ -984,8 +1020,22 @@ 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_;
@@ -997,8 +1047,9 @@ 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*>
@@ -1158,20 +1209,28 @@ 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; }
 
+  // Set DB identity file, and write DB ID to manifest if necessary.
+  Status SetDBId();
+
   // REQUIRES: db mutex held when calling this function, but the db mutex can
   // be released and re-acquired. Db mutex will be held when the function
   // returns.
-  // After best-efforts recovery, there may be SST files in db/cf paths that are
-  // not referenced in the MANIFEST. We delete these SST files. In the
+  // After recovery, there may be SST files in db/cf paths that are
+  // not referenced in the MANIFEST (e.g.
+  // 1. It's best effort recovery;
+  // 2. The VersionEdits referencing the SST files are appended to
+  // MANIFEST, DB crashes when syncing the MANIFEST, the VersionEdits are
+  // still not synced to MANIFEST during recovery.)
+  // We delete these SST files. In the
   // meantime, we find out the largest file number present in the paths, and
   // bump up the version set's next_file_number_ to be 1 + largest_file_number.
-  Status FinishBestEffortsRecovery();
+  Status DeleteUnreferencedSstFiles();
 
   // SetDbSessionId() should be called in the constuctor DBImpl()
   // to ensure that db_session_id_ gets updated every time the DB is opened
@@ -1351,7 +1410,7 @@ class DBImpl : public DB {
   // Required: DB mutex held
   Status PersistentStatsProcessFormatVersion();
 
-  Status ResumeImpl();
+  Status ResumeImpl(DBRecoverContext context);
 
   void MaybeIgnoreError(Status* s) const;
 
@@ -1539,6 +1598,10 @@ 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);
 
@@ -1620,18 +1683,12 @@ 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,
@@ -1653,7 +1710,9 @@ class DBImpl : public DB {
       std::unique_ptr<TaskLimiterToken>* token, LogBuffer* log_buffer);
 
   // helper function to call after some of the logs_ were synced
-  void MarkLogsSynced(uint64_t up_to, bool synced_dir, const Status& status);
+  Status MarkLogsSynced(uint64_t up_to, bool synced_dir);
+  // WALs with log number up to up_to are not synced successfully.
+  void MarkLogsNotSynced(uint64_t up_to);
 
   SnapshotImpl* GetSnapshotImpl(bool is_write_conflict_boundary,
                                 bool lock = true);
@@ -1814,7 +1873,14 @@ 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
@@ -2077,13 +2143,13 @@ 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.
@@ -2148,12 +2214,18 @@ extern CompressionType GetCompressionFlush(
 // `memtables_to_flush`) will be flushed and thus will not depend on any WAL
 // file.
 // The function is only applicable to 2pc mode.
-extern uint64_t PrecomputeMinLogNumberToKeep(
+extern uint64_t PrecomputeMinLogNumberToKeep2PC(
     VersionSet* vset, const ColumnFamilyData& cfd_to_flush,
-    autovector<VersionEdit*> edit_list,
+    const autovector<VersionEdit*>& edit_list,
     const autovector<MemTable*>& memtables_to_flush,
     LogsWithPrepTracker* prep_tracker);
 
+// In non-2PC mode, WALs with log number < the returned number can be
+// deleted after the cfd_to_flush column family is flushed successfully.
+extern uint64_t PrecomputeMinLogNumberToKeepNon2PC(
+    VersionSet* vset, const ColumnFamilyData& cfd_to_flush,
+    const autovector<VersionEdit*>& edit_list);
+
 // `cfd_to_flush` is the column family whose memtable will be flushed and thus
 // will not depend on any WAL file. nullptr means no memtable is being flushed.
 // The function is only applicable to 2pc mode.

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"
@@ -124,9 +123,20 @@ IOStatus DBImpl::SyncClosedLogs(JobContext* job_context) {
 
     // "number <= current_log_number - 1" is equivalent to
     // "number < current_log_number".
-    MarkLogsSynced(current_log_number - 1, true, io_s);
+    if (io_s.ok()) {
+      io_s = status_to_io_status(MarkLogsSynced(current_log_number - 1, true));
+    } else {
+      MarkLogsNotSynced(current_log_number - 1);
+    }
     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 +153,7 @@ 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 +166,7 @@ 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 +179,7 @@ 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 +190,11 @@ 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");
   }
@@ -204,10 +220,28 @@ 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()) {
@@ -218,14 +252,30 @@ Status DBImpl::FlushMemTableToOutputFile(
       // 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::kFlush
-                            : BackgroundErrorReason::kManifestWrite;
-      error_handler_.SetBGError(io_s, err_reason);
+      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
@@ -245,7 +295,9 @@ 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
@@ -356,7 +408,7 @@ 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();
   }
 
@@ -541,16 +593,36 @@ 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;
@@ -573,8 +645,9 @@ 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();
         }
       }
     }
@@ -590,13 +663,25 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
       // 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::kFlush
-                            : BackgroundErrorReason::kManifestWrite;
-      error_handler_.SetBGError(io_s, err_reason);
+      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();
     }
   }
 
@@ -711,7 +796,9 @@ Status DBImpl::CompactRange(const CompactRangeOptions& options,
     // changes to RangesOverlapWithMemtables.
     Range range(*begin, *end);
     SuperVersion* super_version = cfd->GetReferencedSuperVersion(this);
-    cfd->RangesOverlapWithMemtables({range}, super_version, &flush_needed);
+    cfd->RangesOverlapWithMemtables({range}, super_version,
+                                    immutable_db_options_.allow_data_in_errors,
+                                    &flush_needed);
     CleanupSuperVersion(super_version);
   }
 
@@ -792,6 +879,7 @@ 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
@@ -824,9 +912,19 @@ 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;
         }
@@ -846,13 +944,22 @@ 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().
+      Status temp_s = ContinueBackgroundWork();
+      assert(temp_s.ok());
     }
-    ContinueBackgroundWork();
+    EnableManualCompaction();
   }
   LogFlush(immutable_db_options_.info_log);
 
@@ -959,7 +1066,7 @@ 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);
   }
 
@@ -1046,13 +1153,14 @@ Status DBImpl::CompactFilesImpl(
       job_context->job_id, c.get(), immutable_db_options_,
       file_options_for_compaction_, versions_.get(), &shutting_down_,
       preserve_deletes_seqnum_.load(), log_buffer, directories_.GetDbDir(),
-      GetDataDir(c->column_family_data(), c->output_path_id()), stats_, &mutex_,
-      &error_handler_, snapshot_seqs, earliest_write_conflict_snapshot,
-      snapshot_checker, table_cache_, &event_logger_,
+      GetDataDir(c->column_family_data(), c->output_path_id()),
+      GetDataDir(c->column_family_data(), 0), stats_, &mutex_, &error_handler_,
+      snapshot_seqs, 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, 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
@@ -1073,10 +1181,15 @@ 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
@@ -1111,9 +1224,11 @@ Status DBImpl::CompactFilesImpl(
                    job_context->job_id, status.ToString().c_str());
     IOStatus io_s = compaction_job.io_status();
     if (!io_s.ok()) {
-      error_handler_.SetBGError(io_s, BackgroundErrorReason::kCompaction);
+      error_handler_.SetBGError(io_s, BackgroundErrorReason::kCompaction)
+          .PermitUncheckedError();
     } else {
-      error_handler_.SetBGError(status, BackgroundErrorReason::kCompaction);
+      error_handler_.SetBGError(status, BackgroundErrorReason::kCompaction)
+          .PermitUncheckedError();
     }
   }
 
@@ -1180,7 +1295,7 @@ 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();
@@ -1190,46 +1305,11 @@ 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();
@@ -1254,7 +1334,7 @@ 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();
@@ -1313,20 +1393,32 @@ 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());
@@ -1615,6 +1707,7 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
       return s;
     }
   }
+
   FlushRequest flush_req;
   {
     WriteContext context;
@@ -1631,7 +1724,11 @@ 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() ||
@@ -1639,7 +1736,8 @@ 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);
@@ -1668,7 +1766,6 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
         }
       }
     }
-
     if (s.ok() && !flush_req.empty()) {
       for (auto& elem : flush_req) {
         ColumnFamilyData* loop_cfd = elem.first;
@@ -1704,8 +1801,10 @@ 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();
@@ -1763,7 +1862,8 @@ 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();
@@ -1806,8 +1906,10 @@ 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();
@@ -1917,6 +2019,13 @@ 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.
@@ -1965,11 +2074,21 @@ 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() {
@@ -2528,7 +2647,7 @@ 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 {
@@ -2570,12 +2689,13 @@ 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(
@@ -2841,23 +2961,24 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
         job_context->job_id, c.get(), immutable_db_options_,
         file_options_for_compaction_, versions_.get(), &shutting_down_,
         preserve_deletes_seqnum_.load(), log_buffer, directories_.GetDbDir(),
-        GetDataDir(c->column_family_data(), c->output_path_id()), stats_,
-        &mutex_, &error_handler_, snapshot_seqs,
-        earliest_write_conflict_snapshot, snapshot_checker, table_cache_,
-        &event_logger_, c->mutable_cf_options()->paranoid_file_checks,
+        GetDataDir(c->column_family_data(), c->output_path_id()),
+        GetDataDir(c->column_family_data(), 0), stats_, &mutex_,
+        &error_handler_, snapshot_seqs, 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();
 
@@ -2875,6 +2996,8 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
 
   if (status.ok() && !io_s.ok()) {
     status = io_s;
+  } else {
+    io_s.PermitUncheckedError();
   }
 
   if (c != nullptr) {
@@ -2911,9 +3034,10 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
       auto err_reason = versions_->io_status().ok()
                             ? BackgroundErrorReason::kCompaction
                             : BackgroundErrorReason::kManifestWrite;
-      error_handler_.SetBGError(io_s, err_reason);
+      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"
 
@@ -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_files.cc

@@ -58,7 +58,7 @@ 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;
+  int saved_counter;  // initialize on all paths
   {
     InstrumentedMutexLock l(&mutex_);
     if (force) {
@@ -67,13 +67,13 @@ Status DBImpl::EnableFileDeletions(bool force) {
     } else if (disable_delete_obsolete_files_ > 0) {
       --disable_delete_obsolete_files_;
     }
-    if (disable_delete_obsolete_files_ == 0) {
-      file_deletion_enabled = true;
+    saved_counter = disable_delete_obsolete_files_;
+    if (saved_counter == 0) {
       FindObsoleteFiles(&job_context, true);
       bg_cv_.SignalAll();
     }
   }
-  if (file_deletion_enabled) {
+  if (saved_counter == 0) {
     ROCKS_LOG_INFO(immutable_db_options_.info_log, "File Deletions Enabled");
     if (job_context.HaveSomethingToDelete()) {
       PurgeObsoleteFiles(job_context);
@@ -81,7 +81,7 @@ Status DBImpl::EnableFileDeletions(bool force) {
   } else {
     ROCKS_LOG_WARN(immutable_db_options_.info_log,
                    "File Deletions Enable, but not really enabled. Counter: %d",
-                   disable_delete_obsolete_files_);
+                   saved_counter);
   }
   job_context.Clean();
   LogFlush(immutable_db_options_.info_log);
@@ -215,7 +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);
@@ -226,7 +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);
@@ -317,7 +319,7 @@ void DBImpl::DeleteObsoleteFileImpl(int job_id, const std::string& fname,
                            const_cast<std::string*>(&fname));
 
   Status file_deletion_status;
-  if (type == kTableFile || type == kBlobFile || type == kLogFile) {
+  if (type == kTableFile || type == kBlobFile || type == kWalFile) {
     file_deletion_status =
         DeleteDBFile(&immutable_db_options_, fname, path_to_sync,
                      /*force_bg=*/false, /*force_fg=*/!wal_in_db_path_);
@@ -447,7 +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();
@@ -463,7 +466,7 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
 
     bool keep = true;
     switch (type) {
-      case kLogFile:
+      case kWalFile:
         keep = ((number >= state.log_number) ||
                 (number == state.prev_log_number) ||
                 (log_recycle_files_set.find(number) !=
@@ -543,7 +546,7 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
       dir_to_sync = candidate_file.file_path;
     } else {
       dir_to_sync =
-          (type == kLogFile) ? immutable_db_options_.wal_dir : dbname_;
+          (type == kWalFile) ? immutable_db_options_.wal_dir : dbname_;
       fname = dir_to_sync +
               ((!dir_to_sync.empty() && dir_to_sync.back() == '/') ||
                        (!to_delete.empty() && to_delete.front() == '/')
@@ -553,7 +556,7 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
     }
 
 #ifndef ROCKSDB_LITE
-    if (type == kLogFile && (immutable_db_options_.wal_ttl_seconds > 0 ||
+    if (type == kWalFile && (immutable_db_options_.wal_ttl_seconds > 0 ||
                              immutable_db_options_.wal_size_limit_mb > 0)) {
       wal_manager_.ArchiveWALFile(fname, number);
       continue;
@@ -566,7 +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);
@@ -678,16 +680,10 @@ uint64_t FindMinPrepLogReferencedByMemTable(
   return min_log;
 }
 
-uint64_t PrecomputeMinLogNumberToKeep(
+uint64_t PrecomputeMinLogNumberToKeepNon2PC(
     VersionSet* vset, const ColumnFamilyData& cfd_to_flush,
-    autovector<VersionEdit*> edit_list,
-    const autovector<MemTable*>& memtables_to_flush,
-    LogsWithPrepTracker* prep_tracker) {
+    const autovector<VersionEdit*>& edit_list) {
   assert(vset != nullptr);
-  assert(prep_tracker != nullptr);
-  // Calculate updated min_log_number_to_keep
-  // Since the function should only be called in 2pc mode, log number in
-  // the version edit should be sufficient.
 
   // Precompute the min log number containing unflushed data for the column
   // family being flushed (`cfd_to_flush`).
@@ -711,6 +707,22 @@ uint64_t PrecomputeMinLogNumberToKeep(
     min_log_number_to_keep =
         std::min(cf_min_log_number_to_keep, min_log_number_to_keep);
   }
+  return min_log_number_to_keep;
+}
+
+uint64_t PrecomputeMinLogNumberToKeep2PC(
+    VersionSet* vset, const ColumnFamilyData& cfd_to_flush,
+    const autovector<VersionEdit*>& edit_list,
+    const autovector<MemTable*>& memtables_to_flush,
+    LogsWithPrepTracker* prep_tracker) {
+  assert(vset != nullptr);
+  assert(prep_tracker != nullptr);
+  // Calculate updated min_log_number_to_keep
+  // Since the function should only be called in 2pc mode, log number in
+  // the version edit should be sufficient.
+
+  uint64_t min_log_number_to_keep =
+      PrecomputeMinLogNumberToKeepNon2PC(vset, cfd_to_flush, edit_list);
 
   // if are 2pc we must consider logs containing prepared
   // sections of outstanding transactions.
@@ -739,7 +751,43 @@ uint64_t PrecomputeMinLogNumberToKeep(
   return min_log_number_to_keep;
 }
 
-Status DBImpl::FinishBestEffortsRecovery() {
+Status DBImpl::SetDBId() {
+  Status s;
+  // Happens when immutable_db_options_.write_dbid_to_manifest is set to true
+  // the very first time.
+  if (db_id_.empty()) {
+    // Check for the IDENTITY file and create it if not there.
+    s = fs_->FileExists(IdentityFileName(dbname_), IOOptions(), nullptr);
+    // Typically Identity file is created in NewDB() and for some reason if
+    // it is no longer available then at this point DB ID is not in Identity
+    // file or Manifest.
+    if (s.IsNotFound()) {
+      s = SetIdentityFile(env_, dbname_);
+      if (!s.ok()) {
+        return s;
+      }
+    } else if (!s.ok()) {
+      assert(s.IsIOError());
+      return s;
+    }
+    s = GetDbIdentityFromIdentityFile(&db_id_);
+    if (immutable_db_options_.write_dbid_to_manifest && s.ok()) {
+      VersionEdit edit;
+      edit.SetDBId(db_id_);
+      Options options;
+      MutableCFOptions mutable_cf_options(options);
+      versions_->db_id_ = db_id_;
+      s = versions_->LogAndApply(versions_->GetColumnFamilySet()->GetDefault(),
+                                 mutable_cf_options, &edit, &mutex_, nullptr,
+                                 /* new_descriptor_log */ false);
+    }
+  } else {
+    s = SetIdentityFile(env_, dbname_, db_id_);
+  }
+  return s;
+}
+
+Status DBImpl::DeleteUnreferencedSstFiles() {
   mutex_.AssertHeld();
   std::vector<std::string> paths;
   paths.push_back(NormalizePath(dbname_ + std::string(1, kFilePathSeparator)));
@@ -760,9 +808,13 @@ 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;
@@ -778,6 +830,10 @@ Status DBImpl::FinishBestEffortsRecovery() {
       }
     }
   }
+  if (!s.ok()) {
+    return s;
+  }
+
   if (largest_file_number > next_file_number) {
     versions_->next_file_number_.store(largest_file_number + 1);
   }
@@ -787,9 +843,7 @@ Status DBImpl::FinishBestEffortsRecovery() {
   assert(versions_->GetColumnFamilySet());
   ColumnFamilyData* default_cfd = versions_->GetColumnFamilySet()->GetDefault();
   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;
     }
@@ -280,8 +290,8 @@ Status DBImpl::NewDB(std::vector<std::string>* new_filenames) {
     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);
@@ -354,7 +364,7 @@ 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();
 
@@ -391,7 +401,7 @@ Status DBImpl::Recover(
       }
       for (const std::string& file : files_in_dbname) {
         uint64_t number = 0;
-        FileType type = kLogFile;  // initialize
+        FileType type = kWalFile;  // initialize
         if (ParseFileName(file, &number, &type) && type == kDescriptorFile) {
           // Found MANIFEST (descriptor log), thus best-efforts recovery does
           // not have to treat the db as empty.
@@ -469,42 +479,14 @@ Status DBImpl::Recover(
       // TryRecover may delete previous column_family_set_.
       column_family_memtables_.reset(
           new ColumnFamilyMemTablesImpl(versions_->GetColumnFamilySet()));
-      s = FinishBestEffortsRecovery();
     }
   }
   if (!s.ok()) {
     return s;
   }
-  // Happens when immutable_db_options_.write_dbid_to_manifest is set to true
-  // the very first time.
-  if (db_id_.empty()) {
-    // Check for the IDENTITY file and create it if not there.
-    s = fs_->FileExists(IdentityFileName(dbname_), IOOptions(), nullptr);
-    // Typically Identity file is created in NewDB() and for some reason if
-    // it is no longer available then at this point DB ID is not in Identity
-    // file or Manifest.
-    if (s.IsNotFound()) {
-      s = SetIdentityFile(env_, dbname_);
-      if (!s.ok()) {
-        return s;
-      }
-    } else if (!s.ok()) {
-      assert(s.IsIOError());
-      return s;
-    }
-    s = GetDbIdentityFromIdentityFile(&db_id_);
-    if (immutable_db_options_.write_dbid_to_manifest && s.ok()) {
-      VersionEdit edit;
-      edit.SetDBId(db_id_);
-      Options options;
-      MutableCFOptions mutable_cf_options(options);
-      versions_->db_id_ = db_id_;
-      s = versions_->LogAndApply(versions_->GetColumnFamilySet()->GetDefault(),
-                             mutable_cf_options, &edit, &mutex_, nullptr,
-                             false);
-    }
-  } else {
-    s = SetIdentityFile(env_, dbname_, db_id_);
+  s = SetDBId();
+  if (s.ok() && !read_only) {
+    s = DeleteUnreferencedSstFiles();
   }
 
   if (immutable_db_options_.paranoid_checks && s.ok()) {
@@ -526,7 +508,7 @@ Status DBImpl::Recover(
 
   std::vector<std::string> files_in_wal_dir;
   if (s.ok()) {
-    // Initial max_total_in_memory_state_ before recovery logs. Log recovery
+    // Initial max_total_in_memory_state_ before recovery wals. Log recovery
     // may check this value to decide whether to flush.
     max_total_in_memory_state_ = 0;
     for (auto cfd : *versions_->GetColumnFamilySet()) {
@@ -561,50 +543,77 @@ Status DBImpl::Recover(
       return s;
     }
 
-    std::vector<uint64_t> logs;
+    std::unordered_map<uint64_t, std::string> wal_files;
     for (const auto& file : files_in_wal_dir) {
       uint64_t number;
       FileType type;
-      if (ParseFileName(file, &number, &type) && type == kLogFile) {
+      if (ParseFileName(file, &number, &type) && type == kWalFile) {
         if (is_new_db) {
           return Status::Corruption(
               "While creating a new Db, wal_dir contains "
               "existing log file: ",
               file);
         } else {
-          logs.push_back(number);
+          wal_files[number] =
+              LogFileName(immutable_db_options_.wal_dir, number);
         }
       }
     }
 
-    if (logs.size() > 0) {
-      if (error_if_log_file_exist) {
+    if (immutable_db_options_.track_and_verify_wals_in_manifest) {
+      if (!immutable_db_options_.best_efforts_recovery) {
+        // Verify WALs in MANIFEST.
+        s = versions_->GetWalSet().CheckWals(env_, wal_files);
+      }  // else since best effort recovery does not recover from WALs, no need
+         // to check WALs.
+    } else if (!versions_->GetWalSet().GetWals().empty()) {
+      // Tracking is disabled, clear previously tracked WALs from MANIFEST,
+      // otherwise, in the future, if WAL tracking is enabled again,
+      // since the WALs deleted when WAL tracking is disabled are not persisted
+      // into MANIFEST, WAL check may fail.
+      VersionEdit edit;
+      WalNumber max_wal_number =
+          versions_->GetWalSet().GetWals().rbegin()->first;
+      edit.DeleteWalsBefore(max_wal_number + 1);
+      s = versions_->LogAndApplyToDefaultColumnFamily(&edit, &mutex_);
+    }
+    if (!s.ok()) {
+      return s;
+    }
+
+    if (!wal_files.empty()) {
+      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) {
-        for (auto& log : logs) {
-          std::string fname = LogFileName(immutable_db_options_.wal_dir, log);
+            "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& wal_file : wal_files) {
           uint64_t bytes;
-          s = env_->GetFileSize(fname, &bytes);
+          s = env_->GetFileSize(wal_file.second, &bytes);
           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.");
             }
           }
         }
       }
     }
 
-    if (!logs.empty()) {
-      // Recover in the order in which the logs were generated
-      std::sort(logs.begin(), logs.end());
-      bool corrupted_log_found = false;
-      s = RecoverLogFiles(logs, &next_sequence, read_only,
-                          &corrupted_log_found);
-      if (corrupted_log_found && recovered_seq != nullptr) {
+    if (!wal_files.empty()) {
+      // Recover in the order in which the wals were generated
+      std::vector<uint64_t> wals;
+      wals.reserve(wal_files.size());
+      for (const auto& wal_file : wal_files) {
+        wals.push_back(wal_file.first);
+      }
+      std::sort(wals.begin(), wals.end());
+
+      bool corrupted_wal_found = false;
+      s = RecoverLogFiles(wals, &next_sequence, read_only,
+                          &corrupted_wal_found);
+      if (corrupted_wal_found && recovered_seq != nullptr) {
         *recovered_seq = next_sequence;
       }
       if (!s.ok()) {
@@ -675,41 +684,56 @@ 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;
@@ -742,10 +766,10 @@ Status DBImpl::InitPersistStatsColumnFamily() {
   return s;
 }
 
-// REQUIRES: log_numbers are sorted in ascending order
-Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
+// REQUIRES: wal_numbers are sorted in ascending order
+Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& wal_numbers,
                                SequenceNumber* next_sequence, bool read_only,
-                               bool* corrupted_log_found) {
+                               bool* corrupted_wal_found) {
   struct LogReporter : public log::Reader::Reporter {
     Env* env;
     Logger* info_log;
@@ -775,10 +799,10 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
     auto stream = event_logger_.Log();
     stream << "job" << job_id << "event"
            << "recovery_started";
-    stream << "log_files";
+    stream << "wal_files";
     stream.StartArray();
-    for (auto log_number : log_numbers) {
-      stream << log_number;
+    for (auto wal_number : wal_numbers) {
+      stream << wal_number;
     }
     stream.EndArray();
   }
@@ -801,25 +825,25 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
   bool stop_replay_by_wal_filter = false;
   bool stop_replay_for_corruption = false;
   bool flushed = false;
-  uint64_t corrupted_log_number = kMaxSequenceNumber;
-  uint64_t min_log_number = MinLogNumberToKeep();
-  for (auto log_number : log_numbers) {
-    if (log_number < min_log_number) {
+  uint64_t corrupted_wal_number = kMaxSequenceNumber;
+  uint64_t min_wal_number = MinLogNumberToKeep();
+  for (auto wal_number : wal_numbers) {
+    if (wal_number < min_wal_number) {
       ROCKS_LOG_INFO(immutable_db_options_.info_log,
                      "Skipping log #%" PRIu64
                      " since it is older than min log to keep #%" PRIu64,
-                     log_number, min_log_number);
+                     wal_number, min_wal_number);
       continue;
     }
     // The previous incarnation may not have written any MANIFEST
     // records after allocating this log number.  So we manually
     // update the file number allocation counter in VersionSet.
-    versions_->MarkFileNumberUsed(log_number);
+    versions_->MarkFileNumberUsed(wal_number);
     // Open the log file
-    std::string fname = LogFileName(immutable_db_options_.wal_dir, log_number);
+    std::string fname = LogFileName(immutable_db_options_.wal_dir, wal_number);
 
     ROCKS_LOG_INFO(immutable_db_options_.info_log,
-                   "Recovering log #%" PRIu64 " mode %d", log_number,
+                   "Recovering log #%" PRIu64 " mode %d", wal_number,
                    static_cast<int>(immutable_db_options_.wal_recovery_mode));
     auto logFileDropped = [this, &fname]() {
       uint64_t bytes;
@@ -851,7 +875,8 @@ 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.
@@ -871,7 +896,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
     // to be skipped instead of propagating bad information (like overly
     // large sequence numbers).
     log::Reader reader(immutable_db_options_.info_log, std::move(file_reader),
-                       &reporter, true /*checksum*/, log_number);
+                       &reporter, true /*checksum*/, wal_number);
 
     // Determine if we should tolerate incomplete records at the tail end of the
     // Read all the records and add to a memtable
@@ -890,7 +915,11 @@ 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 ==
@@ -915,7 +944,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
 
         WalFilter::WalProcessingOption wal_processing_option =
             immutable_db_options_.wal_filter->LogRecordFound(
-                log_number, fname, batch, &new_batch, &batch_changed);
+                wal_number, fname, batch, &new_batch, &batch_changed);
 
         switch (wal_processing_option) {
           case WalFilter::WalProcessingOption::kContinueProcessing:
@@ -967,7 +996,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
                 " mode %d log filter %s returned "
                 "more records (%d) than original (%d) which is not allowed. "
                 "Aborting recovery.",
-                log_number,
+                wal_number,
                 static_cast<int>(immutable_db_options_.wal_recovery_mode),
                 immutable_db_options_.wal_filter->Name(), new_count,
                 original_count);
@@ -994,7 +1023,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
       bool has_valid_writes = false;
       status = WriteBatchInternal::InsertInto(
           &batch, column_family_memtables_.get(), &flush_scheduler_,
-          &trim_history_scheduler_, true, log_number, this,
+          &trim_history_scheduler_, true, wal_number, this,
           false /* concurrent_memtable_writes */, next_sequence,
           &has_valid_writes, seq_per_batch_, batch_per_txn_);
       MaybeIgnoreError(&status);
@@ -1014,7 +1043,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
           cfd->UnrefAndTryDelete();
           // If this asserts, it means that InsertInto failed in
           // filtering updates to already-flushed column families
-          assert(cfd->GetLogNumber() <= log_number);
+          assert(cfd->GetLogNumber() <= wal_number);
           auto iter = version_edits.find(cfd->GetID());
           assert(iter != version_edits.end());
           VersionEdit* edit = &iter->second;
@@ -1051,21 +1080,21 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
                           " seq #%" PRIu64
                           ". %s. This likely mean loss of synced WAL, "
                           "thus recovery fails.",
-                          log_number, *next_sequence,
+                          wal_number, *next_sequence,
                           status.ToString().c_str());
           return status;
         }
         // We should ignore the error but not continue replaying
         status = Status::OK();
         stop_replay_for_corruption = true;
-        corrupted_log_number = log_number;
-        if (corrupted_log_found != nullptr) {
-          *corrupted_log_found = true;
+        corrupted_wal_number = wal_number;
+        if (corrupted_wal_found != nullptr) {
+          *corrupted_wal_found = true;
         }
         ROCKS_LOG_INFO(immutable_db_options_.info_log,
                        "Point in time recovered to log #%" PRIu64
                        " seq #%" PRIu64,
-                       log_number, *next_sequence);
+                       wal_number, *next_sequence);
       } else {
         assert(immutable_db_options_.wal_recovery_mode ==
                    WALRecoveryMode::kTolerateCorruptedTailRecords ||
@@ -1091,7 +1120,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
   // corruption. This could during PIT recovery when the WAL is corrupted and
   // some (but not all) CFs are flushed
   // Exclude the PIT case where no log is dropped after the corruption point.
-  // This is to cover the case for empty logs after corrupted log, in which we
+  // This is to cover the case for empty wals after corrupted log, in which we
   // don't reset stop_replay_for_corruption.
   if (stop_replay_for_corruption == true &&
       (immutable_db_options_.wal_recovery_mode ==
@@ -1099,7 +1128,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
        immutable_db_options_.wal_recovery_mode ==
            WALRecoveryMode::kTolerateCorruptedTailRecords)) {
     for (auto cfd : *versions_->GetColumnFamilySet()) {
-      if (cfd->GetLogNumber() > corrupted_log_number) {
+      if (cfd->GetLogNumber() > corrupted_wal_number) {
         ROCKS_LOG_ERROR(immutable_db_options_.info_log,
                         "Column family inconsistency: SST file contains data"
                         " beyond the point of corruption.");
@@ -1114,16 +1143,16 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
   if (!read_only) {
     // no need to refcount since client still doesn't have access
     // to the DB and can not drop column families while we iterate
-    auto max_log_number = log_numbers.back();
+    const WalNumber max_wal_number = wal_numbers.back();
     for (auto cfd : *versions_->GetColumnFamilySet()) {
       auto iter = version_edits.find(cfd->GetID());
       assert(iter != version_edits.end());
       VersionEdit* edit = &iter->second;
 
-      if (cfd->GetLogNumber() > max_log_number) {
+      if (cfd->GetLogNumber() > max_wal_number) {
         // Column family cfd has already flushed the data
-        // from all logs. Memtable has to be empty because
-        // we filter the updates based on log_number
+        // from all wals. Memtable has to be empty because
+        // we filter the updates based on wal_number
         // (in WriteBatch::InsertInto)
         assert(cfd->mem()->GetFirstSequenceNumber() == 0);
         assert(edit->NumEntries() == 0);
@@ -1155,13 +1184,13 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
       // Update the log number info in the version edit corresponding to this
       // column family. Note that the version edits will be written to MANIFEST
       // together later.
-      // writing log_number in the manifest means that any log file
-      // with number strongly less than (log_number + 1) is already
+      // writing wal_number in the manifest means that any log file
+      // with number strongly less than (wal_number + 1) is already
       // recovered and should be ignored on next reincarnation.
-      // Since we already recovered max_log_number, we want all logs
-      // with numbers `<= max_log_number` (includes this one) to be ignored
+      // Since we already recovered max_wal_number, we want all wals
+      // with numbers `<= max_wal_number` (includes this one) to be ignored
       if (flushed || cfd->mem()->GetFirstSequenceNumber() == 0) {
-        edit->SetLogNumber(max_log_number + 1);
+        edit->SetLogNumber(max_wal_number + 1);
       }
     }
     if (status.ok()) {
@@ -1169,7 +1198,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
       // not actually used. that is because VersionSet assumes
       // VersionSet::next_file_number_ always to be strictly greater than any
       // log number
-      versions_->MarkFileNumberUsed(max_log_number + 1);
+      versions_->MarkFileNumberUsed(max_wal_number + 1);
 
       autovector<ColumnFamilyData*> cfds;
       autovector<const MutableCFOptions*> cf_opts;
@@ -1181,6 +1210,14 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
         assert(iter != version_edits.end());
         edit_lists.push_back({&iter->second});
       }
+
+      std::unique_ptr<VersionEdit> wal_deletion;
+      if (immutable_db_options_.track_and_verify_wals_in_manifest) {
+        wal_deletion.reset(new VersionEdit);
+        wal_deletion->DeleteWalsBefore(max_wal_number + 1);
+        edit_lists.back().push_back(wal_deletion.get());
+      }
+
       // write MANIFEST with update
       status = versions_->LogAndApply(cfds, cf_opts, edit_lists, &mutex_,
                                       directories_.GetDbDir(),
@@ -1189,7 +1226,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
   }
 
   if (status.ok() && data_seen && !flushed) {
-    status = RestoreAliveLogFiles(log_numbers);
+    status = RestoreAliveLogFiles(wal_numbers);
   }
 
   event_logger_.Log() << "job" << job_id << "event"
@@ -1198,8 +1235,8 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& log_numbers,
   return status;
 }
 
-Status DBImpl::RestoreAliveLogFiles(const std::vector<uint64_t>& log_numbers) {
-  if (log_numbers.empty()) {
+Status DBImpl::RestoreAliveLogFiles(const std::vector<uint64_t>& wal_numbers) {
+  if (wal_numbers.empty()) {
     return Status::OK();
   }
   Status s;
@@ -1212,20 +1249,20 @@ Status DBImpl::RestoreAliveLogFiles(const std::vector<uint64_t>& log_numbers) {
   // FindObsoleteFiles()
   total_log_size_ = 0;
   log_empty_ = false;
-  for (auto log_number : log_numbers) {
-    LogFileNumberSize log(log_number);
-    std::string fname = LogFileName(immutable_db_options_.wal_dir, log_number);
-    // This gets the appear size of the logs, not including preallocated space.
+  for (auto wal_number : wal_numbers) {
+    LogFileNumberSize log(wal_number);
+    std::string fname = LogFileName(immutable_db_options_.wal_dir, wal_number);
+    // This gets the appear size of the wals, not including preallocated space.
     s = env_->GetFileSize(fname, &log.size);
     if (!s.ok()) {
       break;
     }
     total_log_size_ += log.size;
     alive_log_files_.push_back(log);
-    // We preallocate space for logs, but then after a crash and restart, those
+    // We preallocate space for wals, but then after a crash and restart, those
     // preallocated space are not needed anymore. It is likely only the last
     // log has such preallocated space, so we only truncate for the last log.
-    if (log_number == log_numbers.back()) {
+    if (wal_number == wal_numbers.back()) {
       std::unique_ptr<FSWritableFile> last_log;
       Status truncate_status = fs_->ReopenWritableFile(
           fname,
@@ -1242,7 +1279,7 @@ Status DBImpl::RestoreAliveLogFiles(const std::vector<uint64_t>& log_numbers) {
       // Not a critical error if fail to truncate.
       if (!truncate_status.ok()) {
         ROCKS_LOG_WARN(immutable_db_options_.info_log,
-                       "Failed to truncate log #%" PRIu64 ": %s", log_number,
+                       "Failed to truncate log #%" PRIu64 ": %s", wal_number,
                        truncate_status.ToString().c_str());
       }
     }
@@ -1257,7 +1294,10 @@ 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()));
@@ -1281,7 +1321,8 @@ 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;
 
@@ -1303,20 +1344,23 @@ 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(), immutable_db_options_, *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]"
@@ -1330,23 +1374,39 @@ 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;
+
+  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();
+    stats.num_output_files = 1;
+  }
+
+  const auto& blobs = edit->GetBlobFileAdditions();
+  for (const auto& blob : blobs) {
+    stats.bytes_written += blob.GetTotalBlobBytes();
+  }
+
+  stats.num_output_files += static_cast<int>(blobs.size());
+
   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;
 }
@@ -1420,9 +1480,9 @@ IOStatus DBImpl::CreateWAL(uint64_t log_file_num, uint64_t recycle_log_number,
     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);
@@ -1434,7 +1494,7 @@ 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;
   }
@@ -1528,7 +1588,7 @@ 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;
           }
         }
@@ -1557,7 +1617,7 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
       // In WritePrepared there could be gap in sequence numbers. This breaks
       // the trick we use in kPointInTimeRecovery which assumes the first seq in
       // the log right after the corrupted log is one larger than the last seq
-      // we read from the logs. To let this trick keep working, we add a dummy
+      // we read from the wals. To let this trick keep working, we add a dummy
       // entry with the expected sequence to the first log right after recovery.
       // In non-WritePrepared case also the new log after recovery could be
       // empty, and thus missing the consecutive seq hint to distinguish
@@ -1581,7 +1641,7 @@ 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();
   }
 
@@ -1625,6 +1685,8 @@ 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();
 
@@ -1673,20 +1735,24 @@ 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();
           }
         }
       }
@@ -1700,6 +1766,7 @@ 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()) {
@@ -1714,11 +1781,14 @@ 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

@@ -86,10 +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;
 }
@@ -118,10 +118,10 @@ Status DBImplReadOnly::NewIterators(
         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);
   }
@@ -151,7 +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()) {
@@ -188,7 +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()) {
@@ -197,14 +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();
 
@@ -212,14 +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_));
@@ -253,7 +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.");
 }
 
@@ -261,7 +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);
@@ -112,7 +112,7 @@ Status DBImplSecondary::FindNewLogNumbers(std::vector<uint64_t>* logs) {
   for (size_t i = 0; i < filenames.size(); i++) {
     uint64_t number;
     FileType type;
-    if (ParseFileName(filenames[i], &number, &type) && type == kLogFile &&
+    if (ParseFileName(filenames[i], &number, &type) && type == kWalFile &&
         number >= log_number_min) {
       logs->push_back(number);
     }
@@ -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;
     }
@@ -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_));

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

@@ -77,7 +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) {
@@ -101,12 +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;
     }
   }
 
@@ -126,13 +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;
@@ -164,6 +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
 
@@ -204,6 +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
@@ -252,6 +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
@@ -423,7 +426,11 @@ Status DBImpl::WriteImpl(const WriteOptions& write_options,
 
   if (need_log_sync) {
     mutex_.Lock();
-    MarkLogsSynced(logfile_number_, need_log_dir_sync, status);
+    if (status.ok()) {
+      status = MarkLogsSynced(logfile_number_, need_log_dir_sync);
+    } else {
+      MarkLogsNotSynced(logfile_number_);
+    }
     mutex_.Unlock();
     // Requesting sync with two_write_queues_ is expected to be very rare. We
     // hence provide a simple implementation that is not necessarily efficient.
@@ -470,6 +477,7 @@ 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()) {
@@ -547,7 +555,11 @@ Status DBImpl::PipelinedWriteImpl(const WriteOptions& write_options,
 
     if (need_log_sync) {
       mutex_.Lock();
-      MarkLogsSynced(logfile_number_, need_log_dir_sync, w.status);
+      if (w.status.ok()) {
+        w.status = MarkLogsSynced(logfile_number_, need_log_dir_sync);
+      } else {
+        MarkLogsNotSynced(logfile_number_);
+      }
       mutex_.Unlock();
     }
 
@@ -623,8 +635,6 @@ 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);
     }
@@ -639,6 +649,7 @@ 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();
@@ -689,7 +700,7 @@ 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;
@@ -830,13 +841,30 @@ 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();
   }
 }
@@ -844,10 +872,13 @@ 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();
   }
 }
@@ -861,7 +892,9 @@ 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();
   }
 }
@@ -975,8 +1008,10 @@ 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;
@@ -1546,11 +1581,14 @@ 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_);
@@ -1752,10 +1790,15 @@ 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
+    // Should handle return error?
     if (!io_s.ok()) {
-      error_handler_.SetBGError(io_s, BackgroundErrorReason::kMemTable);
+      // Should handle return error?
+      error_handler_.SetBGError(io_s, BackgroundErrorReason::kMemTable)
+          .PermitUncheckedError();
     } else {
-      error_handler_.SetBGError(s, BackgroundErrorReason::kMemTable);
+      // 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();
@@ -1789,6 +1832,10 @@ 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;
 }
 
@@ -1879,7 +1926,10 @@ 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);
 }
 
@@ -1887,7 +1937,10 @@ 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

@@ -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_;
@@ -97,14 +97,14 @@ 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) {
     uint64_t number;
     FileType type;
     if (ParseFileName(file, &number, &type)) {
-      log_cnt += (type == kLogFile);
+      log_cnt += (type == kWalFile);
       sst_cnt += (type == kTableFile);
       manifest_cnt += (type == kDescriptorFile);
     }
@@ -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) {
@@ -880,6 +883,7 @@ TEST_F(DBSecondaryTest, StartFromInconsistent) {
       });
   SyncPoint::GetInstance()->EnableProcessing();
   Options options1;
+  options1.env = env_;
   Status s = TryOpenSecondary(options1);
   ASSERT_TRUE(s.IsCorruption());
 }
@@ -891,6 +895,7 @@ TEST_F(DBSecondaryTest, InconsistencyDuringCatchUp) {
   ASSERT_OK(Flush());
 
   Options options1;
+  options1.env = env_;
   OpenSecondary(options1);
 
   {

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(" ; ");
+      case kWalFile:
+        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) {
@@ -110,12 +118,15 @@ void DumpDBFileSummary(const ImmutableDBOptions& options,
     wal_info.clear();
     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 (type == kWalFile) {
+          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) {

db/db_io_failure_test.cc

@@ -15,7 +15,8 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBIOFailureTest : public DBTestBase {
  public:
-  DBIOFailureTest() : DBTestBase("/db_io_failure_test") {}
+  DBIOFailureTest()
+      : DBTestBase("/db_io_failure_test", /*env_do_fsync=*/true) {}
 };
 
 #ifndef ROCKSDB_LITE
@@ -34,7 +35,7 @@ TEST_F(DBIOFailureTest, DropWrites) {
     // Force out-of-space errors
     env_->drop_writes_.store(true, std::memory_order_release);
     env_->sleep_counter_.Reset();
-    env_->no_slowdown_ = true;
+    env_->SetMockSleep();
     for (int i = 0; i < 5; i++) {
       if (option_config_ != kUniversalCompactionMultiLevel &&
           option_config_ != kUniversalSubcompactions) {

db/db_iter.cc

@@ -107,11 +107,12 @@ Status DBIter::GetProperty(std::string prop_name, std::string* prop) {
 }
 
 bool DBIter::ParseKey(ParsedInternalKey* ikey) {
-  if (!ParseInternalKey(iter_.key(), ikey)) {
-    status_ = Status::Corruption("corrupted internal key in DBIter");
+  Status s =
+      ParseInternalKey(iter_.key(), ikey, false /* log_err_key */);  // TODO
+  if (!s.ok()) {
+    status_ = Status::Corruption("In DBIter: ", s.getState());
     valid_ = false;
-    ROCKS_LOG_ERROR(logger_, "corrupted internal key in DBIter: %s",
-                    iter_.key().ToString(true).c_str());
+    ROCKS_LOG_ERROR(logger_, "In DBIter: %s", status_.getState());
     return false;
   } else {
     return true;
@@ -147,13 +148,13 @@ void DBIter::Next() {
 
   local_stats_.next_count_++;
   if (ok && iter_.Valid()) {
-    Slice prefix;
     if (prefix_same_as_start_) {
       assert(prefix_extractor_ != nullptr);
-      prefix = prefix_.GetUserKey();
+      const Slice prefix = prefix_.GetUserKey();
+      FindNextUserEntry(true /* skipping the current user key */, &prefix);
+    } else {
+      FindNextUserEntry(true /* skipping the current user key */, nullptr);
     }
-    FindNextUserEntry(true /* skipping the current user key */,
-                      prefix_same_as_start_ ? &prefix : nullptr);
   } else {
     is_key_seqnum_zero_ = false;
     valid_ = false;
@@ -223,11 +224,13 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
 
     is_key_seqnum_zero_ = (ikey_.sequence == 0);
 
-    assert(iterate_upper_bound_ == nullptr || iter_.MayBeOutOfUpperBound() ||
+    assert(iterate_upper_bound_ == nullptr ||
+           iter_.UpperBoundCheckResult() != IterBoundCheck::kInbound ||
            user_comparator_.CompareWithoutTimestamp(
                ikey_.user_key, /*a_has_ts=*/true, *iterate_upper_bound_,
                /*b_has_ts=*/false) < 0);
-    if (iterate_upper_bound_ != nullptr && iter_.MayBeOutOfUpperBound() &&
+    if (iterate_upper_bound_ != nullptr &&
+        iter_.UpperBoundCheckResult() != IterBoundCheck::kInbound &&
         user_comparator_.CompareWithoutTimestamp(
             ikey_.user_key, /*a_has_ts=*/true, *iterate_upper_bound_,
             /*b_has_ts=*/false) >= 0) {
@@ -246,11 +249,10 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
     }
 
     assert(ikey_.user_key.size() >= timestamp_size_);
-    Slice ts;
+    Slice ts = timestamp_size_ > 0 ? ExtractTimestampFromUserKey(
+                                         ikey_.user_key, timestamp_size_)
+                                   : Slice();
     bool more_recent = false;
-    if (timestamp_size_ > 0) {
-      ts = ExtractTimestampFromUserKey(ikey_.user_key, timestamp_size_);
-    }
     if (IsVisible(ikey_.sequence, ts, &more_recent)) {
       // If the previous entry is of seqnum 0, the current entry will not
       // possibly be skipped. This condition can potentially be relaxed to
@@ -285,7 +287,20 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
             // 2) return ikey only if ikey.seqnum >= start_seqnum_
             // note that if deletion seqnum is < start_seqnum_ we
             // just skip it like in normal iterator.
-            if (start_seqnum_ > 0 && ikey_.sequence >= start_seqnum_)  {
+            if (start_seqnum_ > 0) {
+              if (ikey_.sequence >= start_seqnum_) {
+                saved_key_.SetInternalKey(ikey_);
+                valid_ = true;
+                return true;
+              } else {
+                saved_key_.SetUserKey(
+                    ikey_.user_key,
+                    !pin_thru_lifetime_ ||
+                        !iter_.iter()->IsKeyPinned() /* copy */);
+                skipping_saved_key = true;
+                PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
+              }
+            } else if (timestamp_lb_) {
               saved_key_.SetInternalKey(ikey_);
               valid_ = true;
               return true;
@@ -300,10 +315,8 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
           case kTypeValue:
           case kTypeBlobIndex:
             if (start_seqnum_ > 0) {
-              // we are taking incremental snapshot here
-              // incremental snapshots aren't supported on DB with range deletes
-              assert(ikey_.type != kTypeBlobIndex);
               if (ikey_.sequence >= start_seqnum_) {
+                assert(ikey_.type != kTypeBlobIndex);
                 saved_key_.SetInternalKey(ikey_);
                 valid_ = true;
                 return true;
@@ -316,6 +329,10 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
                         !iter_.iter()->IsKeyPinned() /* copy */);
                 skipping_saved_key = true;
               }
+            } else if (timestamp_lb_) {
+              saved_key_.SetInternalKey(ikey_);
+              valid_ = true;
+              return true;
             } else {
               saved_key_.SetUserKey(
                   ikey_.user_key, !pin_thru_lifetime_ ||
@@ -368,8 +385,11 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
             }
             break;
           default:
-            assert(false);
-            break;
+            valid_ = false;
+            status_ = Status::Corruption(
+                "Unknown value type: " +
+                std::to_string(static_cast<unsigned int>(ikey_.type)));
+            return false;
         }
       }
     } else {
@@ -417,11 +437,11 @@ bool DBIter::FindNextUserEntryInternal(bool skipping_saved_key,
               &last_key,
               ParsedInternalKey(saved_key_.GetUserKey(), 0, kTypeDeletion));
         } else {
-          std::string min_ts(timestamp_size_, static_cast<char>(0));
+          const std::string kTsMin(timestamp_size_, static_cast<char>(0));
           AppendInternalKeyWithDifferentTimestamp(
               &last_key,
               ParsedInternalKey(saved_key_.GetUserKey(), 0, kTypeDeletion),
-              min_ts);
+              kTsMin);
         }
         // Don't set skipping_saved_key = false because we may still see more
         // user-keys equal to saved_key_.
@@ -540,7 +560,11 @@ bool DBIter::MergeValuesNewToOld() {
       valid_ = false;
       return false;
     } else {
-      assert(false);
+      valid_ = false;
+      status_ = Status::Corruption(
+          "Unrecognized value type: " +
+          std::to_string(static_cast<unsigned int>(ikey.type)));
+      return false;
     }
   }
 
@@ -787,12 +811,19 @@ bool DBIter::FindValueForCurrentKey() {
                 ikey, RangeDelPositioningMode::kBackwardTraversal)) {
           last_key_entry_type = kTypeRangeDeletion;
           PERF_COUNTER_ADD(internal_delete_skipped_count, 1);
-        } else {
-          assert(iter_.iter()->IsValuePinned());
+        } else if (iter_.iter()->IsValuePinned()) {
           pinned_value_ = iter_.value();
+        } else {
+          valid_ = false;
+          status_ = Status::NotSupported(
+              "Backward iteration not supported if underlying iterator's value "
+              "cannot be pinned.");
         }
         merge_context_.Clear();
         last_not_merge_type = last_key_entry_type;
+        if (!status_.ok()) {
+          return false;
+        }
         break;
       case kTypeDeletion:
       case kTypeSingleDeletion:
@@ -816,7 +847,11 @@ bool DBIter::FindValueForCurrentKey() {
         }
         break;
       default:
-        assert(false);
+        valid_ = false;
+        status_ = Status::Corruption(
+            "Unknown value type: " +
+            std::to_string(static_cast<unsigned int>(last_key_entry_type)));
+        return false;
     }
 
     PERF_COUNTER_ADD(internal_key_skipped_count, 1);
@@ -830,6 +865,7 @@ bool DBIter::FindValueForCurrentKey() {
   }
 
   Status s;
+  s.PermitUncheckedError();
   is_blob_ = false;
   switch (last_key_entry_type) {
     case kTypeDeletion:
@@ -881,8 +917,11 @@ bool DBIter::FindValueForCurrentKey() {
       is_blob_ = true;
       break;
     default:
-      assert(false);
-      break;
+      valid_ = false;
+      status_ = Status::Corruption(
+          "Unknown value type: " +
+          std::to_string(static_cast<unsigned int>(last_key_entry_type)));
+      return false;
   }
   if (!s.ok()) {
     valid_ = false;
@@ -1031,7 +1070,11 @@ bool DBIter::FindValueForCurrentKeyUsingSeek() {
       valid_ = false;
       return false;
     } else {
-      assert(false);
+      valid_ = false;
+      status_ = Status::Corruption(
+          "Unknown value type: " +
+          std::to_string(static_cast<unsigned int>(ikey.type)));
+      return false;
     }
   }
 
@@ -1193,7 +1236,8 @@ void DBIter::Seek(const Slice& target) {
 
 #ifndef ROCKSDB_LITE
   if (db_impl_ != nullptr && cfd_ != nullptr) {
-    db_impl_->TraceIteratorSeek(cfd_->GetID(), target);
+    // TODO: What do we do if this returns an error?
+    db_impl_->TraceIteratorSeek(cfd_->GetID(), target).PermitUncheckedError();
   }
 #endif  // ROCKSDB_LITE
 
@@ -1254,7 +1298,9 @@ void DBIter::SeekForPrev(const Slice& target) {
 
 #ifndef ROCKSDB_LITE
   if (db_impl_ != nullptr && cfd_ != nullptr) {
-    db_impl_->TraceIteratorSeekForPrev(cfd_->GetID(), target);
+    // TODO: What do we do if this returns an error?
+    db_impl_->TraceIteratorSeekForPrev(cfd_->GetID(), target)
+        .PermitUncheckedError();
   }
 #endif  // ROCKSDB_LITE
 

db/db_iter.h

@@ -140,10 +140,17 @@ class DBIter final : public Iterator {
   }
   ReadRangeDelAggregator* GetRangeDelAggregator() { return &range_del_agg_; }
 
-  bool Valid() const override { return valid_; }
+  bool Valid() const override {
+#ifdef ROCKSDB_ASSERT_STATUS_CHECKED
+    if (valid_) {
+      status_.PermitUncheckedError();
+    }
+#endif  // ROCKSDB_ASSERT_STATUS_CHECKED
+    return valid_;
+  }
   Slice key() const override {
     assert(valid_);
-    if (start_seqnum_ > 0) {
+    if (start_seqnum_ > 0 || timestamp_lb_) {
       return saved_key_.GetInternalKey();
     } else {
       const Slice ukey_and_ts = saved_key_.GetUserKey();

db/db_iter_test.cc

@@ -99,9 +99,11 @@ class TestIterator : public InternalIterator {
     }
     for (auto it = data_.begin(); it != data_.end(); ++it) {
       ParsedInternalKey ikey;
-      bool ok __attribute__((__unused__)) = ParseInternalKey(it->first, &ikey);
-      assert(ok);
-      if (ikey.user_key != _key) {
+      Status pik_status =
+          ParseInternalKey(it->first, &ikey, true /* log_err_key */);
+      pik_status.PermitUncheckedError();
+      assert(pik_status.ok());
+      if (!pik_status.ok() || ikey.user_key != _key) {
         continue;
       }
       if (valid_ && data_.begin() + iter_ > it) {

db/db_iterator_test.cc

@@ -34,7 +34,7 @@ class DummyReadCallback : public ReadCallback {
 class DBIteratorTest : public DBTestBase,
                        public testing::WithParamInterface<bool> {
  public:
-  DBIteratorTest() : DBTestBase("/db_iterator_test") {}
+  DBIteratorTest() : DBTestBase("/db_iterator_test", /*env_do_fsync=*/true) {}
 
   Iterator* NewIterator(const ReadOptions& read_options,
                         ColumnFamilyHandle* column_family = nullptr) {
@@ -2146,7 +2146,7 @@ TEST_P(DBIteratorTest, ReadAhead) {
   BlockBasedTableOptions table_options;
   table_options.block_size = 1024;
   table_options.no_block_cache = true;
-  options.table_factory.reset(new BlockBasedTableFactory(table_options));
+  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
   Reopen(options);
 
   std::string value(1024, 'a');
@@ -3021,6 +3021,44 @@ TEST_F(DBIteratorWithReadCallbackTest, ReadCallback) {
   delete iter;
 }
 
+TEST_F(DBIteratorTest, BackwardIterationOnInplaceUpdateMemtable) {
+  Options options = CurrentOptions();
+  options.create_if_missing = true;
+  options.inplace_update_support = false;
+  options.env = env_;
+  DestroyAndReopen(options);
+  constexpr int kNumKeys = 10;
+
+  // Write kNumKeys to WAL.
+  for (int i = 0; i < kNumKeys; ++i) {
+    ASSERT_OK(Put(Key(i), "val"));
+  }
+  ReadOptions read_opts;
+  read_opts.total_order_seek = true;
+  {
+    std::unique_ptr<Iterator> iter(db_->NewIterator(read_opts));
+    int count = 0;
+    for (iter->SeekToLast(); iter->Valid(); iter->Prev()) {
+      ++count;
+    }
+    ASSERT_EQ(kNumKeys, count);
+  }
+
+  // Reopen and rebuild the memtable from WAL.
+  options.create_if_missing = false;
+  options.avoid_flush_during_recovery = true;
+  options.inplace_update_support = true;
+  options.allow_concurrent_memtable_write = false;
+  Reopen(options);
+  {
+    std::unique_ptr<Iterator> iter(db_->NewIterator(read_opts));
+    iter->SeekToLast();
+    // Backward iteration not supported due to inplace_update_support = true.
+    ASSERT_TRUE(iter->status().IsNotSupported());
+    ASSERT_FALSE(iter->Valid());
+  }
+}
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {

db/db_log_iter_test.cc

@@ -19,7 +19,8 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBTestXactLogIterator : public DBTestBase {
  public:
-  DBTestXactLogIterator() : DBTestBase("/db_log_iter_test") {}
+  DBTestXactLogIterator()
+      : DBTestBase("/db_log_iter_test", /*env_do_fsync=*/true) {}
 
   std::unique_ptr<TransactionLogIterator> OpenTransactionLogIter(
       const SequenceNumber seq) {

db/db_memtable_test.cc

@@ -17,7 +17,7 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBMemTableTest : public DBTestBase {
  public:
-  DBMemTableTest() : DBTestBase("/db_memtable_test") {}
+  DBMemTableTest() : DBTestBase("/db_memtable_test", /*env_do_fsync=*/true) {}
 };
 
 class MockMemTableRep : public MemTableRep {
@@ -316,6 +316,7 @@ TEST_F(DBMemTableTest, InsertWithHint) {
 TEST_F(DBMemTableTest, ColumnFamilyId) {
   // Verifies MemTableRepFactory is told the right column family id.
   Options options;
+  options.env = CurrentOptions().env;
   options.allow_concurrent_memtable_write = false;
   options.create_if_missing = true;
   options.memtable_factory.reset(new MockMemTableRepFactory());

db/db_merge_operand_test.cc

@@ -21,7 +21,8 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBMergeOperandTest : public DBTestBase {
  public:
-  DBMergeOperandTest() : DBTestBase("/db_merge_operand_test") {}
+  DBMergeOperandTest()
+      : DBTestBase("/db_merge_operand_test", /*env_do_fsync=*/true) {}
 };
 
 TEST_F(DBMergeOperandTest, GetMergeOperandsBasic) {

db/db_merge_operator_test.cc

@@ -36,7 +36,8 @@ class TestReadCallback : public ReadCallback {
 // Test merge operator functionality.
 class DBMergeOperatorTest : public DBTestBase {
  public:
-  DBMergeOperatorTest() : DBTestBase("/db_merge_operator_test") {}
+  DBMergeOperatorTest()
+      : DBTestBase("/db_merge_operator_test", /*env_do_fsync=*/false) {}
 
   std::string GetWithReadCallback(SnapshotChecker* snapshot_checker,
                                   const Slice& key,

db/db_options_test.cc

@@ -27,23 +27,20 @@ namespace ROCKSDB_NAMESPACE {
 
 class DBOptionsTest : public DBTestBase {
  public:
-  DBOptionsTest() : DBTestBase("/db_options_test") {}
+  DBOptionsTest() : DBTestBase("/db_options_test", /*env_do_fsync=*/true) {}
 
 #ifndef ROCKSDB_LITE
   std::unordered_map<std::string, std::string> GetMutableDBOptionsMap(
       const DBOptions& options) {
     std::string options_str;
+    std::unordered_map<std::string, std::string> mutable_map;
     ConfigOptions config_options;
     config_options.delimiter = "; ";
-    GetStringFromDBOptions(config_options, options, &options_str);
-    std::unordered_map<std::string, std::string> options_map;
-    StringToMap(options_str, &options_map);
-    std::unordered_map<std::string, std::string> mutable_map;
-    for (const auto& opt : db_options_type_info) {
-      if (opt.second.IsMutable() && opt.second.ShouldSerialize()) {
-        mutable_map[opt.first] = options_map[opt.first];
-      }
-    }
+
+    EXPECT_OK(GetStringFromMutableDBOptions(
+        config_options, MutableDBOptions(options), &options_str));
+    EXPECT_OK(StringToMap(options_str, &mutable_map));
+
     return mutable_map;
   }
 
@@ -53,15 +50,10 @@ class DBOptionsTest : public DBTestBase {
     ConfigOptions config_options;
     config_options.delimiter = "; ";
 
-    GetStringFromColumnFamilyOptions(config_options, options, &options_str);
-    std::unordered_map<std::string, std::string> options_map;
-    StringToMap(options_str, &options_map);
     std::unordered_map<std::string, std::string> mutable_map;
-    for (const auto& opt : cf_options_type_info) {
-      if (opt.second.IsMutable() && opt.second.ShouldSerialize()) {
-        mutable_map[opt.first] = options_map[opt.first];
-      }
-    }
+    EXPECT_OK(GetStringFromMutableCFOptions(
+        config_options, MutableCFOptions(options), &options_str));
+    EXPECT_OK(StringToMap(options_str, &mutable_map));
     return mutable_map;
   }
 
@@ -87,6 +79,22 @@ class DBOptionsTest : public DBTestBase {
 #endif  // ROCKSDB_LITE
 };
 
+TEST_F(DBOptionsTest, ImmutableTrackAndVerifyWalsInManifest) {
+  Options options;
+  options.env = env_;
+  options.track_and_verify_wals_in_manifest = true;
+
+  ImmutableDBOptions db_options(options);
+  ASSERT_TRUE(db_options.track_and_verify_wals_in_manifest);
+
+  Reopen(options);
+  ASSERT_TRUE(dbfull()->GetDBOptions().track_and_verify_wals_in_manifest);
+
+  Status s =
+      dbfull()->SetDBOptions({{"track_and_verify_wals_in_manifest", "false"}});
+  ASSERT_FALSE(s.ok());
+}
+
 // RocksDB lite don't support dynamic options.
 #ifndef ROCKSDB_LITE
 
@@ -143,7 +151,7 @@ TEST_F(DBOptionsTest, SetBytesPerSync) {
   WriteOptions write_opts;
   // should sync approximately 40MB/1MB ~= 40 times.
   for (i = 0; i < 40; i++) {
-    Put(Key(i), kValue, write_opts);
+    ASSERT_OK(Put(Key(i), kValue, write_opts));
   }
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
   ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
@@ -159,7 +167,7 @@ TEST_F(DBOptionsTest, SetBytesPerSync) {
   // should sync approximately 40MB*2/8MB ~= 10 times.
   // data will be 40*2MB because of previous Puts too.
   for (i = 0; i < 40; i++) {
-    Put(Key(i), kValue, write_opts);
+    ASSERT_OK(Put(Key(i), kValue, write_opts));
   }
   ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
   ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
@@ -190,7 +198,7 @@ TEST_F(DBOptionsTest, SetWalBytesPerSync) {
   const std::string kValue(kValueSize, 'v');
   int i = 0;
   for (; i < 10; i++) {
-    Put(Key(i), kValue);
+    ASSERT_OK(Put(Key(i), kValue));
   }
   // Do not flush. If we flush here, SwitchWAL will reuse old WAL file since its
   // empty and will not get the new wal_bytes_per_sync value.
@@ -201,7 +209,7 @@ TEST_F(DBOptionsTest, SetWalBytesPerSync) {
   counter = 0;
   i = 0;
   for (; i < 10; i++) {
-    Put(Key(i), kValue);
+    ASSERT_OK(Put(Key(i), kValue));
   }
   ASSERT_GT(counter, 0);
   ASSERT_GT(low_bytes_per_sync, 0);
@@ -236,9 +244,9 @@ TEST_F(DBOptionsTest, WritableFileMaxBufferSize) {
   for (; i < 3; i++) {
     ASSERT_OK(Put("foo", ToString(i)));
     ASSERT_OK(Put("bar", ToString(i)));
-    Flush();
+    ASSERT_OK(Flush());
   }
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(unmatch_cnt, 0);
   ASSERT_GE(match_cnt, 11);
 
@@ -254,9 +262,9 @@ TEST_F(DBOptionsTest, WritableFileMaxBufferSize) {
   for (; i < 3; i++) {
     ASSERT_OK(Put("foo", ToString(i)));
     ASSERT_OK(Put("bar", ToString(i)));
-    Flush();
+    ASSERT_OK(Flush());
   }
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(unmatch_cnt, 0);
   ASSERT_GE(match_cnt, 11);
 }
@@ -292,14 +300,14 @@ TEST_F(DBOptionsTest, EnableAutoCompactionAndTriggerStall) {
       DestroyAndReopen(options);
       int i = 0;
       for (; i < 1024; i++) {
-        Put(Key(i), kValue);
+        ASSERT_OK(Put(Key(i), kValue));
       }
-      Flush();
+      ASSERT_OK(Flush());
       for (; i < 1024 * 2; i++) {
-        Put(Key(i), kValue);
+        ASSERT_OK(Put(Key(i), kValue));
       }
-      Flush();
-      dbfull()->TEST_WaitForFlushMemTable();
+      ASSERT_OK(Flush());
+      ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
       ASSERT_EQ(2, NumTableFilesAtLevel(0));
       uint64_t l0_size = SizeAtLevel(0);
 
@@ -321,7 +329,7 @@ TEST_F(DBOptionsTest, EnableAutoCompactionAndTriggerStall) {
           break;
       }
       Reopen(options);
-      dbfull()->TEST_WaitForCompact();
+      ASSERT_OK(dbfull()->TEST_WaitForCompact());
       ASSERT_FALSE(dbfull()->TEST_write_controler().IsStopped());
       ASSERT_FALSE(dbfull()->TEST_write_controler().NeedsDelay());
 
@@ -368,7 +376,7 @@ TEST_F(DBOptionsTest, EnableAutoCompactionAndTriggerStall) {
       TEST_SYNC_POINT("DBOptionsTest::EnableAutoCompactionAndTriggerStall:3");
 
       // Background compaction executed.
-      dbfull()->TEST_WaitForCompact();
+      ASSERT_OK(dbfull()->TEST_WaitForCompact());
       ASSERT_FALSE(dbfull()->TEST_write_controler().IsStopped());
       ASSERT_FALSE(dbfull()->TEST_write_controler().NeedsDelay());
     }
@@ -385,12 +393,12 @@ TEST_F(DBOptionsTest, SetOptionsMayTriggerCompaction) {
     // Need to insert two keys to avoid trivial move.
     ASSERT_OK(Put("foo", ToString(i)));
     ASSERT_OK(Put("bar", ToString(i)));
-    Flush();
+    ASSERT_OK(Flush());
   }
   ASSERT_EQ("3", FilesPerLevel());
   ASSERT_OK(
       dbfull()->SetOptions({{"level0_file_num_compaction_trigger", "3"}}));
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ("0,1", FilesPerLevel());
 }
 
@@ -512,7 +520,7 @@ TEST_F(DBOptionsTest, MaxTotalWalSizeChange) {
   ASSERT_OK(dbfull()->SetDBOptions({{"max_total_wal_size", "10"}}));
 
   for (size_t cf = 0; cf < handles_.size(); ++cf) {
-    dbfull()->TEST_WaitForFlushMemTable(handles_[cf]);
+    ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable(handles_[cf]));
     ASSERT_EQ("1", FilesPerLevel(static_cast<int>(cf)));
   }
 }
@@ -563,10 +571,9 @@ static void assert_candidate_files_empty(DBImpl* dbfull, const bool empty) {
 }
 
 TEST_F(DBOptionsTest, DeleteObsoleteFilesPeriodChange) {
-  SpecialEnv env(env_);
-  env.time_elapse_only_sleep_ = true;
   Options options;
-  options.env = &env;
+  options.env = env_;
+  SetTimeElapseOnlySleepOnReopen(&options);
   options.create_if_missing = true;
   ASSERT_OK(TryReopen(options));
 
@@ -585,10 +592,10 @@ TEST_F(DBOptionsTest, DeleteObsoleteFilesPeriodChange) {
 
   assert_candidate_files_empty(dbfull(), true);
 
-  env.addon_time_.store(20);
+  env_->MockSleepForMicroseconds(20);
   assert_candidate_files_empty(dbfull(), true);
 
-  env.addon_time_.store(21);
+  env_->MockSleepForMicroseconds(1);
   assert_candidate_files_empty(dbfull(), false);
 
   Close();
@@ -615,6 +622,7 @@ TEST_F(DBOptionsTest, MaxOpenFilesChange) {
 
 TEST_F(DBOptionsTest, SanitizeDelayedWriteRate) {
   Options options;
+  options.env = CurrentOptions().env;
   options.delayed_write_rate = 0;
   Reopen(options);
   ASSERT_EQ(16 * 1024 * 1024, dbfull()->GetDBOptions().delayed_write_rate);
@@ -626,6 +634,7 @@ TEST_F(DBOptionsTest, SanitizeDelayedWriteRate) {
 
 TEST_F(DBOptionsTest, SanitizeUniversalTTLCompaction) {
   Options options;
+  options.env = CurrentOptions().env;
   options.compaction_style = kCompactionStyleUniversal;
 
   options.ttl = 0;
@@ -655,6 +664,7 @@ TEST_F(DBOptionsTest, SanitizeUniversalTTLCompaction) {
 
 TEST_F(DBOptionsTest, SanitizeTtlDefault) {
   Options options;
+  options.env = CurrentOptions().env;
   Reopen(options);
   ASSERT_EQ(30 * 24 * 60 * 60, dbfull()->GetOptions().ttl);
 
@@ -671,6 +681,7 @@ TEST_F(DBOptionsTest, SanitizeTtlDefault) {
 TEST_F(DBOptionsTest, SanitizeFIFOPeriodicCompaction) {
   Options options;
   options.compaction_style = kCompactionStyleFIFO;
+  options.env = CurrentOptions().env;
   options.ttl = 0;
   Reopen(options);
   ASSERT_EQ(30 * 24 * 60 * 60, dbfull()->GetOptions().ttl);
@@ -696,17 +707,19 @@ TEST_F(DBOptionsTest, SanitizeFIFOPeriodicCompaction) {
 
 TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
   Options options;
+  options.env = CurrentOptions().env;
   options.compaction_style = kCompactionStyleFIFO;
   options.write_buffer_size = 10 << 10;  // 10KB
   options.arena_block_size = 4096;
   options.compression = kNoCompression;
   options.create_if_missing = true;
   options.compaction_options_fifo.allow_compaction = false;
-  env_->time_elapse_only_sleep_ = false;
+  env_->SetMockSleep();
   options.env = env_;
 
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
   // Test dynamically changing ttl.
-  env_->addon_time_.store(0);
   options.ttl = 1 * 60 * 60;  // 1 hour
   ASSERT_OK(TryReopen(options));
 
@@ -721,23 +734,23 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
 
-  // Add 61 seconds to the time.
-  env_->addon_time_.fetch_add(61);
+  env_->MockSleepForSeconds(61);
 
   // No files should be compacted as ttl is set to 1 hour.
   ASSERT_EQ(dbfull()->GetOptions().ttl, 3600);
-  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
 
   // Set ttl to 1 minute. So all files should get deleted.
   ASSERT_OK(dbfull()->SetOptions({{"ttl", "60"}}));
   ASSERT_EQ(dbfull()->GetOptions().ttl, 60);
-  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(NumTableFilesAtLevel(0), 0);
 
+  // NOTE: Presumed unnecessary and removed: resetting mock time in env
+
   // Test dynamically changing compaction_options_fifo.max_table_files_size
-  env_->addon_time_.store(0);
   options.compaction_options_fifo.max_table_files_size = 500 << 10;  // 00KB
   options.ttl = 0;
   DestroyAndReopen(options);
@@ -747,7 +760,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
     for (int j = 0; j < 10; j++) {
       ASSERT_OK(Put(ToString(i * 20 + j), rnd.RandomString(980)));
     }
-    Flush();
+    ASSERT_OK(Flush());
   }
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
@@ -755,7 +768,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
   // No files should be compacted as max_table_files_size is set to 500 KB.
   ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.max_table_files_size,
             500 << 10);
-  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
 
   // Set max_table_files_size to 12 KB. So only 1 file should remain now.
@@ -763,7 +776,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
       {{"compaction_options_fifo", "{max_table_files_size=12288;}"}}));
   ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.max_table_files_size,
             12 << 10);
-  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(NumTableFilesAtLevel(0), 1);
 
@@ -779,7 +792,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
     for (int j = 0; j < 10; j++) {
       ASSERT_OK(Put(ToString(i * 20 + j), rnd.RandomString(980)));
     }
-    Flush();
+    ASSERT_OK(Flush());
   }
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
@@ -788,7 +801,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
   // allow_compaction is false
   ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.allow_compaction,
             false);
-  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
 
   // Set allow_compaction to true. So number of files should be between 1 and 5.
@@ -796,7 +809,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
       {{"compaction_options_fifo", "{allow_compaction=true;}"}}));
   ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.allow_compaction,
             true);
-  dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr);
+  ASSERT_OK(dbfull()->CompactRange(CompactRangeOptions(), nullptr, nullptr));
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_GE(NumTableFilesAtLevel(0), 1);
   ASSERT_LE(NumTableFilesAtLevel(0), 5);
@@ -817,14 +830,14 @@ TEST_F(DBOptionsTest, CompactionReadaheadSizeChange) {
   ASSERT_OK(dbfull()->SetDBOptions({{"compaction_readahead_size", "256"}}));
   ASSERT_EQ(256, dbfull()->GetDBOptions().compaction_readahead_size);
   for (int i = 0; i < 1024; i++) {
-    Put(Key(i), kValue);
+    ASSERT_OK(Put(Key(i), kValue));
   }
-  Flush();
+  ASSERT_OK(Flush());
   for (int i = 0; i < 1024 * 2; i++) {
-    Put(Key(i), kValue);
+    ASSERT_OK(Put(Key(i), kValue));
   }
-  Flush();
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(Flush());
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(256, env_->compaction_readahead_size_);
   Close();
 }
@@ -834,6 +847,7 @@ TEST_F(DBOptionsTest, FIFOTtlBackwardCompatible) {
   options.compaction_style = kCompactionStyleFIFO;
   options.write_buffer_size = 10 << 10;  // 10KB
   options.create_if_missing = true;
+  options.env = CurrentOptions().env;
 
   ASSERT_OK(TryReopen(options));
 
@@ -843,7 +857,7 @@ TEST_F(DBOptionsTest, FIFOTtlBackwardCompatible) {
     for (int j = 0; j < 10; j++) {
       ASSERT_OK(Put(ToString(i * 20 + j), rnd.RandomString(980)));
     }
-    Flush();
+    ASSERT_OK(Flush());
   }
   ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_EQ(NumTableFilesAtLevel(0), 10);
@@ -887,6 +901,7 @@ TEST_F(DBOptionsTest, ChangeCompression) {
   options.bottommost_compression = CompressionType::kNoCompression;
   options.bottommost_compression_opts.level = 2;
   options.bottommost_compression_opts.parallel_threads = 1;
+  options.env = CurrentOptions().env;
 
   ASSERT_OK(TryReopen(options));
 
@@ -908,7 +923,7 @@ TEST_F(DBOptionsTest, ChangeCompression) {
   ASSERT_OK(Put("foo", "foofoofoo"));
   ASSERT_OK(Put("bar", "foofoofoo"));
   ASSERT_OK(Flush());
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_TRUE(compacted);
   ASSERT_EQ(CompressionType::kNoCompression, compression_used);
   ASSERT_EQ(options.compression_opts.level, compression_opt_used.level);
@@ -926,7 +941,7 @@ TEST_F(DBOptionsTest, ChangeCompression) {
   ASSERT_OK(Put("foo", "foofoofoo"));
   ASSERT_OK(Put("bar", "foofoofoo"));
   ASSERT_OK(Flush());
-  dbfull()->TEST_WaitForCompact();
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
   ASSERT_TRUE(compacted);
   ASSERT_EQ(CompressionType::kSnappyCompression, compression_used);
   ASSERT_EQ(6, compression_opt_used.level);
@@ -937,6 +952,66 @@ TEST_F(DBOptionsTest, ChangeCompression) {
 
 #endif  // ROCKSDB_LITE
 
+TEST_F(DBOptionsTest, BottommostCompressionOptsWithFallbackType) {
+  // Verify the bottommost compression options still take effect even when the
+  // bottommost compression type is left at its default value. Verify for both
+  // automatic and manual compaction.
+  if (!Snappy_Supported() || !LZ4_Supported()) {
+    return;
+  }
+
+  constexpr int kUpperCompressionLevel = 1;
+  constexpr int kBottommostCompressionLevel = 2;
+  constexpr int kNumL0Files = 2;
+
+  Options options = CurrentOptions();
+  options.level0_file_num_compaction_trigger = kNumL0Files;
+  options.compression = CompressionType::kLZ4Compression;
+  options.compression_opts.level = kUpperCompressionLevel;
+  options.bottommost_compression_opts.level = kBottommostCompressionLevel;
+  options.bottommost_compression_opts.enabled = true;
+  Reopen(options);
+
+  CompressionType compression_used = CompressionType::kDisableCompressionOption;
+  CompressionOptions compression_opt_used;
+  bool compacted = false;
+  SyncPoint::GetInstance()->SetCallBack(
+      "CompactionPicker::RegisterCompaction:Registered", [&](void* arg) {
+        Compaction* c = static_cast<Compaction*>(arg);
+        compression_used = c->output_compression();
+        compression_opt_used = c->output_compression_opts();
+        compacted = true;
+      });
+  SyncPoint::GetInstance()->EnableProcessing();
+
+  // First, verify for automatic compaction.
+  for (int i = 0; i < kNumL0Files; ++i) {
+    ASSERT_OK(Put("foo", "foofoofoo"));
+    ASSERT_OK(Put("bar", "foofoofoo"));
+    ASSERT_OK(Flush());
+  }
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+
+  ASSERT_TRUE(compacted);
+  ASSERT_EQ(CompressionType::kLZ4Compression, compression_used);
+  ASSERT_EQ(kBottommostCompressionLevel, compression_opt_used.level);
+
+  // Second, verify for manual compaction.
+  compacted = false;
+  compression_used = CompressionType::kDisableCompressionOption;
+  compression_opt_used = CompressionOptions();
+  CompactRangeOptions cro;
+  cro.bottommost_level_compaction = BottommostLevelCompaction::kForceOptimized;
+  ASSERT_OK(dbfull()->CompactRange(cro, nullptr, nullptr));
+
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
+  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
+
+  ASSERT_TRUE(compacted);
+  ASSERT_EQ(CompressionType::kLZ4Compression, compression_used);
+  ASSERT_EQ(kBottommostCompressionLevel, compression_opt_used.level);
+}
+
 }  // namespace ROCKSDB_NAMESPACE
 
 int main(int argc, char** argv) {