Merge branch 'main' into create-view

Bumps [rustls](https://github.com/rustls/rustls) from 0.22.3 to 0.22.4.
- [Release notes](https://github.com/rustls/rustls/releases)
- [Changelog](https://github.com/rustls/rustls/blob/main/CHANGELOG.md)
- [Commits](https://github.com/rustls/rustls/compare/v/0.22.3...v/0.22.4)

---
updated-dependencies:
- dependency-name: rustls
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

.cargo/config.toml

@@ -3,13 +3,3 @@ linker = "aarch64-linux-gnu-gcc"
 
 [alias]
 sqlness = "run --bin sqlness-runner --"
-
-
-[build]
-rustflags = [
-    # lints
-    # TODO: use lint configuration in cargo https://github.com/rust-lang/cargo/issues/5034
-    "-Wclippy::print_stdout",
-    "-Wclippy::print_stderr",
-    "-Wclippy::implicit_clone",
-]

.editorconfig

@@ -0,0 +1,10 @@
+root = true
+
+[*]
+end_of_line = lf
+indent_style = space
+insert_final_newline = true
+trim_trailing_whitespace = true
+
+[{Makefile,**.mk}]
+indent_style = tab

.env.example

@@ -21,3 +21,6 @@ GT_GCS_CREDENTIAL_PATH = GCS credential path
 GT_GCS_ENDPOINT = GCS end point
 # Settings for kafka wal test
 GT_KAFKA_ENDPOINTS = localhost:9092
+
+# Setting for fuzz tests
+GT_MYSQL_ADDR = localhost:4002

.github/CODEOWNERS

@@ -0,0 +1,27 @@
+# GreptimeDB CODEOWNERS
+
+# These owners will be the default owners for everything in the repo.
+
+* @GreptimeTeam/db-approver
+
+## [Module] Databse Engine
+/src/index @zhongzc
+/src/mito2 @evenyag @v0y4g3r @waynexia
+/src/query @evenyag
+
+## [Module] Distributed
+/src/common/meta @MichaelScofield
+/src/common/procedure @MichaelScofield
+/src/meta-client @MichaelScofield
+/src/meta-srv @MichaelScofield
+
+## [Module] Write Ahead Log
+/src/log-store @v0y4g3r
+/src/store-api @v0y4g3r
+
+## [Module] Metrics Engine
+/src/metric-engine @waynexia
+/src/promql @waynexia
+
+## [Module] Flow
+/src/flow @zhongzc @waynexia

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -39,7 +39,7 @@ body:
         - Query Engine
         - Table Engine
         - Write Protocols
-        - MetaSrv
+        - Metasrv
         - Frontend
         - Datanode
         - Other

.github/actions/build-greptime-images/action.yml

@@ -53,7 +53,7 @@ runs:
       uses: docker/setup-buildx-action@v2
 
     - name: Download amd64 artifacts
-      uses: actions/download-artifact@v3
+      uses: actions/download-artifact@v4
       with:
         name: ${{ inputs.amd64-artifact-name }}
 
@@ -66,7 +66,7 @@ runs:
         mv ${{ inputs.amd64-artifact-name }} amd64
 
     - name: Download arm64 artifacts
-      uses: actions/download-artifact@v3
+      uses: actions/download-artifact@v4
       if: ${{ inputs.arm64-artifact-name }}
       with:
         name: ${{ inputs.arm64-artifact-name }}

.github/actions/build-linux-artifacts/action.yml

@@ -34,7 +34,7 @@ runs:
 
     - name: Upload sqlness logs
       if: ${{ failure() && inputs.disable-run-tests == 'false' }} # Only upload logs when the integration tests failed.
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
         name: sqlness-logs
         path: /tmp/greptime-*.log

.github/actions/build-macos-artifacts/action.yml

@@ -67,7 +67,7 @@ runs:
 
     - name: Upload sqlness logs
       if: ${{ failure() }} # Only upload logs when the integration tests failed.
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
         name: sqlness-logs
         path: /tmp/greptime-*.log

.github/actions/build-windows-artifacts/action.yml

@@ -25,7 +25,9 @@ inputs:
 runs:
   using: composite
   steps:
-    - uses: arduino/setup-protoc@v1
+    - uses: arduino/setup-protoc@v3
+      with:
+        repo-token: ${{ secrets.GITHUB_TOKEN }}
 
     - name: Install rust toolchain
       uses: dtolnay/rust-toolchain@master
@@ -38,7 +40,7 @@ runs:
       uses: Swatinem/rust-cache@v2
 
     - name: Install Python
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: '3.10'
 
@@ -62,15 +64,15 @@ runs:
 
     - name: Upload sqlness logs
       if: ${{ failure() }} # Only upload logs when the integration tests failed.
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
         name: sqlness-logs
-        path: ${{ runner.temp }}/greptime-*.log
+        path: /tmp/greptime-*.log
         retention-days: 3
 
     - name: Build greptime binary
       shell: pwsh
-      run: cargo build --profile ${{ inputs.cargo-profile }} --features ${{ inputs.features }} --target ${{ inputs.arch }}
+      run: cargo build --profile ${{ inputs.cargo-profile }} --features ${{ inputs.features }} --target ${{ inputs.arch }} --bin greptime
 
     - name: Upload artifacts
       uses: ./.github/actions/upload-artifacts

.github/actions/fuzz-test/action.yaml

@@ -0,0 +1,13 @@
+name: Fuzz Test
+description: 'Fuzz test given setup and service'
+inputs:
+  target:
+    description: "The fuzz target to test"
+runs:
+  using: composite
+  steps:
+  - name: Run Fuzz Test
+    shell: bash
+    run: cargo fuzz run ${{ inputs.target }} --fuzz-dir tests-fuzz -D -s none -- -max_total_time=120
+    env:
+      GT_MYSQL_ADDR: 127.0.0.1:4002

.github/actions/publish-github-release/action.yml

@@ -15,7 +15,7 @@ runs:
     #   |- greptime-darwin-amd64-v0.5.0.sha256sum/greptime-darwin-amd64-v0.5.0.sha256sum
     #   ...
     - name: Download artifacts
-      uses: actions/download-artifact@v3
+      uses: actions/download-artifact@v4
 
     - name: Create git tag for release
       if: ${{ github.event_name != 'push' }} # Meaning this is a scheduled or manual workflow.

.github/actions/release-cn-artifacts/action.yaml

@@ -73,7 +73,7 @@ runs:
   using: composite
   steps:
     - name: Download artifacts
-      uses: actions/download-artifact@v3
+      uses: actions/download-artifact@v4
       with:
         path: ${{ inputs.artifacts-dir }}
 

.github/actions/upload-artifacts/action.yml

@@ -6,7 +6,7 @@ inputs:
     required: true
   target-file:
     description: The path of the target artifact
-    required: true
+    required: false
   version:
     description: Version of the artifact
     required: true
@@ -18,6 +18,7 @@ runs:
   using: composite
   steps:
     - name: Create artifacts directory
+      if: ${{ inputs.target-file != '' }}
       working-directory: ${{ inputs.working-dir }}
       shell: bash
       run: |
@@ -49,15 +50,15 @@ runs:
       run: Get-FileHash ${{ inputs.artifacts-dir }}.tar.gz -Algorithm SHA256 | select -ExpandProperty Hash > ${{ inputs.artifacts-dir }}.sha256sum
 
     # Note: The artifacts will be double zip compressed(related issue: https://github.com/actions/upload-artifact/issues/39).
-    # However, when we use 'actions/download-artifact@v3' to download the artifacts, it will be automatically unzipped.
+    # However, when we use 'actions/download-artifact' to download the artifacts, it will be automatically unzipped.
     - name: Upload artifacts
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
         name: ${{ inputs.artifacts-dir }}
         path: ${{ inputs.working-dir }}/${{ inputs.artifacts-dir }}.tar.gz
 
     - name: Upload checksum
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
         name: ${{ inputs.artifacts-dir }}.sha256sum
         path: ${{ inputs.working-dir }}/${{ inputs.artifacts-dir }}.sha256sum

.github/pull_request_template.md

@@ -1,8 +1,10 @@
-I hereby agree to the terms of the [GreptimeDB CLA](https://gist.github.com/xtang/6378857777706e568c1949c7578592cc)
+I hereby agree to the terms of the [GreptimeDB CLA](https://github.com/GreptimeTeam/.github/blob/main/CLA.md).
+
+## Refer to a related PR or issue link (optional)
 
 ## What's changed and what's your intention?
 
-_PLEASE DO NOT LEAVE THIS EMPTY !!!_
+__!!! DO NOT LEAVE THIS BLOCK EMPTY !!!__
 
 Please explain IN DETAIL what the changes are in this PR and why they are needed:
 
@@ -16,5 +18,3 @@ Please explain IN DETAIL what the changes are in this PR and why they are needed
 - [ ]  I have written the necessary rustdoc comments.
 - [ ]  I have added the necessary unit tests and integration tests.
 - [x]  This PR does not require documentation updates.
-
-## Refer to a related PR or issue link (optional)

.github/workflows/apidoc.yml

@@ -13,14 +13,14 @@ on:
 name: Build API docs
 
 env:
-  RUST_TOOLCHAIN: nightly-2023-12-19
+  RUST_TOOLCHAIN: nightly-2024-04-18
 
 jobs:
   apidoc:
     runs-on: ubuntu-20.04
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
-    - uses: arduino/setup-protoc@v1
+    - uses: arduino/setup-protoc@v3
       with:
         repo-token: ${{ secrets.GITHUB_TOKEN }}
     - uses: dtolnay/rust-toolchain@master
@@ -40,3 +40,4 @@ jobs:
       uses: JamesIves/github-pages-deploy-action@v4
       with:
         folder: target/doc
+        single-commit: true

.github/workflows/dev-build.yml

@@ -101,7 +101,7 @@ jobs:
       version: ${{ steps.create-version.outputs.version }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -155,12 +155,12 @@ jobs:
     runs-on: ${{ needs.allocate-runners.outputs.linux-amd64-runner }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
       - name: Checkout greptimedb
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           repository: ${{ inputs.repository }}
           ref: ${{ inputs.commit }}
@@ -184,12 +184,12 @@ jobs:
     runs-on: ${{ needs.allocate-runners.outputs.linux-arm64-runner }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
       - name: Checkout greptimedb
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           repository: ${{ inputs.repository }}
           ref: ${{ inputs.commit }}
@@ -216,7 +216,7 @@ jobs:
     outputs:
       build-result: ${{ steps.set-build-result.outputs.build-result }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -247,7 +247,7 @@ jobs:
     runs-on: ubuntu-20.04
     continue-on-error: true
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -281,7 +281,7 @@ jobs:
     ]
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -306,7 +306,7 @@ jobs:
     ]
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -330,14 +330,14 @@ jobs:
     env:
       SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
     steps:
-      - name: Notifiy nightly build successful result
+      - name: Notifiy dev build successful result
         uses: slackapi/slack-github-action@v1.23.0
         if: ${{ needs.release-images-to-dockerhub.outputs.build-result == 'success' }}
         with:
           payload: |
             {"text": "GreptimeDB's ${{ env.NEXT_RELEASE_VERSION }} build has completed successfully."}
 
-      - name: Notifiy nightly build failed result
+      - name: Notifiy dev build failed result
         uses: slackapi/slack-github-action@v1.23.0
         if: ${{ needs.release-images-to-dockerhub.outputs.build-result != 'success' }}
         with:

.github/workflows/develop.yml

@@ -9,6 +9,7 @@ on:
       - '.dockerignore'
       - 'docker/**'
       - '.gitignore'
+      - 'grafana/**'
   push:
     branches:
       - main
@@ -19,6 +20,7 @@ on:
       - '.dockerignore'
       - 'docker/**'
       - '.gitignore'
+      - 'grafana/**'
   workflow_dispatch:
 
 name: CI
@@ -28,27 +30,31 @@ concurrency:
   cancel-in-progress: true
 
 env:
-  RUST_TOOLCHAIN: nightly-2023-12-19
+  RUST_TOOLCHAIN: nightly-2024-04-18
 
 jobs:
-  typos:
+  check-typos-and-docs:
-    name: Spell Check with Typos
+    name: Check typos and docs
     runs-on: ubuntu-20.04
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - uses: crate-ci/typos@v1.13.10
+      - name: Check the config docs
+        run: |
+          make config-docs && \
+          git diff --name-only --exit-code ./config/config.md  \
+          || (echo "'config/config.md' is not up-to-date, please run 'make config-docs'." && exit 1)
 
   check:
     name: Check
-    if: github.event.pull_request.draft == false
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ windows-latest-8-cores, ubuntu-20.04 ]
+        os: [ windows-latest, ubuntu-20.04 ]
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -56,37 +62,78 @@ jobs:
           toolchain: ${{ env.RUST_TOOLCHAIN }}
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares across multiple jobs
+          # Shares with `Clippy` job
+          shared-key: "check-lint"
       - name: Run cargo check
         run: cargo check --locked --workspace --all-targets
 
   toml:
     name: Toml Check
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@master
         with:
           toolchain: stable
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares across multiple jobs
+          shared-key: "check-toml"
       - name: Install taplo
-        run: cargo +stable install taplo-cli --version ^0.8 --locked
+        run: cargo +stable install taplo-cli --version ^0.9 --locked
       - name: Run taplo
         run: taplo format --check
 
-  sqlness:
+  build:
-    name: Sqlness Test
+    name: Build GreptimeDB binaries
-    if: github.event.pull_request.draft == false
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ ubuntu-20.04-8-cores ]
+        os: [ ubuntu-20.04 ]
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
+        with:
+          repo-token: ${{ secrets.GITHUB_TOKEN }}
+      - uses: dtolnay/rust-toolchain@master
+        with:
+          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: Swatinem/rust-cache@v2
+        with:
+          # Shares across multiple jobs
+          shared-key: "build-binaries"
+      - name: Build greptime binaries
+        shell: bash
+        run: cargo build --bin greptime --bin sqlness-runner
+      - name: Pack greptime binaries
+        shell: bash
+        run: |
+          mkdir bins && \
+          mv ./target/debug/greptime bins && \
+          mv ./target/debug/sqlness-runner bins
+      - name: Print greptime binaries info
+        run: ls -lh bins
+      - name: Upload artifacts
+        uses: ./.github/actions/upload-artifacts
+        with:
+          artifacts-dir: bins
+          version: current
+
+  fuzztest:
+    name: Fuzz Test
+    needs: build
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        target: [ "fuzz_create_table", "fuzz_alter_table", "fuzz_create_database" ]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -94,55 +141,95 @@ jobs:
           toolchain: ${{ env.RUST_TOOLCHAIN }}
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares across multiple jobs
+          shared-key: "fuzz-test-targets"
+      - name: Set Rust Fuzz
+        shell: bash
+        run: |
+          sudo apt update && sudo apt install -y libfuzzer-14-dev
+          cargo install cargo-fuzz
+      - name: Download pre-built binaries
+        uses: actions/download-artifact@v4
+        with:
+          name: bins
+          path: .
+      - name: Unzip binaries
+        run: tar -xvf ./bins.tar.gz
+      - name: Run GreptimeDB
+        run: |
+          ./bins/greptime standalone start&
+      - name: Fuzz Test
+        uses: ./.github/actions/fuzz-test
+        env:
+          CUSTOM_LIBFUZZER_PATH: /usr/lib/llvm-14/lib/libFuzzer.a
+        with:
+          target: ${{ matrix.target }}
+
+  sqlness:
+    name: Sqlness Test
+    needs: build
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [ ubuntu-20.04 ]
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@v4
+      - name: Download pre-built binaries
+        uses: actions/download-artifact@v4
+        with:
+          name: bins
+          path: .
+      - name: Unzip binaries
+        run: tar -xvf ./bins.tar.gz
       - name: Run sqlness
-        run: cargo sqlness
+        run: RUST_BACKTRACE=1 ./bins/sqlness-runner -c ./tests/cases --bins-dir ./bins
       - name: Upload sqlness logs
         if: always()
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: sqlness-logs
-          path: ${{ runner.temp }}/greptime-*.log
+          path: /tmp/greptime-*.log
           retention-days: 3
 
   sqlness-kafka-wal:
     name: Sqlness Test with Kafka Wal
-    if: github.event.pull_request.draft == false
+    needs: build
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ ubuntu-20.04-8-cores ]
+        os: [ ubuntu-20.04 ]
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - name: Download pre-built binaries
+        uses: actions/download-artifact@v4
         with:
-          repo-token: ${{ secrets.GITHUB_TOKEN }}
+          name: bins
-      - uses: dtolnay/rust-toolchain@master
+          path: .
-        with:
+      - name: Unzip binaries
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+        run: tar -xvf ./bins.tar.gz
-      - name: Rust Cache
-        uses: Swatinem/rust-cache@v2
       - name: Setup kafka server
         working-directory: tests-integration/fixtures/kafka
         run: docker compose -f docker-compose-standalone.yml up -d --wait
       - name: Run sqlness
-        run: cargo sqlness -w kafka -k 127.0.0.1:9092
+        run: RUST_BACKTRACE=1 ./bins/sqlness-runner -w kafka -k 127.0.0.1:9092 -c ./tests/cases --bins-dir ./bins
       - name: Upload sqlness logs
         if: always()
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
-          name: sqlness-logs
+          name: sqlness-logs-with-kafka-wal
-          path: ${{ runner.temp }}/greptime-*.log
+          path: /tmp/greptime-*.log
           retention-days: 3
 
   fmt:
     name: Rustfmt
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -151,17 +238,19 @@ jobs:
           components: rustfmt
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares across multiple jobs
+          shared-key: "check-rust-fmt"
       - name: Run cargo fmt
         run: cargo fmt --all -- --check
 
   clippy:
     name: Clippy
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -170,6 +259,10 @@ jobs:
           components: clippy
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares across multiple jobs
+          # Shares with `Check` job
+          shared-key: "check-lint"
       - name: Run cargo clippy
         run: cargo clippy --workspace --all-targets -- -D warnings
 
@@ -178,8 +271,8 @@ jobs:
     runs-on: ubuntu-20.04-8-cores
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: KyleMayes/install-llvm-action@v1
@@ -192,12 +285,19 @@ jobs:
           components: llvm-tools-preview
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares cross multiple jobs
+          shared-key: "coverage-test"
+      - name: Docker Cache
+        uses: ScribeMD/docker-cache@0.3.7
+        with:
+          key: docker-${{ runner.os }}-coverage
       - name: Install latest nextest release
         uses: taiki-e/install-action@nextest
       - name: Install cargo-llvm-cov
         uses: taiki-e/install-action@cargo-llvm-cov
       - name: Install Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: '3.10'
       - name: Install PyArrow Package
@@ -214,18 +314,36 @@ jobs:
           CARGO_BUILD_RUSTFLAGS: "-C link-arg=-fuse-ld=lld"
           RUST_BACKTRACE: 1
           CARGO_INCREMENTAL: 0
-          GT_S3_BUCKET: ${{ secrets.S3_BUCKET }}
+          GT_S3_BUCKET: ${{ vars.AWS_CI_TEST_BUCKET }}
-          GT_S3_ACCESS_KEY_ID: ${{ secrets.S3_ACCESS_KEY_ID }}
+          GT_S3_ACCESS_KEY_ID: ${{ secrets.AWS_CI_TEST_ACCESS_KEY_ID }}
-          GT_S3_ACCESS_KEY: ${{ secrets.S3_ACCESS_KEY }}
+          GT_S3_ACCESS_KEY: ${{ secrets.AWS_CI_TEST_SECRET_ACCESS_KEY }}
-          GT_S3_REGION: ${{ secrets.S3_REGION }}
+          GT_S3_REGION: ${{ vars.AWS_CI_TEST_BUCKET_REGION }}
           GT_ETCD_ENDPOINTS: http://127.0.0.1:2379
           GT_KAFKA_ENDPOINTS: 127.0.0.1:9092
           UNITTEST_LOG_DIR: "__unittest_logs"
       - name: Codecov upload
-        uses: codecov/codecov-action@v2
+        uses: codecov/codecov-action@v4
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           files: ./lcov.info
           flags: rust
           fail_ci_if_error: false
           verbose: true
+
+  compat:
+    name: Compatibility Test
+    needs: build
+    runs-on: ubuntu-20.04
+    timeout-minutes: 60
+    steps:
+      - uses: actions/checkout@v4
+      - name: Download pre-built binaries
+        uses: actions/download-artifact@v4
+        with:
+          name: bins
+          path: .
+      - name: Unzip binaries
+        run: |
+          mkdir -p ./bins/current
+          tar -xvf ./bins.tar.gz --strip-components=1 -C ./bins/current
+      - run: ./tests/compat/test-compat.sh 0.6.0

.github/workflows/doc-issue.yml

@@ -14,7 +14,7 @@ jobs:
     runs-on: ubuntu-20.04
     steps:
       - name: create an issue in doc repo
-        uses: dacbd/create-issue-action@main
+        uses: dacbd/create-issue-action@v1.2.1
         with:
           owner: GreptimeTeam
           repo: docs
@@ -28,7 +28,7 @@ jobs:
     runs-on: ubuntu-20.04
     steps:
       - name: create an issue in cloud repo
-        uses: dacbd/create-issue-action@main
+        uses: dacbd/create-issue-action@v1.2.1
         with:
           owner: GreptimeTeam
           repo: greptimedb-cloud

.github/workflows/doc-label.yml

@@ -12,14 +12,14 @@ jobs:
     if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
     runs-on: ubuntu-latest
     steps:
-    - uses: github/issue-labeler@v3.3
+    - uses: github/issue-labeler@v3.4
       with:
         configuration-path: .github/doc-label-config.yml
         enable-versioned-regex: false
         repo-token: ${{ secrets.GITHUB_TOKEN }}
         sync-labels: 1
     - name: create an issue in doc repo
-      uses: dacbd/create-issue-action@main
+      uses: dacbd/create-issue-action@v1.2.1
       if: ${{ github.event.action == 'opened' && contains(github.event.pull_request.body, '- [ ]  This PR does not require documentation updates.') }}
       with:
         owner: GreptimeTeam
@@ -29,3 +29,8 @@ jobs:
         body: |
           A document change request is generated from
           ${{ github.event.issue.html_url || github.event.pull_request.html_url }}
+    - name: Check doc labels
+      uses: docker://agilepathway/pull-request-label-checker:latest
+      with:
+        one_of: Doc update required,Doc not needed
+        repo_token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/docs.yml

@@ -9,6 +9,7 @@ on:
       - '.dockerignore'
       - 'docker/**'
       - '.gitignore'
+      - 'grafana/**'
   push:
     branches:
       - main
@@ -19,6 +20,7 @@ on:
       - '.dockerignore'
       - 'docker/**'
       - '.gitignore'
+      - 'grafana/**'
   workflow_dispatch:
 
 name: CI
@@ -31,39 +33,46 @@ jobs:
     name: Spell Check with Typos
     runs-on: ubuntu-20.04
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - uses: crate-ci/typos@v1.13.10
 
   check:
     name: Check
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     steps:
       - run: 'echo "No action required"'
 
   fmt:
     name: Rustfmt
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     steps:
       - run: 'echo "No action required"'
 
   clippy:
     name: Clippy
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     steps:
       - run: 'echo "No action required"'
 
   coverage:
-    if: github.event.pull_request.draft == false
     runs-on: ubuntu-20.04
     steps:
       - run: 'echo "No action required"'
 
   sqlness:
     name: Sqlness Test
-    if: github.event.pull_request.draft == false
+    runs-on: ${{ matrix.os }}
-    runs-on: ubuntu-20.04
+    strategy:
+      matrix:
+        os: [ ubuntu-20.04 ]
+    steps:
+      - run: 'echo "No action required"'
+
+  sqlness-kafka-wal:
+    name: Sqlness Test with Kafka Wal
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [ ubuntu-20.04 ]
     steps:
       - run: 'echo "No action required"'

.github/workflows/license.yaml

@@ -11,6 +11,6 @@ jobs:
     runs-on: ubuntu-20.04
     name: license-header-check
     steps:
-    - uses: actions/checkout@v2
+    - uses: actions/checkout@v4
     - name: Check License Header
-      uses: korandoru/hawkeye@v3
+      uses: korandoru/hawkeye@v5

.github/workflows/nightly-build.yml

@@ -85,7 +85,7 @@ jobs:
       version: ${{ steps.create-version.outputs.version }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -137,7 +137,7 @@ jobs:
     ]
     runs-on: ${{ needs.allocate-runners.outputs.linux-amd64-runner }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -156,7 +156,7 @@ jobs:
     ]
     runs-on: ${{ needs.allocate-runners.outputs.linux-arm64-runner }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -179,7 +179,7 @@ jobs:
     outputs:
       nightly-build-result: ${{ steps.set-nightly-build-result.outputs.nightly-build-result }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -211,7 +211,7 @@ jobs:
     # The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
     continue-on-error: true
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -245,7 +245,7 @@ jobs:
     ]
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -270,7 +270,7 @@ jobs:
     ]
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 

.github/workflows/nightly-ci.yml

@@ -12,7 +12,7 @@ concurrency:
   cancel-in-progress: true
 
 env:
-  RUST_TOOLCHAIN: nightly-2023-12-19
+  RUST_TOOLCHAIN: nightly-2024-04-18
 
 jobs:
   sqlness:
@@ -24,8 +24,8 @@ jobs:
         os: [ windows-latest-8-cores ]
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v4.1.0
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -45,10 +45,10 @@ jobs:
             {"text": "Nightly CI failed for sqlness tests"}
       - name: Upload sqlness logs
         if: always()
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: sqlness-logs
-          path: ${{ runner.temp }}/greptime-*.log
+          path: /tmp/greptime-*.log
           retention-days: 3
 
   test-on-windows:
@@ -57,8 +57,8 @@ jobs:
     timeout-minutes: 60
     steps:
       - run: git config --global core.autocrlf false
-      - uses: actions/checkout@v4.1.0
+      - uses: actions/checkout@v4
-      - uses: arduino/setup-protoc@v1
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - name: Install Rust toolchain
@@ -71,7 +71,7 @@ jobs:
       - name: Install Cargo Nextest
         uses: taiki-e/install-action@nextest
       - name: Install Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v5
         with:
           python-version: '3.10'
       - name: Install PyArrow Package
@@ -85,10 +85,10 @@ jobs:
         env:
           RUST_BACKTRACE: 1
           CARGO_INCREMENTAL: 0
-          GT_S3_BUCKET: ${{ secrets.S3_BUCKET }}
+          GT_S3_BUCKET: ${{ vars.AWS_CI_TEST_BUCKET }}
-          GT_S3_ACCESS_KEY_ID: ${{ secrets.S3_ACCESS_KEY_ID }}
+          GT_S3_ACCESS_KEY_ID: ${{ secrets.AWS_CI_TEST_ACCESS_KEY_ID }}
-          GT_S3_ACCESS_KEY: ${{ secrets.S3_ACCESS_KEY }}
+          GT_S3_ACCESS_KEY: ${{ secrets.AWS_CI_TEST_SECRET_ACCESS_KEY }}
-          GT_S3_REGION: ${{ secrets.S3_REGION }}
+          GT_S3_REGION: ${{ vars.AWS_CI_TEST_BUCKET_REGION }}
           UNITTEST_LOG_DIR: "__unittest_logs"
       - name: Notify slack if failed
         if: failure()

.github/workflows/nightly-funtional-tests.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 

.github/workflows/pr-title-checker.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-20.04
     timeout-minutes: 10
     steps:
-      - uses: thehanimo/pr-title-checker@v1.3.4
+      - uses: thehanimo/pr-title-checker@v1.4.2
         with:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           pass_on_octokit_error: false
@@ -22,7 +22,7 @@ jobs:
     runs-on: ubuntu-20.04
     timeout-minutes: 10
     steps:
-      - uses: thehanimo/pr-title-checker@v1.3.4
+      - uses: thehanimo/pr-title-checker@v1.4.2
         with:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           pass_on_octokit_error: false

.github/workflows/release-dev-builder-images.yaml

@@ -30,7 +30,7 @@ jobs:
     runs-on: ubuntu-20.04-16-cores
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 

.github/workflows/release.yml

@@ -82,7 +82,7 @@ on:
 # Use env variables to control all the release process.
 env:
   # The arguments of building greptime.
-  RUST_TOOLCHAIN: nightly-2023-12-19
+  RUST_TOOLCHAIN: nightly-2024-04-18
   CARGO_PROFILE: nightly
 
   # Controls whether to run tests, include unit-test, integration-test and sqlness.
@@ -91,7 +91,7 @@ env:
   # The scheduled version is '${{ env.NEXT_RELEASE_VERSION }}-nightly-YYYYMMDD', like v0.2.0-nigthly-20230313;
   NIGHTLY_RELEASE_PREFIX: nightly
   # Note: The NEXT_RELEASE_VERSION should be modified manually by every formal release.
-  NEXT_RELEASE_VERSION: v0.7.0
+  NEXT_RELEASE_VERSION: v0.8.0
 
 jobs:
   allocate-runners:
@@ -114,7 +114,7 @@ jobs:
       version: ${{ steps.create-version.outputs.version }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -168,7 +168,7 @@ jobs:
     ]
     runs-on: ${{ needs.allocate-runners.outputs.linux-amd64-runner }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -187,7 +187,7 @@ jobs:
     ]
     runs-on: ${{ needs.allocate-runners.outputs.linux-arm64-runner }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -221,12 +221,14 @@ jobs:
             arch: x86_64-apple-darwin
             artifacts-dir-prefix: greptime-darwin-amd64-pyo3
     runs-on: ${{ matrix.os }}
+    outputs:
+      build-macos-result: ${{ steps.set-build-macos-result.outputs.build-macos-result }}
     needs: [
       allocate-runners,
     ]
     if: ${{ inputs.build_macos_artifacts || github.event_name == 'push' || github.event_name == 'schedule' }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -240,6 +242,11 @@ jobs:
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
           artifacts-dir: ${{ matrix.artifacts-dir-prefix }}-${{ needs.allocate-runners.outputs.version }}
 
+      - name: Set build macos result
+        id: set-build-macos-result
+        run: |
+          echo "build-macos-result=success" >> $GITHUB_OUTPUT    
+
   build-windows-artifacts:
     name: Build Windows artifacts
     strategy:
@@ -255,6 +262,8 @@ jobs:
             features: pyo3_backend,servers/dashboard
             artifacts-dir-prefix: greptime-windows-amd64-pyo3
     runs-on: ${{ matrix.os }}
+    outputs:
+      build-windows-result: ${{ steps.set-build-windows-result.outputs.build-windows-result }}
     needs: [
       allocate-runners,
     ]
@@ -262,7 +271,7 @@ jobs:
     steps:
       - run: git config --global core.autocrlf false
 
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -276,6 +285,11 @@ jobs:
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
           artifacts-dir: ${{ matrix.artifacts-dir-prefix }}-${{ needs.allocate-runners.outputs.version }}
 
+      - name: Set build windows result
+        id: set-build-windows-result
+        run: |
+          echo "build-windows-result=success" >> $Env:GITHUB_OUTPUT
+
   release-images-to-dockerhub:
     name: Build and push images to DockerHub
     if: ${{ inputs.release_images || github.event_name == 'push' || github.event_name == 'schedule' }}
@@ -285,8 +299,10 @@ jobs:
       build-linux-arm64-artifacts,
     ]
     runs-on: ubuntu-2004-16-cores
+    outputs:
+      build-image-result: ${{ steps.set-build-image-result.outputs.build-image-result }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -299,6 +315,11 @@ jobs:
           image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
           version: ${{ needs.allocate-runners.outputs.version }}
 
+      - name: Set build image result
+        id: set-build-image-result
+        run: |
+          echo "build-image-result=success" >> $GITHUB_OUTPUT    
+
   release-cn-artifacts:
     name: Release artifacts to CN region
     if: ${{ inputs.release_images || github.event_name == 'push' || github.event_name == 'schedule' }}
@@ -316,7 +337,7 @@ jobs:
     # The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
     continue-on-error: true
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -352,7 +373,7 @@ jobs:
     ]
     runs-on: ubuntu-20.04
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -375,7 +396,7 @@ jobs:
     ]
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -400,7 +421,7 @@ jobs:
     ]
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -413,3 +434,29 @@ jobs:
           aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
           aws-region: ${{ vars.EC2_RUNNER_REGION }}
           github-token: ${{ secrets.GH_PERSONAL_ACCESS_TOKEN }}
+
+  notification:
+    if: ${{ always() }} # Not requiring successful dependent jobs, always run.
+    name: Send notification to Greptime team
+    needs: [
+      release-images-to-dockerhub,
+      build-macos-artifacts,
+      build-windows-artifacts,
+    ]
+    runs-on: ubuntu-20.04
+    env:
+      SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
+    steps:
+      - name: Notifiy release successful result
+        uses: slackapi/slack-github-action@v1.25.0
+        if: ${{ needs.release-images-to-dockerhub.outputs.build-image-result == 'success' && needs.build-windows-artifacts.outputs.build-windows-result == 'success' && needs.build-macos-artifacts.outputs.build-macos-result == 'success' }}
+        with:
+          payload: |
+            {"text": "GreptimeDB's release version has completed successfully."}
+
+      - name: Notifiy release failed result
+        uses: slackapi/slack-github-action@v1.25.0
+        if: ${{ needs.release-images-to-dockerhub.outputs.build-image-result != 'success' || needs.build-windows-artifacts.outputs.build-windows-result != 'success' || needs.build-macos-artifacts.outputs.build-macos-result != 'success' }}
+        with:
+          payload: |
+            {"text": "GreptimeDB's release version has failed, please check 'https://github.com/GreptimeTeam/greptimedb/actions/workflows/release.yml'."}

.github/workflows/size-label.yml

@@ -1,25 +0,0 @@
-name: size-labeler
-
-on: [pull_request_target]
-
-jobs:
-  labeler:
-    runs-on: ubuntu-latest
-    name: Label the PR size
-    permissions:
-      issues: write
-      pull-requests: write
-    steps:
-      - uses: codelytv/pr-size-labeler@v1
-        with:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          s_label: 'Size: S'
-          s_max_size: '100'
-          m_label: 'Size: M'
-          m_max_size: '500'
-          l_label: 'Size: L'
-          l_max_size: '1000'
-          xl_label: 'Size: XL'
-          fail_if_xl: 'false'
-          message_if_xl: ""
-          files_to_ignore: 'Cargo.lock'

.github/workflows/unassign.yml

@@ -0,0 +1,21 @@
+name: Auto Unassign
+on:
+  schedule:
+    - cron: '4 2 * * *'
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  issues: write
+  pull-requests: write
+
+jobs:
+  auto-unassign:
+    name: Auto Unassign
+    runs-on: ubuntu-latest
+    steps:
+      - name: Auto Unassign
+        uses: tisonspieces/auto-unassign@main
+        with:
+          token: ${{ secrets.GH_PERSONAL_ACCESS_TOKEN }}
+          repository: ${{ github.repository }}

.github/workflows/user-doc-label-checker.yml

@@ -1,19 +0,0 @@
-name: Check user doc labels
-on:
-  pull_request:
-    types:
-      - opened
-      - reopened
-      - labeled
-      - unlabeled
-
-jobs:
-
-  check_labels:
-    name: Check doc labels
-    runs-on: ubuntu-latest
-    steps:
-      - uses: docker://agilepathway/pull-request-label-checker:latest
-        with:
-          one_of: Doc update required,Doc not needed
-          repo_token: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -46,3 +46,7 @@ benchmarks/data
 *.code-workspace
 
 venv/
+
+# Fuzz tests 
+tests-fuzz/artifacts/
+tests-fuzz/corpus/

CODE_OF_CONDUCT.md

@@ -1,132 +0,0 @@
-# Contributor Covenant Code of Conduct
-
-## Our Pledge
-
-We as members, contributors, and leaders pledge to make participation in our
-community a harassment-free experience for everyone, regardless of age, body
-size, visible or invisible disability, ethnicity, sex characteristics, gender
-identity and expression, level of experience, education, socio-economic status,
-nationality, personal appearance, race, caste, color, religion, or sexual
-identity and orientation.
-
-We pledge to act and interact in ways that contribute to an open, welcoming,
-diverse, inclusive, and healthy community.
-
-## Our Standards
-
-Examples of behavior that contributes to a positive environment for our
-community include:
-
-* Demonstrating empathy and kindness toward other people
-* Being respectful of differing opinions, viewpoints, and experiences
-* Giving and gracefully accepting constructive feedback
-* Accepting responsibility and apologizing to those affected by our mistakes,
-  and learning from the experience
-* Focusing on what is best not just for us as individuals, but for the overall
-  community
-
-Examples of unacceptable behavior include:
-
-* The use of sexualized language or imagery, and sexual attention or advances of
-  any kind
-* Trolling, insulting or derogatory comments, and personal or political attacks
-* Public or private harassment
-* Publishing others' private information, such as a physical or email address,
-  without their explicit permission
-* Other conduct which could reasonably be considered inappropriate in a
-  professional setting
-
-## Enforcement Responsibilities
-
-Community leaders are responsible for clarifying and enforcing our standards of
-acceptable behavior and will take appropriate and fair corrective action in
-response to any behavior that they deem inappropriate, threatening, offensive,
-or harmful.
-
-Community leaders have the right and responsibility to remove, edit, or reject
-comments, commits, code, wiki edits, issues, and other contributions that are
-not aligned to this Code of Conduct, and will communicate reasons for moderation
-decisions when appropriate.
-
-## Scope
-
-This Code of Conduct applies within all community spaces, and also applies when
-an individual is officially representing the community in public spaces.
-Examples of representing our community include using an official e-mail address,
-posting via an official social media account, or acting as an appointed
-representative at an online or offline event.
-
-## Enforcement
-
-Instances of abusive, harassing, or otherwise unacceptable behavior may be
-reported to the community leaders responsible for enforcement at
-info@greptime.com.
-All complaints will be reviewed and investigated promptly and fairly.
-
-All community leaders are obligated to respect the privacy and security of the
-reporter of any incident.
-
-## Enforcement Guidelines
-
-Community leaders will follow these Community Impact Guidelines in determining
-the consequences for any action they deem in violation of this Code of Conduct:
-
-### 1. Correction
-
-**Community Impact**: Use of inappropriate language or other behavior deemed
-unprofessional or unwelcome in the community.
-
-**Consequence**: A private, written warning from community leaders, providing
-clarity around the nature of the violation and an explanation of why the
-behavior was inappropriate. A public apology may be requested.
-
-### 2. Warning
-
-**Community Impact**: A violation through a single incident or series of
-actions.
-
-**Consequence**: A warning with consequences for continued behavior. No
-interaction with the people involved, including unsolicited interaction with
-those enforcing the Code of Conduct, for a specified period of time. This
-includes avoiding interactions in community spaces as well as external channels
-like social media. Violating these terms may lead to a temporary or permanent
-ban.
-
-### 3. Temporary Ban
-
-**Community Impact**: A serious violation of community standards, including
-sustained inappropriate behavior.
-
-**Consequence**: A temporary ban from any sort of interaction or public
-communication with the community for a specified period of time. No public or
-private interaction with the people involved, including unsolicited interaction
-with those enforcing the Code of Conduct, is allowed during this period.
-Violating these terms may lead to a permanent ban.
-
-### 4. Permanent Ban
-
-**Community Impact**: Demonstrating a pattern of violation of community
-standards, including sustained inappropriate behavior, harassment of an
-individual, or aggression toward or disparagement of classes of individuals.
-
-**Consequence**: A permanent ban from any sort of public interaction within the
-community.
-
-## Attribution
-
-This Code of Conduct is adapted from the [Contributor Covenant][homepage],
-version 2.1, available at
-[https://www.contributor-covenant.org/version/2/1/code_of_conduct.html][v2.1].
-
-Community Impact Guidelines were inspired by
-[Mozilla's code of conduct enforcement ladder][Mozilla CoC].
-
-For answers to common questions about this code of conduct, see the FAQ at
-[https://www.contributor-covenant.org/faq][FAQ]. Translations are available at
-[https://www.contributor-covenant.org/translations][translations].
-
-[homepage]: https://www.contributor-covenant.org
-[v2.1]: https://www.contributor-covenant.org/version/2/1/code_of_conduct.html
-[Mozilla CoC]: https://github.com/mozilla/diversity
-[FAQ]: https://www.contributor-covenant.org/faq
-[translations]: https://www.contributor-covenant.org/translations

CONTRIBUTING.md

@@ -50,7 +50,7 @@ GreptimeDB uses the [Apache 2.0 license](https://github.com/GreptimeTeam/greptim
 
 - To ensure that community is free and confident in its ability to use your contributions, please sign the Contributor License Agreement (CLA) which will be incorporated in the pull request process.
 - Make sure all files have proper license header (running `docker run --rm -v $(pwd):/github/workspace ghcr.io/korandoru/hawkeye-native:v3 format` from the project root).
-- Make sure all your codes are formatted and follow the [coding style](https://pingcap.github.io/style-guide/rust/).
+- Make sure all your codes are formatted and follow the [coding style](https://pingcap.github.io/style-guide/rust/) and [style guide](http://github.com/greptimeTeam/docs/style-guide.md).
 - Make sure all unit tests are passed (using `cargo test --workspace` or [nextest](https://nexte.st/index.html) `cargo nextest run`).
 - Make sure all clippy warnings are fixed (you can check it locally by running `cargo clippy --workspace --all-targets -- -D warnings`).
 

Cargo.toml

@@ -18,6 +18,7 @@ members = [
     "src/common/grpc-expr",
     "src/common/mem-prof",
     "src/common/meta",
+    "src/common/plugins",
     "src/common/procedure",
     "src/common/procedure-test",
     "src/common/query",
@@ -29,9 +30,11 @@ members = [
     "src/common/time",
     "src/common/decimal",
     "src/common/version",
+    "src/common/wal",
     "src/datanode",
     "src/datatypes",
     "src/file-engine",
+    "src/flow",
     "src/frontend",
     "src/log-store",
     "src/meta-client",
@@ -52,89 +55,118 @@ members = [
     "src/store-api",
     "src/table",
     "src/index",
+    "tests-fuzz",
     "tests-integration",
     "tests/runner",
 ]
 resolver = "2"
 
 [workspace.package]
-version = "0.6.0"
+version = "0.7.2"
 edition = "2021"
 license = "Apache-2.0"
 
+[workspace.lints]
+clippy.print_stdout = "warn"
+clippy.print_stderr = "warn"
+clippy.implicit_clone = "warn"
+clippy.readonly_write_lock = "allow"
+rust.unknown_lints = "deny"
+# Remove this after https://github.com/PyO3/pyo3/issues/4094
+rust.non_local_definitions = "allow"
+
 [workspace.dependencies]
+# We turn off default-features for some dependencies here so the workspaces which inherit them can
+# selectively turn them on if needed, since we can override default-features = true (from false)
+# for the inherited dependency but cannot do the reverse (override from true to false).
+#
+# See for more detaiils: https://github.com/rust-lang/cargo/issues/11329
 ahash = { version = "0.8", features = ["compile-time-rng"] }
 aquamarine = "0.3"
-arrow = { version = "47.0" }
+arrow = { version = "51.0.0", features = ["prettyprint"] }
-arrow-array = "47.0"
+arrow-array = { version = "51.0.0", default-features = false, features = ["chrono-tz"] }
-arrow-flight = "47.0"
+arrow-flight = "51.0"
-arrow-schema = { version = "47.0", features = ["serde"] }
+arrow-ipc = { version = "51.0.0", default-features = false, features = ["lz4"] }
+arrow-schema = { version = "51.0", features = ["serde"] }
 async-stream = "0.3"
 async-trait = "0.1"
+axum = { version = "0.6", features = ["headers"] }
 base64 = "0.21"
 bigdecimal = "0.4.2"
 bitflags = "2.4.1"
 bytemuck = "1.12"
 bytes = { version = "1.5", features = ["serde"] }
 chrono = { version = "0.4", features = ["serde"] }
+clap = { version = "4.4", features = ["derive"] }
 dashmap = "5.4"
-datafusion = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
-datafusion-common = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion-common = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
-datafusion-expr = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion-expr = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
-datafusion-optimizer = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion-functions = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
-datafusion-physical-expr = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion-optimizer = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
-datafusion-sql = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion-physical-expr = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
-datafusion-substrait = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
+datafusion-sql = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
+datafusion-substrait = { git = "https://github.com/apache/arrow-datafusion.git", rev = "34eda15b73a9e278af8844b30ed2f1c21c10359c" }
 derive_builder = "0.12"
-etcd-client = "0.12"
+dotenv = "0.15"
+# TODO(LFC): Wait for https://github.com/etcdv3/etcd-client/pull/76
+etcd-client = { git = "https://github.com/MichaelScofield/etcd-client.git", rev = "4c371e9b3ea8e0a8ee2f9cbd7ded26e54a45df3b" }
 fst = "0.4.7"
 futures = "0.3"
 futures-util = "0.3"
-greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "a31ea166fc015ea7ff111ac94e26c3a5d64364d2" }
+greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "73ac0207ab71dfea48f30259ffdb611501b5ecb8" }
+humantime = "2.1"
 humantime-serde = "1.1"
 itertools = "0.10"
 lazy_static = "1.4"
-meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "abbd357c1e193cd270ea65ee7652334a150b628f" }
+meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "80b72716dcde47ec4161478416a5c6c21343364d" }
 mockall = "0.11.4"
 moka = "0.12"
+notify = "6.1"
+num_cpus = "1.16"
 once_cell = "1.18"
-opentelemetry-proto = { git = "https://github.com/waynexia/opentelemetry-rust.git", rev = "33841b38dda79b15f2024952be5f32533325ca02", features = [
+opentelemetry-proto = { version = "0.5", features = [
     "gen-tonic",
     "metrics",
     "trace",
 ] }
-parquet = "47.0"
+parquet = { version = "51.0.0", default-features = false, features = ["arrow", "async", "object_store"] }
 paste = "1.0"
 pin-project = "1.0"
 prometheus = { version = "0.13.3", features = ["process"] }
 prost = "0.12"
-raft-engine = { git = "https://github.com/tikv/raft-engine.git", rev = "22dfb426cd994602b57725ef080287d3e53db479" }
+raft-engine = { version = "0.4.1", default-features = false }
 rand = "0.8"
 regex = "1.8"
-regex-automata = { version = "0.2", features = ["transducer"] }
+regex-automata = { version = "0.4" }
 reqwest = { version = "0.11", default-features = false, features = [
     "json",
     "rustls-tls-native-roots",
     "stream",
+    "multipart",
 ] }
 rskafka = "0.5"
 rust_decimal = "1.33"
+schemars = "0.8"
 serde = { version = "1.0", features = ["derive"] }
-serde_json = "1.0"
+serde_json = { version = "1.0", features = ["float_roundtrip"] }
+serde_with = "3"
 smallvec = { version = "1", features = ["serde"] }
 snafu = "0.7"
-# on branch v0.38.x
+sysinfo = "0.30"
-sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "6a93567ae38d42be5c8d08b13c8ff4dde26502ef", features = [
+# on branch v0.44.x
+sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "c919990bf62ad38d2b0c0a3bc90b26ad919d51b0", features = [
     "visitor",
 ] }
 strum = { version = "0.25", features = ["derive"] }
 tempfile = "3"
-tokio = { version = "1.28", features = ["full"] }
+tokio = { version = "1.36", features = ["full"] }
 tokio-stream = { version = "0.1" }
 tokio-util = { version = "0.7", features = ["io-util", "compat"] }
 toml = "0.8.8"
-tonic = { version = "0.10", features = ["tls"] }
+tonic = { version = "0.11", features = ["tls"] }
-uuid = { version = "1", features = ["serde", "v4", "fast-rng"] }
+uuid = { version = "1.7", features = ["serde", "v4", "fast-rng"] }
+zstd = "0.13"
 
 ## workspaces members
 api = { path = "src/api" }
@@ -155,7 +187,7 @@ common-grpc-expr = { path = "src/common/grpc-expr" }
 common-macro = { path = "src/common/macro" }
 common-mem-prof = { path = "src/common/mem-prof" }
 common-meta = { path = "src/common/meta" }
-common-pprof = { path = "src/common/pprof" }
+common-plugins = { path = "src/common/plugins" }
 common-procedure = { path = "src/common/procedure" }
 common-procedure-test = { path = "src/common/procedure-test" }
 common-query = { path = "src/common/query" }
@@ -165,6 +197,7 @@ common-telemetry = { path = "src/common/telemetry" }
 common-test-util = { path = "src/common/test-util" }
 common-time = { path = "src/common/time" }
 common-version = { path = "src/common/version" }
+common-wal = { path = "src/common/wal" }
 datanode = { path = "src/datanode" }
 datatypes = { path = "src/datatypes" }
 file-engine = { path = "src/file-engine" }
@@ -192,7 +225,7 @@ table = { path = "src/table" }
 
 [workspace.dependencies.meter-macros]
 git = "https://github.com/GreptimeTeam/greptime-meter.git"
-rev = "abbd357c1e193cd270ea65ee7652334a150b628f"
+rev = "80b72716dcde47ec4161478416a5c6c21343364d"
 
 [profile.release]
 debug = 1

Makefile

@@ -3,6 +3,7 @@ CARGO_PROFILE ?=
 FEATURES ?=
 TARGET_DIR ?=
 TARGET ?=
+BUILD_BIN ?= greptime
 CARGO_BUILD_OPTS := --locked
 IMAGE_REGISTRY ?= docker.io
 IMAGE_NAMESPACE ?= greptime
@@ -45,6 +46,10 @@ ifneq ($(strip $(TARGET)),)
 	CARGO_BUILD_OPTS += --target ${TARGET}
 endif
 
+ifneq ($(strip $(BUILD_BIN)),)
+	CARGO_BUILD_OPTS += --bin ${BUILD_BIN}
+endif
+
 ifneq ($(strip $(RELEASE)),)
 	CARGO_BUILD_OPTS += --release
 endif
@@ -65,7 +70,7 @@ endif
 build: ## Build debug version greptime.
 	cargo ${CARGO_EXTENSION} build ${CARGO_BUILD_OPTS}
 
-.POHNY: build-by-dev-builder
+.PHONY: build-by-dev-builder
 build-by-dev-builder: ## Build greptime by dev-builder.
 	docker run --network=host \
 	-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry \
@@ -144,11 +149,12 @@ multi-platform-buildx: ## Create buildx multi-platform builder.
 	docker buildx inspect ${BUILDX_BUILDER_NAME} || docker buildx create --name ${BUILDX_BUILDER_NAME} --driver docker-container --bootstrap --use
 
 ##@ Test
+.PHONY: test
 test: nextest ## Run unit and integration tests.
 	cargo nextest run ${NEXTEST_OPTS}
 
-.PHONY: nextest ## Install nextest tools.
+.PHONY: nextest
-nextest:
+nextest: ## Install nextest tools.
 	cargo --list | grep nextest || cargo install cargo-nextest --locked
 
 .PHONY: sqlness-test
@@ -163,6 +169,10 @@ check: ## Cargo check all the targets.
 clippy: ## Check clippy rules.
 	cargo clippy --workspace --all-targets --all-features -- -D warnings
 
+.PHONY: fix-clippy
+fix-clippy: ## Fix clippy violations.
+	cargo clippy --workspace --all-targets --all-features --fix
+
 .PHONY: fmt-check
 fmt-check: ## Check code format.
 	cargo fmt --all -- --check
@@ -182,6 +192,16 @@ run-it-in-container: start-etcd ## Run integration tests in dev-builder.
 	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:latest \
 	make test sqlness-test BUILD_JOBS=${BUILD_JOBS}
 
+##@ Docs
+config-docs: ## Generate configuration documentation from toml files.
+	docker run --rm \
+    -v ${PWD}:/greptimedb \
+    -w /greptimedb/config \
+    toml2docs/toml2docs:latest \
+    -p '##' \
+    -t ./config-docs-template.md \
+    -o ./config.md
+
 ##@ General
 
 # The help target prints out all targets with their descriptions organized

README.md

@@ -6,141 +6,154 @@
   </picture>
 </p>
 
+<h1 align="center">Cloud-scale, Fast and Efficient Time Series Database</h1>
 
+<div align="center">
 <h3 align="center">
-    The next-generation hybrid time-series/analytics processing database in the cloud
+  <a href="https://greptime.com/product/cloud">GreptimeCloud</a> |
-</h3>
+  <a href="https://docs.greptime.com/">User guide</a> |
+  <a href="https://greptimedb.rs/">API Docs</a> |
+  <a href="https://github.com/GreptimeTeam/greptimedb/issues/3412">Roadmap 2024</a>
+</h4>
 
-<p align="center">
+<a href="https://github.com/GreptimeTeam/greptimedb/releases/latest">
-    <a href="https://codecov.io/gh/GrepTimeTeam/greptimedb"><img src="https://codecov.io/gh/GrepTimeTeam/greptimedb/branch/main/graph/badge.svg?token=FITFDI3J3C"></img></a>
+<img src="https://img.shields.io/github/v/release/GreptimeTeam/greptimedb.svg" alt="Version"/>
-    &nbsp;
+</a>
-    <a href="https://github.com/GreptimeTeam/greptimedb/actions/workflows/develop.yml"><img src="https://github.com/GreptimeTeam/greptimedb/actions/workflows/develop.yml/badge.svg" alt="CI"></img></a>
+<a href="https://github.com/GreptimeTeam/greptimedb/releases/latest">
-    &nbsp;
+<img src="https://img.shields.io/github/release-date/GreptimeTeam/greptimedb.svg" alt="Releases"/>
-    <a href="https://github.com/greptimeTeam/greptimedb/blob/main/LICENSE"><img src="https://img.shields.io/github/license/greptimeTeam/greptimedb"></a>
+</a>
-</p>
+<a href="https://hub.docker.com/r/greptime/greptimedb/">
+<img src="https://img.shields.io/docker/pulls/greptime/greptimedb.svg" alt="Docker Pulls"/>
+</a>
+<a href="https://github.com/GreptimeTeam/greptimedb/actions/workflows/develop.yml">
+<img src="https://github.com/GreptimeTeam/greptimedb/actions/workflows/develop.yml/badge.svg" alt="GitHub Actions"/>
+</a>
+<a href="https://codecov.io/gh/GrepTimeTeam/greptimedb">
+<img src="https://codecov.io/gh/GrepTimeTeam/greptimedb/branch/main/graph/badge.svg?token=FITFDI3J3C" alt="Codecov"/>
+</a>
+<a href="https://github.com/greptimeTeam/greptimedb/blob/main/LICENSE">
+<img src="https://img.shields.io/github/license/greptimeTeam/greptimedb" alt="License"/>
+</a>
 
-<p align="center">
+<br/>
-    <a href="https://twitter.com/greptime"><img src="https://img.shields.io/badge/twitter-follow_us-1d9bf0.svg"></a>
-    &nbsp;
-    <a href="https://www.linkedin.com/company/greptime/"><img src="https://img.shields.io/badge/linkedin-connect_with_us-0a66c2.svg"></a>
-    &nbsp;
-    <a href="https://greptime.com/slack"><img src="https://img.shields.io/badge/slack-GreptimeDB-0abd59?logo=slack" alt="slack" /></a>
-</p>
 
-## What is GreptimeDB
+<a href="https://greptime.com/slack">
+<img src="https://img.shields.io/badge/slack-GreptimeDB-0abd59?logo=slack&style=for-the-badge" alt="Slack"/>
+</a>
+<a href="https://twitter.com/greptime">
+<img src="https://img.shields.io/badge/twitter-follow_us-1d9bf0.svg?style=for-the-badge" alt="Twitter"/>
+</a>
+<a href="https://www.linkedin.com/company/greptime/">
+<img src="https://img.shields.io/badge/linkedin-connect_with_us-0a66c2.svg?style=for-the-badge" alt="LinkedIn"/>
+</a>
+</div>
 
-GreptimeDB is an open-source time-series database focusing on efficiency, scalability, and analytical capabilities.
+## Introduction
-It's designed to work on infrastructure of the cloud era, and users benefit from its elasticity and commodity storage.
 
-Our core developers have been building time-series data platforms for years. Based on their best-practices, GreptimeDB is born to give you:
+**GreptimeDB** is an open-source time-series database focusing on efficiency, scalability, and analytical capabilities.
+Designed to work on infrastructure of the cloud era, GreptimeDB benefits users with its elasticity and commodity storage, offering a fast and cost-effective **alternative to InfluxDB** and a **long-term storage for Prometheus**.
 
-- Optimized columnar layout for handling time-series data; compacted, compressed, and stored on various storage backends, particularly cloud object storage with 50x cost efficiency.
+## Why GreptimeDB
-- Fully open-source distributed cluster architecture that harnesses the power of cloud-native elastic computing resources.
-- Seamless scalability from a standalone binary at edge to a robust, highly available distributed cluster in cloud, with a transparent experience for both developers and administrators.
-- Native SQL and PromQL for queries, and Python scripting to facilitate complex analytical tasks.
-- Flexible indexing capabilities and distributed, parallel-processing query engine, tackling high cardinality issues down.
-- Widely adopted database protocols and APIs, including MySQL, PostgreSQL, and Prometheus Remote Storage, etc.
 
-## Quick Start
+Our core developers have been building time-series data platforms for years. Based on our best-practices, GreptimeDB is born to give you:
 
-### [GreptimePlay](https://greptime.com/playground)
+* **Easy horizontal scaling**
+
+  Seamless scalability from a standalone binary at edge to a robust, highly available distributed cluster in cloud, with a transparent experience for both developers and administrators.
+
+* **Analyzing time-series data**
+
+  Query your time-series data with SQL and PromQL. Use Python scripts to facilitate complex analytical tasks.
+
+* **Cloud-native distributed database**
+
+  Fully open-source distributed cluster architecture that harnesses the power of cloud-native elastic computing resources.
+
+* **Performance and Cost-effective**
+
+  Flexible indexing capabilities and distributed, parallel-processing query engine, tackling high cardinality issues down. Optimized columnar layout for handling time-series data; compacted, compressed, and stored on various storage backends, particularly cloud object storage with 50x cost efficiency.
+
+* **Compatible with InfluxDB, Prometheus and more protocols**
+
+  Widely adopted database protocols and APIs, including MySQL, PostgreSQL, and Prometheus Remote Storage, etc. [Read more](https://docs.greptime.com/user-guide/clients/overview).
+
+## Try GreptimeDB
+
+### 1. [GreptimePlay](https://greptime.com/playground)
 
 Try out the features of GreptimeDB right from your browser.
 
-### Build
+### 2. [GreptimeCloud](https://console.greptime.cloud/)
 
-#### Build from Source
+Start instantly with a free cluster.
 
-To compile GreptimeDB from source, you'll need:
+### 3. Docker Image
 
-- C/C++ Toolchain: provides basic tools for compiling and linking. This is
+To install GreptimeDB locally, the recommended way is via Docker:
-  available as `build-essential` on ubuntu and similar name on other platforms.
-- Rust: the easiest way to install Rust is to use
-  [`rustup`](https://rustup.rs/), which will check our `rust-toolchain` file and
-  install correct Rust version for you.
-- Protobuf: `protoc` is required for compiling `.proto` files. `protobuf` is
-  available from major package manager on macos and linux distributions. You can
-  find an installation instructions [here](https://grpc.io/docs/protoc-installation/).
-  **Note that `protoc` version needs to be >= 3.15** because we have used the `optional`
-  keyword. You can check it with `protoc --version`.
-- python3-dev or python3-devel(Optional feature, only needed if you want to run scripts
-  in CPython, and also need to enable `pyo3_backend` feature when compiling(by `cargo run -F pyo3_backend` or add `pyo3_backend` to src/script/Cargo.toml 's `features.default` like `default = ["python", "pyo3_backend]`)): this install a Python shared library required for running Python
-  scripting engine(In CPython Mode). This is available as `python3-dev` on
-  ubuntu, you can install it with `sudo apt install python3-dev`, or
-  `python3-devel` on RPM based distributions (e.g. Fedora, Red Hat, SuSE). Mac's
-  `Python3` package should have this shared library by default. More detail for compiling with PyO3 can be found in [PyO3](https://pyo3.rs/v0.18.1/building_and_distribution#configuring-the-python-version)'s documentation.
 
-#### Build with Docker
+```shell
-
+docker pull greptime/greptimedb
-A docker image with necessary dependencies is provided:
-
-```
-docker build --network host -f docker/Dockerfile -t greptimedb .
 ```
 
-### Run
+Start a GreptimeDB container with:
-
-Start GreptimeDB from source code, in standalone mode:
 
+```shell
+docker run --rm --name greptime --net=host greptime/greptimedb standalone start
 ```
+
+Read more about [Installation](https://docs.greptime.com/getting-started/installation/overview) on docs.
+
+## Getting Started
+
+* [Quickstart](https://docs.greptime.com/getting-started/quick-start/overview)
+* [Write Data](https://docs.greptime.com/user-guide/clients/overview)
+* [Query Data](https://docs.greptime.com/user-guide/query-data/overview)
+* [Operations](https://docs.greptime.com/user-guide/operations/overview)
+
+## Build
+
+Check the prerequisite:
+
+* [Rust toolchain](https://www.rust-lang.org/tools/install) (nightly)
+* [Protobuf compiler](https://grpc.io/docs/protoc-installation/) (>= 3.15)
+* Python toolchain (optional): Required only if built with PyO3 backend. More detail for compiling with PyO3 can be found in its [documentation](https://pyo3.rs/v0.18.1/building_and_distribution#configuring-the-python-version).
+
+Build GreptimeDB binary:
+
+```shell
+make
+```
+
+Run a standalone server:
+
+```shell
 cargo run -- standalone start
 ```
 
-Or if you built from docker:
+## Extension
-
-```
-docker run -p 4002:4002 -v "$(pwd):/tmp/greptimedb" greptime/greptimedb standalone start
-```
-
-Please see the online document site for more installation options and [operations info](https://docs.greptime.com/user-guide/operations/overview).
-
-### Get started
-
-Read the [complete getting started guide](https://docs.greptime.com/getting-started/overview) on our [official document site](https://docs.greptime.com/).
-
-To write and query data, GreptimeDB is compatible with multiple [protocols and clients](https://docs.greptime.com/user-guide/clients/overview).
-
-## Resources
-
-### Installation
-
-- [Pre-built Binaries](https://greptime.com/download):
-  For Linux and macOS, you can easily download pre-built binaries including official releases and nightly builds that are ready to use.
-  In most cases, downloading the version without PyO3 is sufficient. However, if you plan to run scripts in CPython (and use Python packages like NumPy and Pandas), you will need to download the version with PyO3 and install a Python with the same version as the Python in the PyO3 version.
-  We recommend using virtualenv for the installation process to manage multiple Python versions.
-- [Docker Images](https://hub.docker.com/r/greptime/greptimedb)(**recommended**): pre-built
-  Docker images, this is the easiest way to try GreptimeDB. By default it runs CPython script with `pyo3_backend` enabled.
-- [`gtctl`](https://github.com/GreptimeTeam/gtctl): the command-line tool for
-  Kubernetes deployment
-
-### Documentation
-
-- GreptimeDB [User Guide](https://docs.greptime.com/user-guide/concepts/overview)
-- GreptimeDB [Developer
-  Guide](https://docs.greptime.com/developer-guide/overview.html)
-- GreptimeDB [internal code document](https://greptimedb.rs)
 
 ### Dashboard
+
 - [The dashboard UI for GreptimeDB](https://github.com/GreptimeTeam/dashboard)
 
 ### SDK
 
-- [GreptimeDB C++ Client](https://github.com/GreptimeTeam/greptimedb-client-cpp)
+- [GreptimeDB Go Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-go)
-- [GreptimeDB Erlang Client](https://github.com/GreptimeTeam/greptimedb-client-erl)
+- [GreptimeDB Java Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-java)
-- [GreptimeDB Go Client](https://github.com/GreptimeTeam/greptimedb-client-go)
+- [GreptimeDB C++ Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-cpp)
-- [GreptimeDB Java Client](https://github.com/GreptimeTeam/greptimedb-client-java)
+- [GreptimeDB Erlang Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-erl)
-- [GreptimeDB Python Client](https://github.com/GreptimeTeam/greptimedb-client-py) (WIP)
+- [GreptimeDB Rust Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-rust)
-- [GreptimeDB Rust Client](https://github.com/GreptimeTeam/greptimedb-client-rust)
+- [GreptimeDB JavaScript Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-js)
-- [GreptimeDB JavaScript Client](https://github.com/GreptimeTeam/greptime-js-sdk)
+
+### Grafana Dashboard
+
+Our official Grafana dashboard is available at [grafana](grafana/README.md) directory.
 
 ## Project Status
 
-This project is in its early stage and under heavy development. We move fast and
+The current version has not yet reached General Availability version standards.
-break things. Benchmark on development branch may not represent its potential
+In line with our Greptime 2024 Roadmap, we plan to achieve a production-level
-performance. We release pre-built binaries constantly for functional
+version with the update to v1.0 in August. [[Join Force]](https://github.com/GreptimeTeam/greptimedb/issues/3412)
-evaluation. Do not use it in production at the moment.
-
-For future plans, check out [GreptimeDB roadmap](https://github.com/GreptimeTeam/greptimedb/issues/669).
 
 ## Community
 
@@ -150,29 +163,28 @@ and what went wrong. If you have any questions or if you would like to get invol
 community, please check out:
 
 - GreptimeDB Community on [Slack](https://greptime.com/slack)
-- GreptimeDB GitHub [Discussions](https://github.com/GreptimeTeam/greptimedb/discussions)
+- GreptimeDB [GitHub Discussions forum](https://github.com/GreptimeTeam/greptimedb/discussions)
-- Greptime official [Website](https://greptime.com)
+- Greptime official [website](https://greptime.com)
 
 In addition, you may:
 
-- View our official [Blog](https://greptime.com/blogs/index)
+- View our official [Blog](https://greptime.com/blogs/)
 - Connect us with [Linkedin](https://www.linkedin.com/company/greptime/)
 - Follow us on [Twitter](https://twitter.com/greptime)
 
 ## License
 
-GreptimeDB uses the [Apache 2.0 license][1] to strike a balance between
+GreptimeDB uses the [Apache License 2.0](https://apache.org/licenses/LICENSE-2.0.txt) to strike a balance between
 open contributions and allowing you to use the software however you want.
 
-[1]: <https://github.com/greptimeTeam/greptimedb/blob/main/LICENSE>
-
 ## Contributing
 
-Please refer to [contribution guidelines](CONTRIBUTING.md) for more information.
+Please refer to [contribution guidelines](CONTRIBUTING.md) and [internal concepts docs](https://docs.greptime.com/contributor-guide/overview.html) for more information.
 
 ## Acknowledgement
-- GreptimeDB uses [Apache Arrow](https://arrow.apache.org/) as the memory model and [Apache Parquet](https://parquet.apache.org/) as the persistent file format.
+
-- GreptimeDB's query engine is powered by [Apache Arrow DataFusion](https://github.com/apache/arrow-datafusion).
+- GreptimeDB uses [Apache Arrow™](https://arrow.apache.org/) as the memory model and [Apache Parquet™](https://parquet.apache.org/) as the persistent file format.
-- [Apache OpenDAL (incubating)](https://opendal.apache.org) gives GreptimeDB a very general and elegant data access abstraction layer.
+- GreptimeDB's query engine is powered by [Apache Arrow DataFusion™](https://arrow.apache.org/datafusion/).
+- [Apache OpenDAL™](https://opendal.apache.org) gives GreptimeDB a very general and elegant data access abstraction layer.
 - GreptimeDB's meta service is based on [etcd](https://etcd.io/).
 - GreptimeDB uses [RustPython](https://github.com/RustPython/RustPython) for experimental embedded python scripting.

benchmarks/Cargo.toml

@@ -4,13 +4,35 @@ version.workspace = true
 edition.workspace = true
 license.workspace = true
 
+[lints]
+workspace = true
+
 [dependencies]
+api.workspace = true
 arrow.workspace = true
 chrono.workspace = true
-clap = { version = "4.0", features = ["derive"] }
+clap.workspace = true
 client.workspace = true
+common-base.workspace = true
+common-telemetry.workspace = true
+common-wal.workspace = true
+dotenv.workspace = true
+futures.workspace = true
 futures-util.workspace = true
+humantime.workspace = true
+humantime-serde.workspace = true
 indicatif = "0.17.1"
 itertools.workspace = true
+lazy_static.workspace = true
+log-store.workspace = true
+mito2.workspace = true
+num_cpus.workspace = true
 parquet.workspace = true
+prometheus.workspace = true
+rand.workspace = true
+rskafka.workspace = true
+serde.workspace = true
+store-api.workspace = true
 tokio.workspace = true
+toml.workspace = true
+uuid.workspace = true

benchmarks/README.md

@@ -0,0 +1,11 @@
+Benchmarkers for GreptimeDB
+--------------------------------
+
+## Wal Benchmarker
+The wal benchmarker serves to evaluate the performance of GreptimeDB's Write-Ahead Log (WAL) component. It meticulously assesses the read/write performance of the WAL under diverse workloads generated by the benchmarker. 
+
+
+### How to use
+To compile the benchmarker, navigate to the `greptimedb/benchmarks` directory and execute `cargo build --release`. Subsequently, you'll find the compiled target located at `greptimedb/target/release/wal_bench`.
+
+The `./wal_bench -h` command reveals numerous arguments that the target accepts. Among these, a notable one is the `cfg-file` argument. By utilizing a configuration file in the TOML format, you can bypass the need to repeatedly specify cumbersome arguments.

benchmarks/config/wal_bench.example.toml

@@ -0,0 +1,21 @@
+# Refers to the documents of `Args` in benchmarks/src/wal.rs`.
+wal_provider = "kafka"
+bootstrap_brokers = ["localhost:9092"]
+num_workers = 10
+num_topics = 32
+num_regions = 1000
+num_scrapes = 1000
+num_rows = 5
+col_types = "ifs"
+max_batch_size = "512KB"
+linger = "1ms"
+backoff_init = "10ms"
+backoff_max = "1ms"
+backoff_base = 2
+backoff_deadline = "3s"
+compression = "zstd"
+rng_seed = 42
+skip_read = false
+skip_write = false
+random_topics = true
+report_metrics = false

benchmarks/src/bin/nyc-taxi.rs

@@ -29,7 +29,7 @@ use client::api::v1::column::Values;
 use client::api::v1::{
     Column, ColumnDataType, ColumnDef, CreateTableExpr, InsertRequest, InsertRequests, SemanticType,
 };
-use client::{Client, Database, Output, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
+use client::{Client, Database, OutputData, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
 use futures_util::TryStreamExt;
 use indicatif::{MultiProgress, ProgressBar, ProgressStyle};
 use parquet::arrow::arrow_reader::ParquetRecordBatchReaderBuilder;
@@ -215,37 +215,7 @@ fn build_values(column: &ArrayRef) -> (Values, ColumnDataType) {
                 ColumnDataType::String,
             )
         }
-        DataType::Null
+        _ => unimplemented!(),
-        | DataType::Boolean
-        | DataType::Int8
-        | DataType::Int16
-        | DataType::Int32
-        | DataType::UInt8
-        | DataType::UInt16
-        | DataType::UInt32
-        | DataType::UInt64
-        | DataType::Float16
-        | DataType::Float32
-        | DataType::Date32
-        | DataType::Date64
-        | DataType::Time32(_)
-        | DataType::Time64(_)
-        | DataType::Duration(_)
-        | DataType::Interval(_)
-        | DataType::Binary
-        | DataType::FixedSizeBinary(_)
-        | DataType::LargeBinary
-        | DataType::LargeUtf8
-        | DataType::List(_)
-        | DataType::FixedSizeList(_, _)
-        | DataType::LargeList(_)
-        | DataType::Struct(_)
-        | DataType::Union(_, _)
-        | DataType::Dictionary(_, _)
-        | DataType::Decimal128(_, _)
-        | DataType::Decimal256(_, _)
-        | DataType::RunEndEncoded(_, _)
-        | DataType::Map(_, _) => todo!(),
     }
 }
 
@@ -258,7 +228,7 @@ fn create_table_expr(table_name: &str) -> CreateTableExpr {
         catalog_name: CATALOG_NAME.to_string(),
         schema_name: SCHEMA_NAME.to_string(),
         table_name: table_name.to_string(),
-        desc: "".to_string(),
+        desc: String::default(),
         column_defs: vec![
             ColumnDef {
                 name: "VendorID".to_string(),
@@ -502,9 +472,9 @@ async fn do_query(num_iter: usize, db: &Database, table_name: &str) {
         for i in 0..num_iter {
             let now = Instant::now();
             let res = db.sql(&query).await.unwrap();
-            match res {
+            match res.data {
-                Output::AffectedRows(_) | Output::RecordBatches(_) => (),
+                OutputData::AffectedRows(_) | OutputData::RecordBatches(_) => (),
-                Output::Stream(stream) => {
+                OutputData::Stream(stream) => {
                     stream.try_collect::<Vec<_>>().await.unwrap();
                 }
             }

benchmarks/src/bin/wal_bench.rs

@@ -0,0 +1,326 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#![feature(int_roundings)]
+
+use std::fs;
+use std::sync::Arc;
+use std::time::Instant;
+
+use api::v1::{ColumnDataType, ColumnSchema, SemanticType};
+use benchmarks::metrics;
+use benchmarks::wal_bench::{Args, Config, Region, WalProvider};
+use clap::Parser;
+use common_telemetry::info;
+use common_wal::config::kafka::common::BackoffConfig;
+use common_wal::config::kafka::DatanodeKafkaConfig as KafkaConfig;
+use common_wal::config::raft_engine::RaftEngineConfig;
+use common_wal::options::{KafkaWalOptions, WalOptions};
+use itertools::Itertools;
+use log_store::kafka::log_store::KafkaLogStore;
+use log_store::raft_engine::log_store::RaftEngineLogStore;
+use mito2::wal::Wal;
+use prometheus::{Encoder, TextEncoder};
+use rand::distributions::{Alphanumeric, DistString};
+use rand::rngs::SmallRng;
+use rand::SeedableRng;
+use rskafka::client::partition::Compression;
+use rskafka::client::ClientBuilder;
+use store_api::logstore::LogStore;
+use store_api::storage::RegionId;
+
+async fn run_benchmarker<S: LogStore>(cfg: &Config, topics: &[String], wal: Arc<Wal<S>>) {
+    let chunk_size = cfg.num_regions.div_ceil(cfg.num_workers);
+    let region_chunks = (0..cfg.num_regions)
+        .map(|id| {
+            build_region(
+                id as u64,
+                topics,
+                &mut SmallRng::seed_from_u64(cfg.rng_seed),
+                cfg,
+            )
+        })
+        .chunks(chunk_size as usize)
+        .into_iter()
+        .map(|chunk| Arc::new(chunk.collect::<Vec<_>>()))
+        .collect::<Vec<_>>();
+
+    let mut write_elapsed = 0;
+    let mut read_elapsed = 0;
+
+    if !cfg.skip_write {
+        info!("Benchmarking write ...");
+
+        let num_scrapes = cfg.num_scrapes;
+        let timer = Instant::now();
+        futures::future::join_all((0..cfg.num_workers).map(|i| {
+            let wal = wal.clone();
+            let regions = region_chunks[i as usize].clone();
+            tokio::spawn(async move {
+                for _ in 0..num_scrapes {
+                    let mut wal_writer = wal.writer();
+                    regions
+                        .iter()
+                        .for_each(|region| region.add_wal_entry(&mut wal_writer));
+                    wal_writer.write_to_wal().await.unwrap();
+                }
+            })
+        }))
+        .await;
+        write_elapsed += timer.elapsed().as_millis();
+    }
+
+    if !cfg.skip_read {
+        info!("Benchmarking read ...");
+
+        let timer = Instant::now();
+        futures::future::join_all((0..cfg.num_workers).map(|i| {
+            let wal = wal.clone();
+            let regions = region_chunks[i as usize].clone();
+            tokio::spawn(async move {
+                for region in regions.iter() {
+                    region.replay(&wal).await;
+                }
+            })
+        }))
+        .await;
+        read_elapsed = timer.elapsed().as_millis();
+    }
+
+    dump_report(cfg, write_elapsed, read_elapsed);
+}
+
+fn build_region(id: u64, topics: &[String], rng: &mut SmallRng, cfg: &Config) -> Region {
+    let wal_options = match cfg.wal_provider {
+        WalProvider::Kafka => {
+            assert!(!topics.is_empty());
+            WalOptions::Kafka(KafkaWalOptions {
+                topic: topics.get(id as usize % topics.len()).cloned().unwrap(),
+            })
+        }
+        WalProvider::RaftEngine => WalOptions::RaftEngine,
+    };
+    Region::new(
+        RegionId::from_u64(id),
+        build_schema(&parse_col_types(&cfg.col_types), rng),
+        wal_options,
+        cfg.num_rows,
+        cfg.rng_seed,
+    )
+}
+
+fn build_schema(col_types: &[ColumnDataType], mut rng: &mut SmallRng) -> Vec<ColumnSchema> {
+    col_types
+        .iter()
+        .map(|col_type| ColumnSchema {
+            column_name: Alphanumeric.sample_string(&mut rng, 5),
+            datatype: *col_type as i32,
+            semantic_type: SemanticType::Field as i32,
+            datatype_extension: None,
+        })
+        .chain(vec![ColumnSchema {
+            column_name: "ts".to_string(),
+            datatype: ColumnDataType::TimestampMillisecond as i32,
+            semantic_type: SemanticType::Tag as i32,
+            datatype_extension: None,
+        }])
+        .collect()
+}
+
+fn dump_report(cfg: &Config, write_elapsed: u128, read_elapsed: u128) {
+    let cost_report = format!(
+        "write costs: {} ms, read costs: {} ms",
+        write_elapsed, read_elapsed,
+    );
+
+    let total_written_bytes = metrics::METRIC_WAL_WRITE_BYTES_TOTAL.get() as u128;
+    let write_throughput = if write_elapsed > 0 {
+        (total_written_bytes * 1000).div_floor(write_elapsed)
+    } else {
+        0
+    };
+    let total_read_bytes = metrics::METRIC_WAL_READ_BYTES_TOTAL.get() as u128;
+    let read_throughput = if read_elapsed > 0 {
+        (total_read_bytes * 1000).div_floor(read_elapsed)
+    } else {
+        0
+    };
+
+    let throughput_report = format!(
+        "total written bytes: {} bytes, total read bytes: {} bytes, write throuput: {} bytes/s ({} mb/s), read throughput: {} bytes/s ({} mb/s)",
+        total_written_bytes,
+        total_read_bytes,
+        write_throughput,
+        write_throughput.div_floor(1 << 20),
+        read_throughput,
+        read_throughput.div_floor(1 << 20),
+    );
+
+    let metrics_report = if cfg.report_metrics {
+        let mut buffer = Vec::new();
+        let encoder = TextEncoder::new();
+        let metrics = prometheus::gather();
+        encoder.encode(&metrics, &mut buffer).unwrap();
+        String::from_utf8(buffer).unwrap()
+    } else {
+        String::new()
+    };
+
+    info!(
+        r#"
+Benchmark config: 
+{cfg:?}
+
+Benchmark report:
+{cost_report}
+{throughput_report}
+{metrics_report}"#
+    );
+}
+
+async fn create_topics(cfg: &Config) -> Vec<String> {
+    // Creates topics.
+    let client = ClientBuilder::new(cfg.bootstrap_brokers.clone())
+        .build()
+        .await
+        .unwrap();
+    let ctrl_client = client.controller_client().unwrap();
+    let (topics, tasks): (Vec<_>, Vec<_>) = (0..cfg.num_topics)
+        .map(|i| {
+            let topic = if cfg.random_topics {
+                format!(
+                    "greptime_wal_bench_topic_{}_{}",
+                    uuid::Uuid::new_v4().as_u128(),
+                    i
+                )
+            } else {
+                format!("greptime_wal_bench_topic_{}", i)
+            };
+            let task = ctrl_client.create_topic(
+                topic.clone(),
+                1,
+                cfg.bootstrap_brokers.len() as i16,
+                2000,
+            );
+            (topic, task)
+        })
+        .unzip();
+    // Must ignore errors since we allow topics being created more than once.
+    let _ = futures::future::try_join_all(tasks).await;
+
+    topics
+}
+
+fn parse_compression(comp: &str) -> Compression {
+    match comp {
+        "no" => Compression::NoCompression,
+        "gzip" => Compression::Gzip,
+        "lz4" => Compression::Lz4,
+        "snappy" => Compression::Snappy,
+        "zstd" => Compression::Zstd,
+        other => unreachable!("Unrecognized compression {other}"),
+    }
+}
+
+fn parse_col_types(col_types: &str) -> Vec<ColumnDataType> {
+    let parts = col_types.split('x').collect::<Vec<_>>();
+    assert!(parts.len() <= 2);
+
+    let pattern = parts[0];
+    let repeat = parts
+        .get(1)
+        .map(|r| r.parse::<usize>().unwrap())
+        .unwrap_or(1);
+
+    pattern
+        .chars()
+        .map(|c| match c {
+            'i' | 'I' => ColumnDataType::Int64,
+            'f' | 'F' => ColumnDataType::Float64,
+            's' | 'S' => ColumnDataType::String,
+            other => unreachable!("Cannot parse {other} as a column data type"),
+        })
+        .cycle()
+        .take(pattern.len() * repeat)
+        .collect()
+}
+
+fn main() {
+    // Sets the global logging to INFO and suppress loggings from rskafka other than ERROR and upper ones.
+    std::env::set_var("UNITTEST_LOG_LEVEL", "info,rskafka=error");
+    common_telemetry::init_default_ut_logging();
+
+    let args = Args::parse();
+    let cfg = if !args.cfg_file.is_empty() {
+        toml::from_str(&fs::read_to_string(&args.cfg_file).unwrap()).unwrap()
+    } else {
+        Config::from(args)
+    };
+
+    // Validates arguments.
+    if cfg.num_regions < cfg.num_workers {
+        panic!("num_regions must be greater than or equal to num_workers");
+    }
+    if cfg
+        .num_workers
+        .min(cfg.num_topics)
+        .min(cfg.num_regions)
+        .min(cfg.num_scrapes)
+        .min(cfg.max_batch_size.as_bytes() as u32)
+        .min(cfg.bootstrap_brokers.len() as u32)
+        == 0
+    {
+        panic!("Invalid arguments");
+    }
+
+    tokio::runtime::Builder::new_multi_thread()
+        .enable_all()
+        .build()
+        .unwrap()
+        .block_on(async {
+            match cfg.wal_provider {
+                WalProvider::Kafka => {
+                    let topics = create_topics(&cfg).await;
+                    let kafka_cfg = KafkaConfig {
+                        broker_endpoints: cfg.bootstrap_brokers.clone(),
+                        max_batch_size: cfg.max_batch_size,
+                        linger: cfg.linger,
+                        backoff: BackoffConfig {
+                            init: cfg.backoff_init,
+                            max: cfg.backoff_max,
+                            base: cfg.backoff_base,
+                            deadline: Some(cfg.backoff_deadline),
+                        },
+                        compression: parse_compression(&cfg.compression),
+                        ..Default::default()
+                    };
+                    let store = Arc::new(KafkaLogStore::try_new(&kafka_cfg).await.unwrap());
+                    let wal = Arc::new(Wal::new(store));
+                    run_benchmarker(&cfg, &topics, wal).await;
+                }
+                WalProvider::RaftEngine => {
+                    // The benchmarker assumes the raft engine directory exists.
+                    let store = RaftEngineLogStore::try_new(
+                        "/tmp/greptimedb/raft-engine-wal".to_string(),
+                        RaftEngineConfig::default(),
+                    )
+                    .await
+                    .map(Arc::new)
+                    .unwrap();
+                    let wal = Arc::new(Wal::new(store));
+                    run_benchmarker(&cfg, &[], wal).await;
+                }
+            }
+        });
+}

benchmarks/src/lib.rs

@@ -0,0 +1,16 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+pub mod metrics;
+pub mod wal_bench;

benchmarks/src/metrics.rs

@@ -0,0 +1,39 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use lazy_static::lazy_static;
+use prometheus::*;
+
+/// Logstore label.
+pub const LOGSTORE_LABEL: &str = "logstore";
+/// Operation type label.
+pub const OPTYPE_LABEL: &str = "optype";
+
+lazy_static! {
+    /// Counters of bytes of each operation on a logstore.
+    pub static ref METRIC_WAL_OP_BYTES_TOTAL: IntCounterVec = register_int_counter_vec!(
+        "greptime_bench_wal_op_bytes_total",
+        "wal operation bytes total",
+        &[OPTYPE_LABEL],
+    )
+    .unwrap();
+    /// Counter of bytes of the append_batch operation.
+    pub static ref METRIC_WAL_WRITE_BYTES_TOTAL: IntCounter = METRIC_WAL_OP_BYTES_TOTAL.with_label_values(
+        &["write"],
+    );
+    /// Counter of bytes of the read operation.
+    pub static ref METRIC_WAL_READ_BYTES_TOTAL: IntCounter = METRIC_WAL_OP_BYTES_TOTAL.with_label_values(
+        &["read"],
+    );
+}

benchmarks/src/wal_bench.rs

@@ -0,0 +1,361 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::mem::size_of;
+use std::sync::atomic::{AtomicI64, AtomicU64, Ordering};
+use std::sync::{Arc, Mutex};
+use std::time::Duration;
+
+use api::v1::value::ValueData;
+use api::v1::{ColumnDataType, ColumnSchema, Mutation, OpType, Row, Rows, Value, WalEntry};
+use clap::{Parser, ValueEnum};
+use common_base::readable_size::ReadableSize;
+use common_wal::options::WalOptions;
+use futures::StreamExt;
+use mito2::wal::{Wal, WalWriter};
+use rand::distributions::{Alphanumeric, DistString, Uniform};
+use rand::rngs::SmallRng;
+use rand::{Rng, SeedableRng};
+use serde::{Deserialize, Serialize};
+use store_api::logstore::LogStore;
+use store_api::storage::RegionId;
+
+use crate::metrics;
+
+/// The wal provider.
+#[derive(Clone, ValueEnum, Default, Debug, PartialEq, Serialize, Deserialize)]
+#[serde(rename_all = "snake_case")]
+pub enum WalProvider {
+    #[default]
+    RaftEngine,
+    Kafka,
+}
+
+#[derive(Parser)]
+pub struct Args {
+    /// The provided configuration file.
+    /// The example configuration file can be found at `greptimedb/benchmarks/config/wal_bench.example.toml`.
+    #[clap(long, short = 'c')]
+    pub cfg_file: String,
+
+    /// The wal provider.
+    #[clap(long, value_enum, default_value_t = WalProvider::default())]
+    pub wal_provider: WalProvider,
+
+    /// The advertised addresses of the kafka brokers.
+    /// If there're multiple bootstrap brokers, their addresses should be separated by comma, for e.g. "localhost:9092,localhost:9093".
+    #[clap(long, short = 'b', default_value = "localhost:9092")]
+    pub bootstrap_brokers: String,
+
+    /// The number of workers each running in a dedicated thread.
+    #[clap(long, default_value_t = num_cpus::get() as u32)]
+    pub num_workers: u32,
+
+    /// The number of kafka topics to be created.
+    #[clap(long, default_value_t = 32)]
+    pub num_topics: u32,
+
+    /// The number of regions.
+    #[clap(long, default_value_t = 1000)]
+    pub num_regions: u32,
+
+    /// The number of times each region is scraped.
+    #[clap(long, default_value_t = 1000)]
+    pub num_scrapes: u32,
+
+    /// The number of rows in each wal entry.
+    /// Each time a region is scraped, a wal entry containing will be produced.
+    #[clap(long, default_value_t = 5)]
+    pub num_rows: u32,
+
+    /// The column types of the schema for each region.
+    /// Currently, three column types are supported:
+    /// - i = ColumnDataType::Int64
+    /// - f = ColumnDataType::Float64
+    /// - s = ColumnDataType::String  
+    /// For e.g., "ifs" will be parsed as three columns: i64, f64, and string.
+    ///
+    /// Additionally, a "x" sign can be provided to repeat the column types for a given number of times.
+    /// For e.g., "iix2" will be parsed as 4 columns: i64, i64, i64, and i64.
+    /// This feature is useful if you want to specify many columns.
+    #[clap(long, default_value = "ifs")]
+    pub col_types: String,
+
+    /// The maximum size of a batch of kafka records.
+    /// The default value is 1mb.
+    #[clap(long, default_value = "512KB")]
+    pub max_batch_size: ReadableSize,
+
+    /// The minimum latency the kafka client issues a batch of kafka records.
+    /// However, a batch of kafka records would be immediately issued if a record cannot be fit into the batch.
+    #[clap(long, default_value = "1ms")]
+    pub linger: String,
+
+    /// The initial backoff delay of the kafka consumer.
+    #[clap(long, default_value = "10ms")]
+    pub backoff_init: String,
+
+    /// The maximum backoff delay of the kafka consumer.
+    #[clap(long, default_value = "1s")]
+    pub backoff_max: String,
+
+    /// The exponential backoff rate of the kafka consumer. The next back off = base * the current backoff.
+    #[clap(long, default_value_t = 2)]
+    pub backoff_base: u32,
+
+    /// The deadline of backoff. The backoff ends if the total backoff delay reaches the deadline.
+    #[clap(long, default_value = "3s")]
+    pub backoff_deadline: String,
+
+    /// The client-side compression algorithm for kafka records.
+    #[clap(long, default_value = "zstd")]
+    pub compression: String,
+
+    /// The seed of random number generators.
+    #[clap(long, default_value_t = 42)]
+    pub rng_seed: u64,
+
+    /// Skips the read phase, aka. region replay, if set to true.
+    #[clap(long, default_value_t = false)]
+    pub skip_read: bool,
+
+    /// Skips the write phase if set to true.
+    #[clap(long, default_value_t = false)]
+    pub skip_write: bool,
+
+    /// Randomly generates topic names if set to true.
+    /// Useful when you want to run the benchmarker without worrying about the topics created before.
+    #[clap(long, default_value_t = false)]
+    pub random_topics: bool,
+
+    /// Logs out the gathered prometheus metrics when the benchmarker ends.
+    #[clap(long, default_value_t = false)]
+    pub report_metrics: bool,
+}
+
+/// Benchmarker config.
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Config {
+    pub wal_provider: WalProvider,
+    pub bootstrap_brokers: Vec<String>,
+    pub num_workers: u32,
+    pub num_topics: u32,
+    pub num_regions: u32,
+    pub num_scrapes: u32,
+    pub num_rows: u32,
+    pub col_types: String,
+    pub max_batch_size: ReadableSize,
+    #[serde(with = "humantime_serde")]
+    pub linger: Duration,
+    #[serde(with = "humantime_serde")]
+    pub backoff_init: Duration,
+    #[serde(with = "humantime_serde")]
+    pub backoff_max: Duration,
+    pub backoff_base: u32,
+    #[serde(with = "humantime_serde")]
+    pub backoff_deadline: Duration,
+    pub compression: String,
+    pub rng_seed: u64,
+    pub skip_read: bool,
+    pub skip_write: bool,
+    pub random_topics: bool,
+    pub report_metrics: bool,
+}
+
+impl From<Args> for Config {
+    fn from(args: Args) -> Self {
+        let cfg = Self {
+            wal_provider: args.wal_provider,
+            bootstrap_brokers: args
+                .bootstrap_brokers
+                .split(',')
+                .map(ToString::to_string)
+                .collect::<Vec<_>>(),
+            num_workers: args.num_workers.min(num_cpus::get() as u32),
+            num_topics: args.num_topics,
+            num_regions: args.num_regions,
+            num_scrapes: args.num_scrapes,
+            num_rows: args.num_rows,
+            col_types: args.col_types,
+            max_batch_size: args.max_batch_size,
+            linger: humantime::parse_duration(&args.linger).unwrap(),
+            backoff_init: humantime::parse_duration(&args.backoff_init).unwrap(),
+            backoff_max: humantime::parse_duration(&args.backoff_max).unwrap(),
+            backoff_base: args.backoff_base,
+            backoff_deadline: humantime::parse_duration(&args.backoff_deadline).unwrap(),
+            compression: args.compression,
+            rng_seed: args.rng_seed,
+            skip_read: args.skip_read,
+            skip_write: args.skip_write,
+            random_topics: args.random_topics,
+            report_metrics: args.report_metrics,
+        };
+
+        cfg
+    }
+}
+
+/// The region used for wal benchmarker.
+pub struct Region {
+    id: RegionId,
+    schema: Vec<ColumnSchema>,
+    wal_options: WalOptions,
+    next_sequence: AtomicU64,
+    next_entry_id: AtomicU64,
+    next_timestamp: AtomicI64,
+    rng: Mutex<Option<SmallRng>>,
+    num_rows: u32,
+}
+
+impl Region {
+    /// Creates a new region.
+    pub fn new(
+        id: RegionId,
+        schema: Vec<ColumnSchema>,
+        wal_options: WalOptions,
+        num_rows: u32,
+        rng_seed: u64,
+    ) -> Self {
+        Self {
+            id,
+            schema,
+            wal_options,
+            next_sequence: AtomicU64::new(1),
+            next_entry_id: AtomicU64::new(1),
+            next_timestamp: AtomicI64::new(1655276557000),
+            rng: Mutex::new(Some(SmallRng::seed_from_u64(rng_seed))),
+            num_rows,
+        }
+    }
+
+    /// Scrapes the region and adds the generated entry to wal.
+    pub fn add_wal_entry<S: LogStore>(&self, wal_writer: &mut WalWriter<S>) {
+        let mutation = Mutation {
+            op_type: OpType::Put as i32,
+            sequence: self
+                .next_sequence
+                .fetch_add(self.num_rows as u64, Ordering::Relaxed),
+            rows: Some(self.build_rows()),
+        };
+        let entry = WalEntry {
+            mutations: vec![mutation],
+        };
+        metrics::METRIC_WAL_WRITE_BYTES_TOTAL.inc_by(Self::entry_estimated_size(&entry) as u64);
+
+        wal_writer
+            .add_entry(
+                self.id,
+                self.next_entry_id.fetch_add(1, Ordering::Relaxed),
+                &entry,
+                &self.wal_options,
+            )
+            .unwrap();
+    }
+
+    /// Replays the region.
+    pub async fn replay<S: LogStore>(&self, wal: &Arc<Wal<S>>) {
+        let mut wal_stream = wal.scan(self.id, 0, &self.wal_options).unwrap();
+        while let Some(res) = wal_stream.next().await {
+            let (_, entry) = res.unwrap();
+            metrics::METRIC_WAL_READ_BYTES_TOTAL.inc_by(Self::entry_estimated_size(&entry) as u64);
+        }
+    }
+
+    /// Computes the estimated size in bytes of the entry.
+    pub fn entry_estimated_size(entry: &WalEntry) -> usize {
+        let wrapper_size = size_of::<WalEntry>()
+            + entry.mutations.capacity() * size_of::<Mutation>()
+            + size_of::<Rows>();
+
+        let rows = entry.mutations[0].rows.as_ref().unwrap();
+
+        let schema_size = rows.schema.capacity() * size_of::<ColumnSchema>()
+            + rows
+                .schema
+                .iter()
+                .map(|s| s.column_name.capacity())
+                .sum::<usize>();
+        let values_size = (rows.rows.capacity() * size_of::<Row>())
+            + rows
+                .rows
+                .iter()
+                .map(|r| r.values.capacity() * size_of::<Value>())
+                .sum::<usize>();
+
+        wrapper_size + schema_size + values_size
+    }
+
+    fn build_rows(&self) -> Rows {
+        let cols = self
+            .schema
+            .iter()
+            .map(|col_schema| {
+                let col_data_type = ColumnDataType::try_from(col_schema.datatype).unwrap();
+                self.build_col(&col_data_type, self.num_rows)
+            })
+            .collect::<Vec<_>>();
+
+        let rows = (0..self.num_rows)
+            .map(|i| {
+                let values = cols.iter().map(|col| col[i as usize].clone()).collect();
+                Row { values }
+            })
+            .collect();
+
+        Rows {
+            schema: self.schema.clone(),
+            rows,
+        }
+    }
+
+    fn build_col(&self, col_data_type: &ColumnDataType, num_rows: u32) -> Vec<Value> {
+        let mut rng_guard = self.rng.lock().unwrap();
+        let rng = rng_guard.as_mut().unwrap();
+        match col_data_type {
+            ColumnDataType::TimestampMillisecond => (0..num_rows)
+                .map(|_| {
+                    let ts = self.next_timestamp.fetch_add(1000, Ordering::Relaxed);
+                    Value {
+                        value_data: Some(ValueData::TimestampMillisecondValue(ts)),
+                    }
+                })
+                .collect(),
+            ColumnDataType::Int64 => (0..num_rows)
+                .map(|_| {
+                    let v = rng.sample(Uniform::new(0, 10_000));
+                    Value {
+                        value_data: Some(ValueData::I64Value(v)),
+                    }
+                })
+                .collect(),
+            ColumnDataType::Float64 => (0..num_rows)
+                .map(|_| {
+                    let v = rng.sample(Uniform::new(0.0, 5000.0));
+                    Value {
+                        value_data: Some(ValueData::F64Value(v)),
+                    }
+                })
+                .collect(),
+            ColumnDataType::String => (0..num_rows)
+                .map(|_| {
+                    let v = Alphanumeric.sample_string(rng, 10);
+                    Value {
+                        value_data: Some(ValueData::StringValue(v)),
+                    }
+                })
+                .collect(),
+            _ => unreachable!(),
+        }
+    }
+}

cliff.toml

@@ -0,0 +1,127 @@
+# https://git-cliff.org/docs/configuration
+
+[remote.github]
+owner = "GreptimeTeam"
+repo = "greptimedb"
+
+[changelog]
+header = ""
+footer = ""
+# template for the changelog body
+# https://keats.github.io/tera/docs/#introduction
+body = """
+# {{ version }}
+
+Release date: {{ timestamp | date(format="%B %d, %Y") }}
+
+{%- set breakings = commits | filter(attribute="breaking", value=true) -%}
+{%- if breakings | length > 0 %}
+
+## Breaking changes
+    {% for commit in breakings %}
+      * {{ commit.github.pr_title }}\
+        {% if commit.github.username %} by \
+          {% set author = commit.github.username -%}
+          [@{{ author }}](https://github.com/{{ author }})
+        {%- endif -%}
+        {% if commit.github.pr_number %} in \
+          {% set number = commit.github.pr_number -%}
+          [#{{ number }}]({{ self::remote_url() }}/pull/{{ number }})
+        {%- endif %}
+    {%- endfor %}
+{%- endif -%}
+
+{%- set grouped_commits = commits | filter(attribute="breaking", value=false) | group_by(attribute="group") -%}
+{% for group, commits in grouped_commits %}
+
+    ### {{ group | striptags | trim | upper_first }}
+    {% for commit in commits %}
+        * {{ commit.github.pr_title }}\
+            {% if commit.github.username %} by \
+              {% set author = commit.github.username -%}
+              [@{{ author }}](https://github.com/{{ author }})
+            {%- endif -%}
+            {% if commit.github.pr_number %} in \
+              {% set number = commit.github.pr_number -%}
+              [#{{ number }}]({{ self::remote_url() }}/pull/{{ number }})
+            {%- endif %}
+    {%- endfor -%}
+{% endfor %}
+
+{%- if github.contributors | filter(attribute="is_first_time", value=true) | length != 0 %}
+  {% raw %}\n{% endraw -%}
+  ## New Contributors
+{% endif -%}
+{% for contributor in github.contributors | filter(attribute="is_first_time", value=true) %}
+  * [@{{ contributor.username }}](https://github.com/{{ contributor.username }}) made their first contribution
+    {%- if contributor.pr_number %} in \
+      [#{{ contributor.pr_number }}]({{ self::remote_url() }}/pull/{{ contributor.pr_number }}) \
+    {%- endif %}
+{%- endfor -%}
+
+{% if github.contributors | length != 0 %}
+  {% raw %}\n{% endraw -%}
+## All Contributors
+
+We would like to thank the following contributors from the GreptimeDB community:
+
+{%- set contributors = github.contributors | sort(attribute="username") | map(attribute="username") -%}
+{%- set bots = ['dependabot[bot]'] %}
+
+{% for contributor in contributors %}
+{%- if bots is containing(contributor) -%}{% continue %}{%- endif -%}
+{%- if loop.first -%}
+  [@{{ contributor }}](https://github.com/{{ contributor }})
+{%- else -%}
+  , [@{{ contributor }}](https://github.com/{{ contributor }})
+{%- endif -%}
+{%- endfor %}
+{%- endif %}
+{% raw %}\n{% endraw %}
+
+{%- macro remote_url() -%}
+  https://github.com/{{ remote.github.owner }}/{{ remote.github.repo }}
+{%- endmacro -%}
+"""
+trim = true
+
+[git]
+# parse the commits based on https://www.conventionalcommits.org
+conventional_commits = true
+# filter out the commits that are not conventional
+filter_unconventional = true
+# process each line of a commit as an individual commit
+split_commits = false
+# regex for parsing and grouping commits
+commit_parsers = [
+  { message = "^feat", group = "<!-- 0 -->🚀 Features" },
+  { message = "^fix", group = "<!-- 1 -->🐛 Bug Fixes" },
+  { message = "^doc", group = "<!-- 3 -->📚 Documentation" },
+  { message = "^perf", group = "<!-- 4 -->⚡ Performance" },
+  { message = "^refactor", group = "<!-- 2 -->🚜 Refactor" },
+  { message = "^style", group = "<!-- 5 -->🎨 Styling" },
+  { message = "^test", group = "<!-- 6 -->🧪 Testing" },
+  { message = "^chore\\(release\\): prepare for", skip = true },
+  { message = "^chore\\(deps.*\\)", skip = true },
+  { message = "^chore\\(pr\\)", skip = true },
+  { message = "^chore\\(pull\\)", skip = true },
+  { message = "^chore|^ci", group = "<!-- 7 -->⚙️ Miscellaneous Tasks" },
+  { body = ".*security", group = "<!-- 8 -->🛡️ Security" },
+  { message = "^revert", group = "<!-- 9 -->◀️ Revert" },
+]
+# protect breaking changes from being skipped due to matching a skipping commit_parser
+protect_breaking_commits = false
+# filter out the commits that are not matched by commit parsers
+filter_commits = false
+# regex for matching git tags
+# tag_pattern = "v[0-9].*"
+# regex for skipping tags
+# skip_tags = ""
+# regex for ignoring tags
+ignore_tags = ".*-nightly-.*"
+# sort the tags topologically
+topo_order = false
+# sort the commits inside sections by oldest/newest order
+sort_commits = "oldest"
+# limit the number of commits included in the changelog.
+# limit_commits = 42

codecov.yml

@@ -8,5 +8,6 @@ coverage:
 ignore:
   - "**/error*.rs" # ignore all error.rs files
   - "tests/runner/*.rs" # ignore integration test runner
+  - "tests-integration/**/*.rs" # ignore integration tests
 comment:                  # this is a top-level key
   layout: "diff"

config/config-docs-template.md

@@ -0,0 +1,19 @@
+# Configurations
+
+## Standalone Mode
+
+{{ toml2docs "./standalone.example.toml" }}
+
+## Cluster Mode
+
+### Frontend
+
+{{ toml2docs "./frontend.example.toml" }}
+
+### Metasrv
+
+{{ toml2docs "./metasrv.example.toml" }}
+
+### Datanode
+
+{{ toml2docs "./datanode.example.toml" }}

config/config.md

@@ -0,0 +1,376 @@
+# Configurations
+
+## Standalone Mode
+
+| Key | Type | Default | Descriptions |
+| --- | -----| ------- | ----------- |
+| `mode` | String | `standalone` | The running mode of the datanode. It can be `standalone` or `distributed`. |
+| `enable_telemetry` | Bool | `true` | Enable telemetry to collect anonymous usage data. |
+| `default_timezone` | String | `None` | The default timezone of the server. |
+| `http` | -- | -- | The HTTP server options. |
+| `http.addr` | String | `127.0.0.1:4000` | The address to bind the HTTP server. |
+| `http.timeout` | String | `30s` | HTTP request timeout. |
+| `http.body_limit` | String | `64MB` | HTTP request body limit.<br/>Support the following units are supported: `B`, `KB`, `KiB`, `MB`, `MiB`, `GB`, `GiB`, `TB`, `TiB`, `PB`, `PiB`. |
+| `grpc` | -- | -- | The gRPC server options. |
+| `grpc.addr` | String | `127.0.0.1:4001` | The address to bind the gRPC server. |
+| `grpc.runtime_size` | Integer | `8` | The number of server worker threads. |
+| `mysql` | -- | -- | MySQL server options. |
+| `mysql.enable` | Bool | `true` | Whether to enable. |
+| `mysql.addr` | String | `127.0.0.1:4002` | The addr to bind the MySQL server. |
+| `mysql.runtime_size` | Integer | `2` | The number of server worker threads. |
+| `mysql.tls` | -- | -- | -- |
+| `mysql.tls.mode` | String | `disable` | TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html<br/>- `disable` (default value)<br/>- `prefer`<br/>- `require`<br/>- `verify-ca`<br/>- `verify-full` |
+| `mysql.tls.cert_path` | String | `None` | Certificate file path. |
+| `mysql.tls.key_path` | String | `None` | Private key file path. |
+| `mysql.tls.watch` | Bool | `false` | Watch for Certificate and key file change and auto reload |
+| `postgres` | -- | -- | PostgresSQL server options. |
+| `postgres.enable` | Bool | `true` | Whether to enable |
+| `postgres.addr` | String | `127.0.0.1:4003` | The addr to bind the PostgresSQL server. |
+| `postgres.runtime_size` | Integer | `2` | The number of server worker threads. |
+| `postgres.tls` | -- | -- | PostgresSQL server TLS options, see `mysql_options.tls` section. |
+| `postgres.tls.mode` | String | `disable` | TLS mode. |
+| `postgres.tls.cert_path` | String | `None` | Certificate file path. |
+| `postgres.tls.key_path` | String | `None` | Private key file path. |
+| `postgres.tls.watch` | Bool | `false` | Watch for Certificate and key file change and auto reload |
+| `opentsdb` | -- | -- | OpenTSDB protocol options. |
+| `opentsdb.enable` | Bool | `true` | Whether to enable |
+| `opentsdb.addr` | String | `127.0.0.1:4242` | OpenTSDB telnet API server address. |
+| `opentsdb.runtime_size` | Integer | `2` | The number of server worker threads. |
+| `influxdb` | -- | -- | InfluxDB protocol options. |
+| `influxdb.enable` | Bool | `true` | Whether to enable InfluxDB protocol in HTTP API. |
+| `prom_store` | -- | -- | Prometheus remote storage options |
+| `prom_store.enable` | Bool | `true` | Whether to enable Prometheus remote write and read in HTTP API. |
+| `prom_store.with_metric_engine` | Bool | `true` | Whether to store the data from Prometheus remote write in metric engine. |
+| `wal` | -- | -- | The WAL options. |
+| `wal.provider` | String | `raft_engine` | The provider of the WAL.<br/>- `raft_engine`: the wal is stored in the local file system by raft-engine.<br/>- `kafka`: it's remote wal that data is stored in Kafka. |
+| `wal.dir` | String | `None` | The directory to store the WAL files.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.file_size` | String | `256MB` | The size of the WAL segment file.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.purge_threshold` | String | `4GB` | The threshold of the WAL size to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.purge_interval` | String | `10m` | The interval to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.read_batch_size` | Integer | `128` | The read batch size.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.sync_write` | Bool | `false` | Whether to use sync write.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.enable_log_recycle` | Bool | `true` | Whether to reuse logically truncated log files.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.prefill_log_files` | Bool | `false` | Whether to pre-create log files on start up.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.sync_period` | String | `10s` | Duration for fsyncing log files.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.broker_endpoints` | Array | -- | The Kafka broker endpoints.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.max_batch_size` | String | `1MB` | The max size of a single producer batch.<br/>Warning: Kafka has a default limit of 1MB per message in a topic.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.linger` | String | `200ms` | The linger duration of a kafka batch producer.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.consumer_wait_timeout` | String | `100ms` | The consumer wait timeout.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_init` | String | `500ms` | The initial backoff delay.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_max` | String | `10s` | The maximum backoff delay.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_base` | Integer | `2` | The exponential backoff rate, i.e. next backoff = base * current backoff.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_deadline` | String | `5mins` | The deadline of retries.<br/>**It's only used when the provider is `kafka`**. |
+| `metadata_store` | -- | -- | Metadata storage options. |
+| `metadata_store.file_size` | String | `256MB` | Kv file size in bytes. |
+| `metadata_store.purge_threshold` | String | `4GB` | Kv purge threshold. |
+| `procedure` | -- | -- | Procedure storage options. |
+| `procedure.max_retry_times` | Integer | `3` | Procedure max retry time. |
+| `procedure.retry_delay` | String | `500ms` | Initial retry delay of procedures, increases exponentially |
+| `storage` | -- | -- | The data storage options. |
+| `storage.data_home` | String | `/tmp/greptimedb/` | The working home directory. |
+| `storage.type` | String | `File` | The storage type used to store the data.<br/>- `File`: the data is stored in the local file system.<br/>- `S3`: the data is stored in the S3 object storage.<br/>- `Gcs`: the data is stored in the Google Cloud Storage.<br/>- `Azblob`: the data is stored in the Azure Blob Storage.<br/>- `Oss`: the data is stored in the Aliyun OSS. |
+| `storage.cache_path` | String | `None` | Cache configuration for object storage such as 'S3' etc.<br/>The local file cache directory. |
+| `storage.cache_capacity` | String | `None` | The local file cache capacity in bytes. |
+| `storage.bucket` | String | `None` | The S3 bucket name.<br/>**It's only used when the storage type is `S3`, `Oss` and `Gcs`**. |
+| `storage.root` | String | `None` | The S3 data will be stored in the specified prefix, for example, `s3://${bucket}/${root}`.<br/>**It's only used when the storage type is `S3`, `Oss` and `Azblob`**. |
+| `storage.access_key_id` | String | `None` | The access key id of the aws account.<br/>It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.<br/>**It's only used when the storage type is `S3` and `Oss`**. |
+| `storage.secret_access_key` | String | `None` | The secret access key of the aws account.<br/>It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.<br/>**It's only used when the storage type is `S3`**. |
+| `storage.access_key_secret` | String | `None` | The secret access key of the aliyun account.<br/>**It's only used when the storage type is `Oss`**. |
+| `storage.account_name` | String | `None` | The account key of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.account_key` | String | `None` | The account key of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.scope` | String | `None` | The scope of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
+| `storage.credential_path` | String | `None` | The credential path of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
+| `storage.container` | String | `None` | The container of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.sas_token` | String | `None` | The sas token of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.endpoint` | String | `None` | The endpoint of the S3 service.<br/>**It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**. |
+| `storage.region` | String | `None` | The region of the S3 service.<br/>**It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**. |
+| `[[region_engine]]` | -- | -- | The region engine options. You can configure multiple region engines. |
+| `region_engine.mito` | -- | -- | The Mito engine options. |
+| `region_engine.mito.num_workers` | Integer | `8` | Number of region workers. |
+| `region_engine.mito.worker_channel_size` | Integer | `128` | Request channel size of each worker. |
+| `region_engine.mito.worker_request_batch_size` | Integer | `64` | Max batch size for a worker to handle requests. |
+| `region_engine.mito.manifest_checkpoint_distance` | Integer | `10` | Number of meta action updated to trigger a new checkpoint for the manifest. |
+| `region_engine.mito.compress_manifest` | Bool | `false` | Whether to compress manifest and checkpoint file by gzip (default false). |
+| `region_engine.mito.max_background_jobs` | Integer | `4` | Max number of running background jobs |
+| `region_engine.mito.auto_flush_interval` | String | `1h` | Interval to auto flush a region if it has not flushed yet. |
+| `region_engine.mito.global_write_buffer_size` | String | `1GB` | Global write buffer size for all regions. If not set, it's default to 1/8 of OS memory with a max limitation of 1GB. |
+| `region_engine.mito.global_write_buffer_reject_size` | String | `2GB` | Global write buffer size threshold to reject write requests. If not set, it's default to 2 times of `global_write_buffer_size` |
+| `region_engine.mito.sst_meta_cache_size` | String | `128MB` | Cache size for SST metadata. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/32 of OS memory with a max limitation of 128MB. |
+| `region_engine.mito.vector_cache_size` | String | `512MB` | Cache size for vectors and arrow arrays. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
+| `region_engine.mito.page_cache_size` | String | `512MB` | Cache size for pages of SST row groups. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
+| `region_engine.mito.sst_write_buffer_size` | String | `8MB` | Buffer size for SST writing. |
+| `region_engine.mito.scan_parallelism` | Integer | `0` | Parallelism to scan a region (default: 1/4 of cpu cores).<br/>- `0`: using the default value (1/4 of cpu cores).<br/>- `1`: scan in current thread.<br/>- `n`: scan in parallelism n. |
+| `region_engine.mito.parallel_scan_channel_size` | Integer | `32` | Capacity of the channel to send data from parallel scan tasks to the main task. |
+| `region_engine.mito.allow_stale_entries` | Bool | `false` | Whether to allow stale WAL entries read during replay. |
+| `region_engine.mito.inverted_index` | -- | -- | The options for inverted index in Mito engine. |
+| `region_engine.mito.inverted_index.create_on_flush` | String | `auto` | Whether to create the index on flush.<br/>- `auto`: automatically<br/>- `disable`: never |
+| `region_engine.mito.inverted_index.create_on_compaction` | String | `auto` | Whether to create the index on compaction.<br/>- `auto`: automatically<br/>- `disable`: never |
+| `region_engine.mito.inverted_index.apply_on_query` | String | `auto` | Whether to apply the index on query<br/>- `auto`: automatically<br/>- `disable`: never |
+| `region_engine.mito.inverted_index.mem_threshold_on_create` | String | `64M` | Memory threshold for performing an external sort during index creation.<br/>Setting to empty will disable external sorting, forcing all sorting operations to happen in memory. |
+| `region_engine.mito.inverted_index.intermediate_path` | String | `""` | File system path to store intermediate files for external sorting (default `{data_home}/index_intermediate`). |
+| `region_engine.mito.memtable` | -- | -- | -- |
+| `region_engine.mito.memtable.type` | String | `time_series` | Memtable type.<br/>- `time_series`: time-series memtable<br/>- `partition_tree`: partition tree memtable (experimental) |
+| `region_engine.mito.memtable.index_max_keys_per_shard` | Integer | `8192` | The max number of keys in one shard.<br/>Only available for `partition_tree` memtable. |
+| `region_engine.mito.memtable.data_freeze_threshold` | Integer | `32768` | The max rows of data inside the actively writing buffer in one shard.<br/>Only available for `partition_tree` memtable. |
+| `region_engine.mito.memtable.fork_dictionary_bytes` | String | `1GiB` | Max dictionary bytes.<br/>Only available for `partition_tree` memtable. |
+| `logging` | -- | -- | The logging options. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
+| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
+| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
+| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
+| `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
+| `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
+| `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import.db` | String | `None` | -- |
+| `export_metrics.remote_write` | -- | -- | -- |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
+
+
+## Cluster Mode
+
+### Frontend
+
+| Key | Type | Default | Descriptions |
+| --- | -----| ------- | ----------- |
+| `mode` | String | `standalone` | The running mode of the datanode. It can be `standalone` or `distributed`. |
+| `default_timezone` | String | `None` | The default timezone of the server. |
+| `heartbeat` | -- | -- | The heartbeat options. |
+| `heartbeat.interval` | String | `18s` | Interval for sending heartbeat messages to the metasrv. |
+| `heartbeat.retry_interval` | String | `3s` | Interval for retrying to send heartbeat messages to the metasrv. |
+| `http` | -- | -- | The HTTP server options. |
+| `http.addr` | String | `127.0.0.1:4000` | The address to bind the HTTP server. |
+| `http.timeout` | String | `30s` | HTTP request timeout. |
+| `http.body_limit` | String | `64MB` | HTTP request body limit.<br/>Support the following units are supported: `B`, `KB`, `KiB`, `MB`, `MiB`, `GB`, `GiB`, `TB`, `TiB`, `PB`, `PiB`. |
+| `grpc` | -- | -- | The gRPC server options. |
+| `grpc.addr` | String | `127.0.0.1:4001` | The address to bind the gRPC server. |
+| `grpc.runtime_size` | Integer | `8` | The number of server worker threads. |
+| `mysql` | -- | -- | MySQL server options. |
+| `mysql.enable` | Bool | `true` | Whether to enable. |
+| `mysql.addr` | String | `127.0.0.1:4002` | The addr to bind the MySQL server. |
+| `mysql.runtime_size` | Integer | `2` | The number of server worker threads. |
+| `mysql.tls` | -- | -- | -- |
+| `mysql.tls.mode` | String | `disable` | TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html<br/>- `disable` (default value)<br/>- `prefer`<br/>- `require`<br/>- `verify-ca`<br/>- `verify-full` |
+| `mysql.tls.cert_path` | String | `None` | Certificate file path. |
+| `mysql.tls.key_path` | String | `None` | Private key file path. |
+| `mysql.tls.watch` | Bool | `false` | Watch for Certificate and key file change and auto reload |
+| `postgres` | -- | -- | PostgresSQL server options. |
+| `postgres.enable` | Bool | `true` | Whether to enable |
+| `postgres.addr` | String | `127.0.0.1:4003` | The addr to bind the PostgresSQL server. |
+| `postgres.runtime_size` | Integer | `2` | The number of server worker threads. |
+| `postgres.tls` | -- | -- | PostgresSQL server TLS options, see `mysql_options.tls` section. |
+| `postgres.tls.mode` | String | `disable` | TLS mode. |
+| `postgres.tls.cert_path` | String | `None` | Certificate file path. |
+| `postgres.tls.key_path` | String | `None` | Private key file path. |
+| `postgres.tls.watch` | Bool | `false` | Watch for Certificate and key file change and auto reload |
+| `opentsdb` | -- | -- | OpenTSDB protocol options. |
+| `opentsdb.enable` | Bool | `true` | Whether to enable |
+| `opentsdb.addr` | String | `127.0.0.1:4242` | OpenTSDB telnet API server address. |
+| `opentsdb.runtime_size` | Integer | `2` | The number of server worker threads. |
+| `influxdb` | -- | -- | InfluxDB protocol options. |
+| `influxdb.enable` | Bool | `true` | Whether to enable InfluxDB protocol in HTTP API. |
+| `prom_store` | -- | -- | Prometheus remote storage options |
+| `prom_store.enable` | Bool | `true` | Whether to enable Prometheus remote write and read in HTTP API. |
+| `prom_store.with_metric_engine` | Bool | `true` | Whether to store the data from Prometheus remote write in metric engine. |
+| `meta_client` | -- | -- | The metasrv client options. |
+| `meta_client.metasrv_addrs` | Array | -- | The addresses of the metasrv. |
+| `meta_client.timeout` | String | `3s` | Operation timeout. |
+| `meta_client.heartbeat_timeout` | String | `500ms` | Heartbeat timeout. |
+| `meta_client.ddl_timeout` | String | `10s` | DDL timeout. |
+| `meta_client.connect_timeout` | String | `1s` | Connect server timeout. |
+| `meta_client.tcp_nodelay` | Bool | `true` | `TCP_NODELAY` option for accepted connections. |
+| `meta_client.metadata_cache_max_capacity` | Integer | `100000` | The configuration about the cache of the metadata. |
+| `meta_client.metadata_cache_ttl` | String | `10m` | TTL of the metadata cache. |
+| `meta_client.metadata_cache_tti` | String | `5m` | -- |
+| `datanode` | -- | -- | Datanode options. |
+| `datanode.client` | -- | -- | Datanode client options. |
+| `datanode.client.timeout` | String | `10s` | -- |
+| `datanode.client.connect_timeout` | String | `10s` | -- |
+| `datanode.client.tcp_nodelay` | Bool | `true` | -- |
+| `logging` | -- | -- | The logging options. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
+| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
+| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
+| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
+| `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
+| `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
+| `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import.db` | String | `None` | -- |
+| `export_metrics.remote_write` | -- | -- | -- |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
+
+
+### Metasrv
+
+| Key | Type | Default | Descriptions |
+| --- | -----| ------- | ----------- |
+| `data_home` | String | `/tmp/metasrv/` | The working home directory. |
+| `bind_addr` | String | `127.0.0.1:3002` | The bind address of metasrv. |
+| `server_addr` | String | `127.0.0.1:3002` | The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost. |
+| `store_addr` | String | `127.0.0.1:2379` | Etcd server address. |
+| `selector` | String | `lease_based` | Datanode selector type.<br/>- `lease_based` (default value).<br/>- `load_based`<br/>For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector". |
+| `use_memory_store` | Bool | `false` | Store data in memory. |
+| `enable_telemetry` | Bool | `true` | Whether to enable greptimedb telemetry. |
+| `store_key_prefix` | String | `""` | If it's not empty, the metasrv will store all data with this key prefix. |
+| `procedure` | -- | -- | Procedure storage options. |
+| `procedure.max_retry_times` | Integer | `12` | Procedure max retry time. |
+| `procedure.retry_delay` | String | `500ms` | Initial retry delay of procedures, increases exponentially |
+| `procedure.max_metadata_value_size` | String | `1500KiB` | Auto split large value<br/>GreptimeDB procedure uses etcd as the default metadata storage backend.<br/>The etcd the maximum size of any request is 1.5 MiB<br/>1500KiB = 1536KiB (1.5MiB) - 36KiB (reserved size of key)<br/>Comments out the `max_metadata_value_size`, for don't split large value (no limit). |
+| `failure_detector` | -- | -- | -- |
+| `failure_detector.threshold` | Float | `8.0` | -- |
+| `failure_detector.min_std_deviation` | String | `100ms` | -- |
+| `failure_detector.acceptable_heartbeat_pause` | String | `3000ms` | -- |
+| `failure_detector.first_heartbeat_estimate` | String | `1000ms` | -- |
+| `datanode` | -- | -- | Datanode options. |
+| `datanode.client` | -- | -- | Datanode client options. |
+| `datanode.client.timeout` | String | `10s` | -- |
+| `datanode.client.connect_timeout` | String | `10s` | -- |
+| `datanode.client.tcp_nodelay` | Bool | `true` | -- |
+| `wal` | -- | -- | -- |
+| `wal.provider` | String | `raft_engine` | -- |
+| `wal.broker_endpoints` | Array | -- | The broker endpoints of the Kafka cluster. |
+| `wal.num_topics` | Integer | `64` | Number of topics to be created upon start. |
+| `wal.selector_type` | String | `round_robin` | Topic selector type.<br/>Available selector types:<br/>- `round_robin` (default) |
+| `wal.topic_name_prefix` | String | `greptimedb_wal_topic` | A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`. |
+| `wal.replication_factor` | Integer | `1` | Expected number of replicas of each partition. |
+| `wal.create_topic_timeout` | String | `30s` | Above which a topic creation operation will be cancelled. |
+| `wal.backoff_init` | String | `500ms` | The initial backoff for kafka clients. |
+| `wal.backoff_max` | String | `10s` | The maximum backoff for kafka clients. |
+| `wal.backoff_base` | Integer | `2` | Exponential backoff rate, i.e. next backoff = base * current backoff. |
+| `wal.backoff_deadline` | String | `5mins` | Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate. |
+| `logging` | -- | -- | The logging options. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
+| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
+| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
+| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
+| `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
+| `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
+| `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import.db` | String | `None` | -- |
+| `export_metrics.remote_write` | -- | -- | -- |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
+
+
+### Datanode
+
+| Key | Type | Default | Descriptions |
+| --- | -----| ------- | ----------- |
+| `mode` | String | `standalone` | The running mode of the datanode. It can be `standalone` or `distributed`. |
+| `node_id` | Integer | `None` | The datanode identifier and should be unique in the cluster. |
+| `require_lease_before_startup` | Bool | `false` | Start services after regions have obtained leases.<br/>It will block the datanode start if it can't receive leases in the heartbeat from metasrv. |
+| `init_regions_in_background` | Bool | `false` | Initialize all regions in the background during the startup.<br/>By default, it provides services after all regions have been initialized. |
+| `rpc_addr` | String | `127.0.0.1:3001` | The gRPC address of the datanode. |
+| `rpc_hostname` | String | `None` | The hostname of the datanode. |
+| `rpc_runtime_size` | Integer | `8` | The number of gRPC server worker threads. |
+| `rpc_max_recv_message_size` | String | `512MB` | The maximum receive message size for gRPC server. |
+| `rpc_max_send_message_size` | String | `512MB` | The maximum send message size for gRPC server. |
+| `enable_telemetry` | Bool | `true` | Enable telemetry to collect anonymous usage data. |
+| `heartbeat` | -- | -- | The heartbeat options. |
+| `heartbeat.interval` | String | `3s` | Interval for sending heartbeat messages to the metasrv. |
+| `heartbeat.retry_interval` | String | `3s` | Interval for retrying to send heartbeat messages to the metasrv. |
+| `meta_client` | -- | -- | The metasrv client options. |
+| `meta_client.metasrv_addrs` | Array | -- | The addresses of the metasrv. |
+| `meta_client.timeout` | String | `3s` | Operation timeout. |
+| `meta_client.heartbeat_timeout` | String | `500ms` | Heartbeat timeout. |
+| `meta_client.ddl_timeout` | String | `10s` | DDL timeout. |
+| `meta_client.connect_timeout` | String | `1s` | Connect server timeout. |
+| `meta_client.tcp_nodelay` | Bool | `true` | `TCP_NODELAY` option for accepted connections. |
+| `meta_client.metadata_cache_max_capacity` | Integer | `100000` | The configuration about the cache of the metadata. |
+| `meta_client.metadata_cache_ttl` | String | `10m` | TTL of the metadata cache. |
+| `meta_client.metadata_cache_tti` | String | `5m` | -- |
+| `wal` | -- | -- | The WAL options. |
+| `wal.provider` | String | `raft_engine` | The provider of the WAL.<br/>- `raft_engine`: the wal is stored in the local file system by raft-engine.<br/>- `kafka`: it's remote wal that data is stored in Kafka. |
+| `wal.dir` | String | `None` | The directory to store the WAL files.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.file_size` | String | `256MB` | The size of the WAL segment file.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.purge_threshold` | String | `4GB` | The threshold of the WAL size to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.purge_interval` | String | `10m` | The interval to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.read_batch_size` | Integer | `128` | The read batch size.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.sync_write` | Bool | `false` | Whether to use sync write.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.enable_log_recycle` | Bool | `true` | Whether to reuse logically truncated log files.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.prefill_log_files` | Bool | `false` | Whether to pre-create log files on start up.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.sync_period` | String | `10s` | Duration for fsyncing log files.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.broker_endpoints` | Array | -- | The Kafka broker endpoints.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.max_batch_size` | String | `1MB` | The max size of a single producer batch.<br/>Warning: Kafka has a default limit of 1MB per message in a topic.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.linger` | String | `200ms` | The linger duration of a kafka batch producer.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.consumer_wait_timeout` | String | `100ms` | The consumer wait timeout.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_init` | String | `500ms` | The initial backoff delay.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_max` | String | `10s` | The maximum backoff delay.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_base` | Integer | `2` | The exponential backoff rate, i.e. next backoff = base * current backoff.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.backoff_deadline` | String | `5mins` | The deadline of retries.<br/>**It's only used when the provider is `kafka`**. |
+| `storage` | -- | -- | The data storage options. |
+| `storage.data_home` | String | `/tmp/greptimedb/` | The working home directory. |
+| `storage.type` | String | `File` | The storage type used to store the data.<br/>- `File`: the data is stored in the local file system.<br/>- `S3`: the data is stored in the S3 object storage.<br/>- `Gcs`: the data is stored in the Google Cloud Storage.<br/>- `Azblob`: the data is stored in the Azure Blob Storage.<br/>- `Oss`: the data is stored in the Aliyun OSS. |
+| `storage.cache_path` | String | `None` | Cache configuration for object storage such as 'S3' etc.<br/>The local file cache directory. |
+| `storage.cache_capacity` | String | `None` | The local file cache capacity in bytes. |
+| `storage.bucket` | String | `None` | The S3 bucket name.<br/>**It's only used when the storage type is `S3`, `Oss` and `Gcs`**. |
+| `storage.root` | String | `None` | The S3 data will be stored in the specified prefix, for example, `s3://${bucket}/${root}`.<br/>**It's only used when the storage type is `S3`, `Oss` and `Azblob`**. |
+| `storage.access_key_id` | String | `None` | The access key id of the aws account.<br/>It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.<br/>**It's only used when the storage type is `S3` and `Oss`**. |
+| `storage.secret_access_key` | String | `None` | The secret access key of the aws account.<br/>It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.<br/>**It's only used when the storage type is `S3`**. |
+| `storage.access_key_secret` | String | `None` | The secret access key of the aliyun account.<br/>**It's only used when the storage type is `Oss`**. |
+| `storage.account_name` | String | `None` | The account key of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.account_key` | String | `None` | The account key of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.scope` | String | `None` | The scope of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
+| `storage.credential_path` | String | `None` | The credential path of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
+| `storage.container` | String | `None` | The container of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.sas_token` | String | `None` | The sas token of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
+| `storage.endpoint` | String | `None` | The endpoint of the S3 service.<br/>**It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**. |
+| `storage.region` | String | `None` | The region of the S3 service.<br/>**It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**. |
+| `[[region_engine]]` | -- | -- | The region engine options. You can configure multiple region engines. |
+| `region_engine.mito` | -- | -- | The Mito engine options. |
+| `region_engine.mito.num_workers` | Integer | `8` | Number of region workers. |
+| `region_engine.mito.worker_channel_size` | Integer | `128` | Request channel size of each worker. |
+| `region_engine.mito.worker_request_batch_size` | Integer | `64` | Max batch size for a worker to handle requests. |
+| `region_engine.mito.manifest_checkpoint_distance` | Integer | `10` | Number of meta action updated to trigger a new checkpoint for the manifest. |
+| `region_engine.mito.compress_manifest` | Bool | `false` | Whether to compress manifest and checkpoint file by gzip (default false). |
+| `region_engine.mito.max_background_jobs` | Integer | `4` | Max number of running background jobs |
+| `region_engine.mito.auto_flush_interval` | String | `1h` | Interval to auto flush a region if it has not flushed yet. |
+| `region_engine.mito.global_write_buffer_size` | String | `1GB` | Global write buffer size for all regions. If not set, it's default to 1/8 of OS memory with a max limitation of 1GB. |
+| `region_engine.mito.global_write_buffer_reject_size` | String | `2GB` | Global write buffer size threshold to reject write requests. If not set, it's default to 2 times of `global_write_buffer_size` |
+| `region_engine.mito.sst_meta_cache_size` | String | `128MB` | Cache size for SST metadata. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/32 of OS memory with a max limitation of 128MB. |
+| `region_engine.mito.vector_cache_size` | String | `512MB` | Cache size for vectors and arrow arrays. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
+| `region_engine.mito.page_cache_size` | String | `512MB` | Cache size for pages of SST row groups. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
+| `region_engine.mito.sst_write_buffer_size` | String | `8MB` | Buffer size for SST writing. |
+| `region_engine.mito.scan_parallelism` | Integer | `0` | Parallelism to scan a region (default: 1/4 of cpu cores).<br/>- `0`: using the default value (1/4 of cpu cores).<br/>- `1`: scan in current thread.<br/>- `n`: scan in parallelism n. |
+| `region_engine.mito.parallel_scan_channel_size` | Integer | `32` | Capacity of the channel to send data from parallel scan tasks to the main task. |
+| `region_engine.mito.allow_stale_entries` | Bool | `false` | Whether to allow stale WAL entries read during replay. |
+| `region_engine.mito.inverted_index` | -- | -- | The options for inverted index in Mito engine. |
+| `region_engine.mito.inverted_index.create_on_flush` | String | `auto` | Whether to create the index on flush.<br/>- `auto`: automatically<br/>- `disable`: never |
+| `region_engine.mito.inverted_index.create_on_compaction` | String | `auto` | Whether to create the index on compaction.<br/>- `auto`: automatically<br/>- `disable`: never |
+| `region_engine.mito.inverted_index.apply_on_query` | String | `auto` | Whether to apply the index on query<br/>- `auto`: automatically<br/>- `disable`: never |
+| `region_engine.mito.inverted_index.mem_threshold_on_create` | String | `64M` | Memory threshold for performing an external sort during index creation.<br/>Setting to empty will disable external sorting, forcing all sorting operations to happen in memory. |
+| `region_engine.mito.inverted_index.intermediate_path` | String | `""` | File system path to store intermediate files for external sorting (default `{data_home}/index_intermediate`). |
+| `region_engine.mito.memtable` | -- | -- | -- |
+| `region_engine.mito.memtable.type` | String | `time_series` | Memtable type.<br/>- `time_series`: time-series memtable<br/>- `partition_tree`: partition tree memtable (experimental) |
+| `region_engine.mito.memtable.index_max_keys_per_shard` | Integer | `8192` | The max number of keys in one shard.<br/>Only available for `partition_tree` memtable. |
+| `region_engine.mito.memtable.data_freeze_threshold` | Integer | `32768` | The max rows of data inside the actively writing buffer in one shard.<br/>Only available for `partition_tree` memtable. |
+| `region_engine.mito.memtable.fork_dictionary_bytes` | String | `1GiB` | Max dictionary bytes.<br/>Only available for `partition_tree` memtable. |
+| `logging` | -- | -- | The logging options. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
+| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
+| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
+| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
+| `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
+| `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
+| `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import.db` | String | `None` | -- |
+| `export_metrics.remote_write` | -- | -- | -- |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |

config/datanode.example.toml

@@ -1,75 +1,277 @@
-# Node running mode, see `standalone.example.toml`.
+## The running mode of the datanode. It can be `standalone` or `distributed`.
-mode = "distributed"
+mode = "standalone"
-# The datanode identifier, should be unique.
+
+## The datanode identifier and should be unique in the cluster.
+## +toml2docs:none-default
 node_id = 42
-# gRPC server address, "127.0.0.1:3001" by default.
+
-rpc_addr = "127.0.0.1:3001"
+## Start services after regions have obtained leases.
-# Hostname of this node.
+## It will block the datanode start if it can't receive leases in the heartbeat from metasrv.
-rpc_hostname = "127.0.0.1"
-# The number of gRPC server worker threads, 8 by default.
-rpc_runtime_size = 8
-# Start services after regions have obtained leases.
-# It will block the datanode start if it can't receive leases in the heartbeat from metasrv.
 require_lease_before_startup = false
 
-# Initialize all regions in the background during the startup.
+## Initialize all regions in the background during the startup.
-# By default, it provides services after all regions have been initialized.
+## By default, it provides services after all regions have been initialized.
-initialize_region_in_background = false
+init_regions_in_background = false
 
+## The gRPC address of the datanode.
+rpc_addr = "127.0.0.1:3001"
+
+## The hostname of the datanode.
+## +toml2docs:none-default
+rpc_hostname = "127.0.0.1"
+
+## The number of gRPC server worker threads.
+rpc_runtime_size = 8
+
+## The maximum receive message size for gRPC server.
+rpc_max_recv_message_size = "512MB"
+
+## The maximum send message size for gRPC server.
+rpc_max_send_message_size = "512MB"
+
+## Enable telemetry to collect anonymous usage data.
+enable_telemetry = true
+
+## The heartbeat options.
 [heartbeat]
-# Interval for sending heartbeat messages to the Metasrv, 3 seconds by default.
+## Interval for sending heartbeat messages to the metasrv.
 interval = "3s"
 
-# Metasrv client options.
+## Interval for retrying to send heartbeat messages to the metasrv.
+retry_interval = "3s"
+
+## The metasrv client options.
 [meta_client]
-# Metasrv address list.
+## The addresses of the metasrv.
 metasrv_addrs = ["127.0.0.1:3002"]
-# Heartbeat timeout, 500 milliseconds by default.
+
-heartbeat_timeout = "500ms"
+## Operation timeout.
-# Operation timeout, 3 seconds by default.
 timeout = "3s"
-# Connect server timeout, 1 second by default.
+
+## Heartbeat timeout.
+heartbeat_timeout = "500ms"
+
+## DDL timeout.
+ddl_timeout = "10s"
+
+## Connect server timeout.
 connect_timeout = "1s"
-# `TCP_NODELAY` option for accepted connections, true by default.
+
+## `TCP_NODELAY` option for accepted connections.
 tcp_nodelay = true
 
-# WAL options.
+## The configuration about the cache of the metadata.
+metadata_cache_max_capacity = 100000
+
+## TTL of the metadata cache.
+metadata_cache_ttl = "10m"
+
+# TTI of the metadata cache.
+metadata_cache_tti = "5m"
+
+## The WAL options.
 [wal]
+## The provider of the WAL.
+## - `raft_engine`: the wal is stored in the local file system by raft-engine.
+## - `kafka`: it's remote wal that data is stored in Kafka.
 provider = "raft_engine"
 
-# Raft-engine wal options, see `standalone.example.toml`.
+## The directory to store the WAL files.
-# dir = "/tmp/greptimedb/wal"
+## **It's only used when the provider is `raft_engine`**.
+## +toml2docs:none-default
+dir = "/tmp/greptimedb/wal"
+
+## The size of the WAL segment file.
+## **It's only used when the provider is `raft_engine`**.
 file_size = "256MB"
+
+## The threshold of the WAL size to trigger a flush.
+## **It's only used when the provider is `raft_engine`**.
 purge_threshold = "4GB"
+
+## The interval to trigger a flush.
+## **It's only used when the provider is `raft_engine`**.
 purge_interval = "10m"
+
+## The read batch size.
+## **It's only used when the provider is `raft_engine`**.
 read_batch_size = 128
+
+## Whether to use sync write.
+## **It's only used when the provider is `raft_engine`**.
 sync_write = false
 
-# Kafka wal options, see `standalone.example.toml`.
+## Whether to reuse logically truncated log files.
-# broker_endpoints = ["127.0.0.1:9092"]
+## **It's only used when the provider is `raft_engine`**.
-# Warning: Kafka has a default limit of 1MB per message in a topic.
+enable_log_recycle = true
-# max_batch_size = "1MB"
-# linger = "200ms"
-# consumer_wait_timeout = "100ms"
-# backoff_init = "500ms"
-# backoff_max = "10s"
-# backoff_base = 2
-# backoff_deadline = "5mins"
 
-# Storage options, see `standalone.example.toml`.
+## Whether to pre-create log files on start up.
+## **It's only used when the provider is `raft_engine`**.
+prefill_log_files = false
+
+## Duration for fsyncing log files.
+## **It's only used when the provider is `raft_engine`**.
+sync_period = "10s"
+
+## The Kafka broker endpoints.
+## **It's only used when the provider is `kafka`**.
+broker_endpoints = ["127.0.0.1:9092"]
+
+## The max size of a single producer batch.
+## Warning: Kafka has a default limit of 1MB per message in a topic.
+## **It's only used when the provider is `kafka`**.
+max_batch_size = "1MB"
+
+## The linger duration of a kafka batch producer.
+## **It's only used when the provider is `kafka`**.
+linger = "200ms"
+
+## The consumer wait timeout.
+## **It's only used when the provider is `kafka`**.
+consumer_wait_timeout = "100ms"
+
+## The initial backoff delay.
+## **It's only used when the provider is `kafka`**.
+backoff_init = "500ms"
+
+## The maximum backoff delay.
+## **It's only used when the provider is `kafka`**.
+backoff_max = "10s"
+
+## The exponential backoff rate, i.e. next backoff = base * current backoff.
+## **It's only used when the provider is `kafka`**.
+backoff_base = 2
+
+## The deadline of retries.
+## **It's only used when the provider is `kafka`**.
+backoff_deadline = "5mins"
+
+# Example of using S3 as the storage.
+# [storage]
+# type = "S3"
+# bucket = "greptimedb"
+# root = "data"
+# access_key_id = "test"
+# secret_access_key = "123456"
+# endpoint = "https://s3.amazonaws.com"
+# region = "us-west-2"
+
+# Example of using Oss as the storage.
+# [storage]
+# type = "Oss"
+# bucket = "greptimedb"
+# root = "data"
+# access_key_id = "test"
+# access_key_secret = "123456"
+# endpoint = "https://oss-cn-hangzhou.aliyuncs.com"
+
+# Example of using Azblob as the storage.
+# [storage]
+# type = "Azblob"
+# container = "greptimedb"
+# root = "data"
+# account_name = "test"
+# account_key = "123456"
+# endpoint = "https://greptimedb.blob.core.windows.net"
+# sas_token = ""
+
+# Example of using Gcs as the storage.
+# [storage]
+# type = "Gcs"
+# bucket = "greptimedb"
+# root = "data"
+# scope = "test"
+# credential_path = "123456"
+# endpoint = "https://storage.googleapis.com"
+
+## The data storage options.
 [storage]
-# The working home directory.
+## The working home directory.
 data_home = "/tmp/greptimedb/"
-# Storage type.
-type = "File"
-# TTL for all tables. Disabled by default.
-# global_ttl = "7d"
 
-# Cache configuration for object storage such as 'S3' etc.
+## The storage type used to store the data.
-# The local file cache directory
+## - `File`: the data is stored in the local file system.
-# cache_path = "/path/local_cache"
+## - `S3`: the data is stored in the S3 object storage.
-# The local file cache capacity in bytes.
+## - `Gcs`: the data is stored in the Google Cloud Storage.
-# cache_capacity = "256MB"
+## - `Azblob`: the data is stored in the Azure Blob Storage.
+## - `Oss`: the data is stored in the Aliyun OSS.
+type = "File"
+
+## Cache configuration for object storage such as 'S3' etc.
+## The local file cache directory.
+## +toml2docs:none-default
+cache_path = "/path/local_cache"
+
+## The local file cache capacity in bytes.
+## +toml2docs:none-default
+cache_capacity = "256MB"
+
+## The S3 bucket name.
+## **It's only used when the storage type is `S3`, `Oss` and `Gcs`**.
+## +toml2docs:none-default
+bucket = "greptimedb"
+
+## The S3 data will be stored in the specified prefix, for example, `s3://${bucket}/${root}`.
+## **It's only used when the storage type is `S3`, `Oss` and `Azblob`**.
+## +toml2docs:none-default
+root = "greptimedb"
+
+## The access key id of the aws account.
+## It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.
+## **It's only used when the storage type is `S3` and `Oss`**.
+## +toml2docs:none-default
+access_key_id = "test"
+
+## The secret access key of the aws account.
+## It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.
+## **It's only used when the storage type is `S3`**.
+## +toml2docs:none-default
+secret_access_key = "test"
+
+## The secret access key of the aliyun account.
+## **It's only used when the storage type is `Oss`**.
+## +toml2docs:none-default
+access_key_secret = "test"
+
+## The account key of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+account_name = "test"
+
+## The account key of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+account_key = "test"
+
+## The scope of the google cloud storage.
+## **It's only used when the storage type is `Gcs`**.
+## +toml2docs:none-default
+scope = "test"
+
+## The credential path of the google cloud storage.
+## **It's only used when the storage type is `Gcs`**.
+## +toml2docs:none-default
+credential_path = "test"
+
+## The container of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+container = "greptimedb"
+
+## The sas token of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+sas_token = ""
+
+## The endpoint of the S3 service.
+## **It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**.
+## +toml2docs:none-default
+endpoint = "https://s3.amazonaws.com"
+
+## The region of the S3 service.
+## **It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**.
+## +toml2docs:none-default
+region = "us-west-2"
 
 # Custom storage options
 # [[storage.providers]]
@@ -77,61 +279,152 @@ type = "File"
 # [[storage.providers]]
 # type = "Gcs"
 
-# Mito engine options
+## The region engine options. You can configure multiple region engines.
 [[region_engine]]
+
+## The Mito engine options.
 [region_engine.mito]
-# Number of region workers
+
+## Number of region workers.
 num_workers = 8
-# Request channel size of each worker
+
+## Request channel size of each worker.
 worker_channel_size = 128
-# Max batch size for a worker to handle requests
+
+## Max batch size for a worker to handle requests.
 worker_request_batch_size = 64
-# Number of meta action updated to trigger a new checkpoint for the manifest
+
+## Number of meta action updated to trigger a new checkpoint for the manifest.
 manifest_checkpoint_distance = 10
-# Whether to compress manifest and checkpoint file by gzip (default false).
+
+## Whether to compress manifest and checkpoint file by gzip (default false).
 compress_manifest = false
-# Max number of running background jobs
+
+## Max number of running background jobs
 max_background_jobs = 4
-# Interval to auto flush a region if it has not flushed yet.
+
+## Interval to auto flush a region if it has not flushed yet.
 auto_flush_interval = "1h"
-# Global write buffer size for all regions.
+
+## Global write buffer size for all regions. If not set, it's default to 1/8 of OS memory with a max limitation of 1GB.
 global_write_buffer_size = "1GB"
-# Global write buffer size threshold to reject write requests (default 2G).
+
+## Global write buffer size threshold to reject write requests. If not set, it's default to 2 times of `global_write_buffer_size`
 global_write_buffer_reject_size = "2GB"
-# Cache size for SST metadata (default 128MB). Setting it to 0 to disable the cache.
+
+## Cache size for SST metadata. Setting it to 0 to disable the cache.
+## If not set, it's default to 1/32 of OS memory with a max limitation of 128MB.
 sst_meta_cache_size = "128MB"
-# Cache size for vectors and arrow arrays (default 512MB). Setting it to 0 to disable the cache.
+
+## Cache size for vectors and arrow arrays. Setting it to 0 to disable the cache.
+## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
 vector_cache_size = "512MB"
-# Cache size for pages of SST row groups (default 512MB). Setting it to 0 to disable the cache.
+
+## Cache size for pages of SST row groups. Setting it to 0 to disable the cache.
+## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
 page_cache_size = "512MB"
-# Buffer size for SST writing.
+
+## Buffer size for SST writing.
 sst_write_buffer_size = "8MB"
-# Parallelism to scan a region (default: 1/4 of cpu cores).
+
-# - 0: using the default value (1/4 of cpu cores).
+## Parallelism to scan a region (default: 1/4 of cpu cores).
-# - 1: scan in current thread.
+## - `0`: using the default value (1/4 of cpu cores).
-# - n: scan in parallelism n.
+## - `1`: scan in current thread.
+## - `n`: scan in parallelism n.
 scan_parallelism = 0
-# Capacity of the channel to send data from parallel scan tasks to the main task (default 32).
+
+## Capacity of the channel to send data from parallel scan tasks to the main task.
 parallel_scan_channel_size = 32
-# Whether to allow stale WAL entries read during replay.
+
+## Whether to allow stale WAL entries read during replay.
 allow_stale_entries = false
 
-# Log options, see `standalone.example.toml`
+## The options for inverted index in Mito engine.
-# [logging]
+[region_engine.mito.inverted_index]
-# dir = "/tmp/greptimedb/logs"
-# level = "info"
 
-# Datanode export the metrics generated by itself
+## Whether to create the index on flush.
-# encoded to Prometheus remote-write format
+## - `auto`: automatically
-# and send to Prometheus remote-write compatible receiver (e.g. send to `greptimedb` itself)
+## - `disable`: never
-# This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+create_on_flush = "auto"
-# [export_metrics]
+
-# whether enable export metrics, default is false
+## Whether to create the index on compaction.
-# enable = false
+## - `auto`: automatically
-# The interval of export metrics
+## - `disable`: never
-# write_interval = "30s"
+create_on_compaction = "auto"
-# [export_metrics.remote_write]
+
-# The url the metrics send to. The url is empty by default, url example: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`
+## Whether to apply the index on query
-# url = ""
+## - `auto`: automatically
-# HTTP headers of Prometheus remote-write carry
+## - `disable`: never
-# headers = {}
+apply_on_query = "auto"
+
+## Memory threshold for performing an external sort during index creation.
+## Setting to empty will disable external sorting, forcing all sorting operations to happen in memory.
+mem_threshold_on_create = "64M"
+
+## File system path to store intermediate files for external sorting (default `{data_home}/index_intermediate`).
+intermediate_path = ""
+
+[region_engine.mito.memtable]
+## Memtable type.
+## - `time_series`: time-series memtable
+## - `partition_tree`: partition tree memtable (experimental)
+type = "time_series"
+
+## The max number of keys in one shard.
+## Only available for `partition_tree` memtable.
+index_max_keys_per_shard = 8192
+
+## The max rows of data inside the actively writing buffer in one shard.
+## Only available for `partition_tree` memtable.
+data_freeze_threshold = 32768
+
+## Max dictionary bytes.
+## Only available for `partition_tree` memtable.
+fork_dictionary_bytes = "1GiB"
+
+## The logging options.
+[logging]
+## The directory to store the log files.
+dir = "/tmp/greptimedb/logs"
+
+## The log level. Can be `info`/`debug`/`warn`/`error`.
+## +toml2docs:none-default
+level = "info"
+
+## Enable OTLP tracing.
+enable_otlp_tracing = false
+
+## The OTLP tracing endpoint.
+## +toml2docs:none-default
+otlp_endpoint = ""
+
+## Whether to append logs to stdout.
+append_stdout = true
+
+## The percentage of tracing will be sampled and exported.
+## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
+## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
+[logging.tracing_sample_ratio]
+default_ratio = 1.0
+
+## The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.
+## This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+[export_metrics]
+
+## whether enable export metrics.
+enable = false
+
+## The interval of export metrics.
+write_interval = "30s"
+
+## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+[export_metrics.self_import]
+## +toml2docs:none-default
+db = "information_schema"
+
+[export_metrics.remote_write]
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+url = ""
+
+## HTTP headers of Prometheus remote-write carry.
+headers = { }

config/frontend.example.toml

@@ -1,94 +1,192 @@
-# Node running mode, see `standalone.example.toml`.
+## The running mode of the datanode. It can be `standalone` or `distributed`.
-mode = "distributed"
+mode = "standalone"
-# The default timezone of the server
-# default_timezone = "UTC"
 
+## The default timezone of the server.
+## +toml2docs:none-default
+default_timezone = "UTC"
+
+## The heartbeat options.
 [heartbeat]
-# Interval for sending heartbeat task to the Metasrv, 5 seconds by default.
+## Interval for sending heartbeat messages to the metasrv.
-interval = "5s"
+interval = "18s"
-# Interval for retry sending heartbeat task, 5 seconds by default.
-retry_interval = "5s"
 
-# HTTP server options, see `standalone.example.toml`.
+## Interval for retrying to send heartbeat messages to the metasrv.
+retry_interval = "3s"
+
+## The HTTP server options.
 [http]
+## The address to bind the HTTP server.
 addr = "127.0.0.1:4000"
+## HTTP request timeout.
 timeout = "30s"
+## HTTP request body limit.
+## Support the following units are supported: `B`, `KB`, `KiB`, `MB`, `MiB`, `GB`, `GiB`, `TB`, `TiB`, `PB`, `PiB`.
 body_limit = "64MB"
 
-# gRPC server options, see `standalone.example.toml`.
+## The gRPC server options.
 [grpc]
+## The address to bind the gRPC server.
 addr = "127.0.0.1:4001"
+## The number of server worker threads.
 runtime_size = 8
 
-# MySQL server options, see `standalone.example.toml`.
+## MySQL server options.
 [mysql]
+## Whether to enable.
 enable = true
+## The addr to bind the MySQL server.
 addr = "127.0.0.1:4002"
+## The number of server worker threads.
 runtime_size = 2
 
-# MySQL server TLS options, see `standalone.example.toml`.
+# MySQL server TLS options.
 [mysql.tls]
+
+## TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html
+## - `disable` (default value)
+## - `prefer`
+## - `require`
+## - `verify-ca`
+## - `verify-full`
 mode = "disable"
+
+## Certificate file path.
+## +toml2docs:none-default
 cert_path = ""
+
+## Private key file path.
+## +toml2docs:none-default
 key_path = ""
 
-# PostgresSQL server options, see `standalone.example.toml`.
+## Watch for Certificate and key file change and auto reload
+watch = false
+
+## PostgresSQL server options.
 [postgres]
+## Whether to enable
 enable = true
+## The addr to bind the PostgresSQL server.
 addr = "127.0.0.1:4003"
+## The number of server worker threads.
 runtime_size = 2
 
-# PostgresSQL server TLS options, see `standalone.example.toml`.
+## PostgresSQL server TLS options, see `mysql_options.tls` section.
 [postgres.tls]
+## TLS mode.
 mode = "disable"
+
+## Certificate file path.
+## +toml2docs:none-default
 cert_path = ""
+
+## Private key file path.
+## +toml2docs:none-default
 key_path = ""
 
-# OpenTSDB protocol options, see `standalone.example.toml`.
+## Watch for Certificate and key file change and auto reload
+watch = false
+
+## OpenTSDB protocol options.
 [opentsdb]
+## Whether to enable
 enable = true
+## OpenTSDB telnet API server address.
 addr = "127.0.0.1:4242"
+## The number of server worker threads.
 runtime_size = 2
 
-# InfluxDB protocol options, see `standalone.example.toml`.
+## InfluxDB protocol options.
 [influxdb]
+## Whether to enable InfluxDB protocol in HTTP API.
 enable = true
 
-# Prometheus remote storage options, see `standalone.example.toml`.
+## Prometheus remote storage options
 [prom_store]
+## Whether to enable Prometheus remote write and read in HTTP API.
 enable = true
+## Whether to store the data from Prometheus remote write in metric engine.
+with_metric_engine = true
 
-# Metasrv client options, see `datanode.example.toml`.
+## The metasrv client options.
 [meta_client]
+## The addresses of the metasrv.
 metasrv_addrs = ["127.0.0.1:3002"]
+
+## Operation timeout.
 timeout = "3s"
-# DDL timeouts options.
+
+## Heartbeat timeout.
+heartbeat_timeout = "500ms"
+
+## DDL timeout.
 ddl_timeout = "10s"
+
+## Connect server timeout.
 connect_timeout = "1s"
+
+## `TCP_NODELAY` option for accepted connections.
 tcp_nodelay = true
 
-# Log options, see `standalone.example.toml`
+## The configuration about the cache of the metadata.
-# [logging]
+metadata_cache_max_capacity = 100000
-# dir = "/tmp/greptimedb/logs"
-# level = "info"
 
-# Datanode options.
+## TTL of the metadata cache.
+metadata_cache_ttl = "10m"
+
+# TTI of the metadata cache.
+metadata_cache_tti = "5m"
+
+## Datanode options.
 [datanode]
-# Datanode client options.
+## Datanode client options.
 [datanode.client]
 timeout = "10s"
 connect_timeout = "10s"
 tcp_nodelay = true
 
-# Frontend export the metrics generated by itself
+## The logging options.
-# encoded to Prometheus remote-write format
+[logging]
-# and send to Prometheus remote-write compatible receiver (e.g. send to `greptimedb` itself)
+## The directory to store the log files.
-# This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+dir = "/tmp/greptimedb/logs"
-# [export_metrics]
+
-# whether enable export metrics, default is false
+## The log level. Can be `info`/`debug`/`warn`/`error`.
-# enable = false
+## +toml2docs:none-default
-# The interval of export metrics
+level = "info"
-# write_interval = "30s"
+
-# for `frontend`, `self_import` is recommend to collect metrics generated by itself
+## Enable OTLP tracing.
-# [export_metrics.self_import]
+enable_otlp_tracing = false
-# db = "information_schema"
+
+## The OTLP tracing endpoint.
+## +toml2docs:none-default
+otlp_endpoint = ""
+
+## Whether to append logs to stdout.
+append_stdout = true
+
+## The percentage of tracing will be sampled and exported.
+## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
+## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
+[logging.tracing_sample_ratio]
+default_ratio = 1.0
+
+## The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.
+## This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+[export_metrics]
+
+## whether enable export metrics.
+enable = false
+
+## The interval of export metrics.
+write_interval = "30s"
+
+## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+[export_metrics.self_import]
+## +toml2docs:none-default
+db = "information_schema"
+
+[export_metrics.remote_write]
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+url = ""
+
+## HTTP headers of Prometheus remote-write carry.
+headers = { }

config/metasrv.example.toml

@@ -1,35 +1,46 @@
-# The working home directory.
+## The working home directory.
 data_home = "/tmp/metasrv/"
-# The bind address of metasrv, "127.0.0.1:3002" by default.
+
+## The bind address of metasrv.
 bind_addr = "127.0.0.1:3002"
-# The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost.
+
+## The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost.
 server_addr = "127.0.0.1:3002"
-# Etcd server address, "127.0.0.1:2379" by default.
+
+## Etcd server address.
 store_addr = "127.0.0.1:2379"
-# Datanode selector type.
+
-# - "lease_based" (default value).
+## Datanode selector type.
-# - "load_based"
+## - `lease_based` (default value).
-# For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector".
+## - `load_based`
+## For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector".
 selector = "lease_based"
-# Store data in memory, false by default.
+
+## Store data in memory.
 use_memory_store = false
-# Whether to enable greptimedb telemetry, true by default.
+
+## Whether to enable greptimedb telemetry.
 enable_telemetry = true
-# If it's not empty, the metasrv will store all data with this key prefix.
+
+## If it's not empty, the metasrv will store all data with this key prefix.
 store_key_prefix = ""
 
-# Log options, see `standalone.example.toml`
+## Procedure storage options.
-# [logging]
-# dir = "/tmp/greptimedb/logs"
-# level = "info"
-
-# Procedure storage options.
 [procedure]
-# Procedure max retry time.
+
+## Procedure max retry time.
 max_retry_times = 12
-# Initial retry delay of procedures, increases exponentially
+
+## Initial retry delay of procedures, increases exponentially
 retry_delay = "500ms"
 
+## Auto split large value
+## GreptimeDB procedure uses etcd as the default metadata storage backend.
+## The etcd the maximum size of any request is 1.5 MiB
+## 1500KiB = 1536KiB (1.5MiB) - 36KiB (reserved size of key)
+## Comments out the `max_metadata_value_size`, for don't split large value (no limit).
+max_metadata_value_size = "1500KiB"
+
 # Failure detectors options.
 [failure_detector]
 threshold = 8.0
@@ -38,56 +49,95 @@ acceptable_heartbeat_pause = "3000ms"
 first_heartbeat_estimate = "1000ms"
 
 ## Datanode options.
-# [datanode]
+[datanode]
 ## Datanode client options.
-# [datanode.client_options]
+[datanode.client]
-# timeout = "10s"
+timeout = "10s"
-# connect_timeout = "10s"
+connect_timeout = "10s"
-# tcp_nodelay = true
+tcp_nodelay = true
 
 [wal]
 # Available wal providers:
-# - "raft_engine" (default)
+# - `raft_engine` (default): there're none raft-engine wal config since metasrv only involves in remote wal currently.
-# - "kafka"
+# - `kafka`: metasrv **have to be** configured with kafka wal config when using kafka wal provider in datanode.
 provider = "raft_engine"
 
-# There're none raft-engine wal config since meta srv only involves in remote wal currently.
-
 # Kafka wal config.
-# The broker endpoints of the Kafka cluster. ["127.0.0.1:9092"] by default.
-# broker_endpoints = ["127.0.0.1:9092"]
-# Number of topics to be created upon start.
-# num_topics = 64
-# Topic selector type.
-# Available selector types: 
-# - "round_robin" (default)
-# selector_type = "round_robin"
-# A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.
-# topic_name_prefix = "greptimedb_wal_topic"
-# Expected number of replicas of each partition.
-# replication_factor = 1
-# Above which a topic creation operation will be cancelled.
-# create_topic_timeout = "30s"
-# The initial backoff for kafka clients.
-# backoff_init = "500ms"
-# The maximum backoff for kafka clients.
-# backoff_max = "10s"
-# Exponential backoff rate, i.e. next backoff = base * current backoff.
-# backoff_base = 2
-# Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate.
-# backoff_deadline = "5mins"
 
-# Metasrv export the metrics generated by itself
+## The broker endpoints of the Kafka cluster.
-# encoded to Prometheus remote-write format
+broker_endpoints = ["127.0.0.1:9092"]
-# and send to Prometheus remote-write compatible receiver (e.g. send to `greptimedb` itself)
+
-# This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+## Number of topics to be created upon start.
-# [export_metrics]
+num_topics = 64
-# whether enable export metrics, default is false
+
-# enable = false
+## Topic selector type.
-# The interval of export metrics
+## Available selector types:
-# write_interval = "30s"
+## - `round_robin` (default)
-# [export_metrics.remote_write]
+selector_type = "round_robin"
-# The url the metrics send to. The url is empty by default, url example: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`
+
-# url = ""
+## A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.
-# HTTP headers of Prometheus remote-write carry
+topic_name_prefix = "greptimedb_wal_topic"
-# headers = {}
+
+## Expected number of replicas of each partition.
+replication_factor = 1
+
+## Above which a topic creation operation will be cancelled.
+create_topic_timeout = "30s"
+## The initial backoff for kafka clients.
+backoff_init = "500ms"
+
+## The maximum backoff for kafka clients.
+backoff_max = "10s"
+
+## Exponential backoff rate, i.e. next backoff = base * current backoff.
+backoff_base = 2
+
+## Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate.
+backoff_deadline = "5mins"
+
+## The logging options.
+[logging]
+## The directory to store the log files.
+dir = "/tmp/greptimedb/logs"
+
+## The log level. Can be `info`/`debug`/`warn`/`error`.
+## +toml2docs:none-default
+level = "info"
+
+## Enable OTLP tracing.
+enable_otlp_tracing = false
+
+## The OTLP tracing endpoint.
+## +toml2docs:none-default
+otlp_endpoint = ""
+
+## Whether to append logs to stdout.
+append_stdout = true
+
+## The percentage of tracing will be sampled and exported.
+## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
+## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
+[logging.tracing_sample_ratio]
+default_ratio = 1.0
+
+## The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.
+## This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+[export_metrics]
+
+## whether enable export metrics.
+enable = false
+
+## The interval of export metrics.
+write_interval = "30s"
+
+## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+[export_metrics.self_import]
+## +toml2docs:none-default
+db = "information_schema"
+
+[export_metrics.remote_write]
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+url = ""
+
+## HTTP headers of Prometheus remote-write carry.
+headers = { }

config/standalone.example.toml

@@ -1,175 +1,324 @@
-# Node running mode, "standalone" or "distributed".
+## The running mode of the datanode. It can be `standalone` or `distributed`.
 mode = "standalone"
-# Whether to enable greptimedb telemetry, true by default.
-enable_telemetry = true
-# The default timezone of the server
-# default_timezone = "UTC"
 
-# HTTP server options.
+## Enable telemetry to collect anonymous usage data.
+enable_telemetry = true
+
+## The default timezone of the server.
+## +toml2docs:none-default
+default_timezone = "UTC"
+
+## The HTTP server options.
 [http]
-# Server address, "127.0.0.1:4000" by default.
+## The address to bind the HTTP server.
 addr = "127.0.0.1:4000"
-# HTTP request timeout, 30s by default.
+## HTTP request timeout.
 timeout = "30s"
-# HTTP request body limit, 64Mb by default.
+## HTTP request body limit.
-# the following units are supported: B, KB, KiB, MB, MiB, GB, GiB, TB, TiB, PB, PiB
+## Support the following units are supported: `B`, `KB`, `KiB`, `MB`, `MiB`, `GB`, `GiB`, `TB`, `TiB`, `PB`, `PiB`.
 body_limit = "64MB"
 
-# gRPC server options.
+## The gRPC server options.
 [grpc]
-# Server address, "127.0.0.1:4001" by default.
+## The address to bind the gRPC server.
 addr = "127.0.0.1:4001"
-# The number of server worker threads, 8 by default.
+## The number of server worker threads.
 runtime_size = 8
 
-# MySQL server options.
+## MySQL server options.
 [mysql]
-# Whether to enable
+## Whether to enable.
 enable = true
-# Server address, "127.0.0.1:4002" by default.
+## The addr to bind the MySQL server.
 addr = "127.0.0.1:4002"
-# The number of server worker threads, 2 by default.
+## The number of server worker threads.
 runtime_size = 2
 
 # MySQL server TLS options.
 [mysql.tls]
-# TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html
+
-# - "disable" (default value)
+## TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html
-# - "prefer"
+## - `disable` (default value)
-# - "require"
+## - `prefer`
-# - "verify-ca"
+## - `require`
-# - "verify-full"
+## - `verify-ca`
+## - `verify-full`
 mode = "disable"
-# Certificate file path.
+
+## Certificate file path.
+## +toml2docs:none-default
 cert_path = ""
-# Private key file path.
+
+## Private key file path.
+## +toml2docs:none-default
 key_path = ""
 
-# PostgresSQL server options.
+## Watch for Certificate and key file change and auto reload
+watch = false
+
+## PostgresSQL server options.
 [postgres]
-# Whether to enable
+## Whether to enable
 enable = true
-# Server address, "127.0.0.1:4003" by default.
+## The addr to bind the PostgresSQL server.
 addr = "127.0.0.1:4003"
-# The number of server worker threads, 2 by default.
+## The number of server worker threads.
 runtime_size = 2
 
-# PostgresSQL server TLS options, see `[mysql_options.tls]` section.
+## PostgresSQL server TLS options, see `mysql_options.tls` section.
 [postgres.tls]
-# TLS mode.
+## TLS mode.
 mode = "disable"
-# certificate file path.
+
+## Certificate file path.
+## +toml2docs:none-default
 cert_path = ""
-# private key file path.
+
+## Private key file path.
+## +toml2docs:none-default
 key_path = ""
 
-# OpenTSDB protocol options.
+## Watch for Certificate and key file change and auto reload
+watch = false
+
+## OpenTSDB protocol options.
 [opentsdb]
-# Whether to enable
+## Whether to enable
 enable = true
-# OpenTSDB telnet API server address, "127.0.0.1:4242" by default.
+## OpenTSDB telnet API server address.
 addr = "127.0.0.1:4242"
-# The number of server worker threads, 2 by default.
+## The number of server worker threads.
 runtime_size = 2
 
-# InfluxDB protocol options.
+## InfluxDB protocol options.
 [influxdb]
-# Whether to enable InfluxDB protocol in HTTP API, true by default.
+## Whether to enable InfluxDB protocol in HTTP API.
 enable = true
 
-# Prometheus remote storage options
+## Prometheus remote storage options
 [prom_store]
-# Whether to enable Prometheus remote write and read in HTTP API, true by default.
+## Whether to enable Prometheus remote write and read in HTTP API.
 enable = true
+## Whether to store the data from Prometheus remote write in metric engine.
+with_metric_engine = true
 
+## The WAL options.
 [wal]
-# Available wal providers:
+## The provider of the WAL.
-# - "raft_engine" (default)
+## - `raft_engine`: the wal is stored in the local file system by raft-engine.
-# - "kafka"
+## - `kafka`: it's remote wal that data is stored in Kafka.
 provider = "raft_engine"
 
-# There're none raft-engine wal config since meta srv only involves in remote wal currently.
+## The directory to store the WAL files.
+## **It's only used when the provider is `raft_engine`**.
+## +toml2docs:none-default
+dir = "/tmp/greptimedb/wal"
 
-# Kafka wal options.
+## The size of the WAL segment file.
-# The broker endpoints of the Kafka cluster. ["127.0.0.1:9092"] by default.
+## **It's only used when the provider is `raft_engine`**.
-# broker_endpoints = ["127.0.0.1:9092"]
-
-# Number of topics to be created upon start.
-# num_topics = 64
-# Topic selector type.
-# Available selector types: 
-# - "round_robin" (default)
-# selector_type = "round_robin"
-# The prefix of topic name.
-# topic_name_prefix = "greptimedb_wal_topic"
-# The number of replicas of each partition.
-# Warning: the replication factor must be positive and must not be greater than the number of broker endpoints.
-# replication_factor = 1
-
-# The max size of a single producer batch.
-# Warning: Kafka has a default limit of 1MB per message in a topic.
-# max_batch_size = "1MB"
-# The linger duration.
-# linger = "200ms"
-# The consumer wait timeout.
-# consumer_wait_timeout = "100ms"
-# Create topic timeout.
-# create_topic_timeout = "30s"
-
-# The initial backoff delay.
-# backoff_init = "500ms"
-# The maximum backoff delay.
-# backoff_max = "10s"
-# Exponential backoff rate, i.e. next backoff = base * current backoff.
-# backoff_base = 2
-# The deadline of retries.
-# backoff_deadline = "5mins"
-
-# WAL data directory
-# dir = "/tmp/greptimedb/wal"
-# WAL file size in bytes.
 file_size = "256MB"
-# WAL purge threshold.
+
+## The threshold of the WAL size to trigger a flush.
+## **It's only used when the provider is `raft_engine`**.
 purge_threshold = "4GB"
-# WAL purge interval in seconds.
+
+## The interval to trigger a flush.
+## **It's only used when the provider is `raft_engine`**.
 purge_interval = "10m"
-# WAL read batch size.
-read_batch_size = 128
-# Whether to sync log file after every write.
-sync_write = false
-# Whether to reuse logically truncated log files.
-enable_log_recycle = true
-# Whether to pre-create log files on start up
-prefill_log_files = false
-# Duration for fsyncing log files.
-sync_period = "1000ms"
 
-# Metadata storage options.
+## The read batch size.
+## **It's only used when the provider is `raft_engine`**.
+read_batch_size = 128
+
+## Whether to use sync write.
+## **It's only used when the provider is `raft_engine`**.
+sync_write = false
+
+## Whether to reuse logically truncated log files.
+## **It's only used when the provider is `raft_engine`**.
+enable_log_recycle = true
+
+## Whether to pre-create log files on start up.
+## **It's only used when the provider is `raft_engine`**.
+prefill_log_files = false
+
+## Duration for fsyncing log files.
+## **It's only used when the provider is `raft_engine`**.
+sync_period = "10s"
+
+## The Kafka broker endpoints.
+## **It's only used when the provider is `kafka`**.
+broker_endpoints = ["127.0.0.1:9092"]
+
+## The max size of a single producer batch.
+## Warning: Kafka has a default limit of 1MB per message in a topic.
+## **It's only used when the provider is `kafka`**.
+max_batch_size = "1MB"
+
+## The linger duration of a kafka batch producer.
+## **It's only used when the provider is `kafka`**.
+linger = "200ms"
+
+## The consumer wait timeout.
+## **It's only used when the provider is `kafka`**.
+consumer_wait_timeout = "100ms"
+
+## The initial backoff delay.
+## **It's only used when the provider is `kafka`**.
+backoff_init = "500ms"
+
+## The maximum backoff delay.
+## **It's only used when the provider is `kafka`**.
+backoff_max = "10s"
+
+## The exponential backoff rate, i.e. next backoff = base * current backoff.
+## **It's only used when the provider is `kafka`**.
+backoff_base = 2
+
+## The deadline of retries.
+## **It's only used when the provider is `kafka`**.
+backoff_deadline = "5mins"
+
+## Metadata storage options.
 [metadata_store]
-# Kv file size in bytes.
+## Kv file size in bytes.
 file_size = "256MB"
-# Kv purge threshold.
+## Kv purge threshold.
 purge_threshold = "4GB"
 
-# Procedure storage options.
+## Procedure storage options.
 [procedure]
-# Procedure max retry time.
+## Procedure max retry time.
 max_retry_times = 3
-# Initial retry delay of procedures, increases exponentially
+## Initial retry delay of procedures, increases exponentially
 retry_delay = "500ms"
 
-# Storage options.
+# Example of using S3 as the storage.
+# [storage]
+# type = "S3"
+# bucket = "greptimedb"
+# root = "data"
+# access_key_id = "test"
+# secret_access_key = "123456"
+# endpoint = "https://s3.amazonaws.com"
+# region = "us-west-2"
+
+# Example of using Oss as the storage.
+# [storage]
+# type = "Oss"
+# bucket = "greptimedb"
+# root = "data"
+# access_key_id = "test"
+# access_key_secret = "123456"
+# endpoint = "https://oss-cn-hangzhou.aliyuncs.com"
+
+# Example of using Azblob as the storage.
+# [storage]
+# type = "Azblob"
+# container = "greptimedb"
+# root = "data"
+# account_name = "test"
+# account_key = "123456"
+# endpoint = "https://greptimedb.blob.core.windows.net"
+# sas_token = ""
+
+# Example of using Gcs as the storage.
+# [storage]
+# type = "Gcs"
+# bucket = "greptimedb"
+# root = "data"
+# scope = "test"
+# credential_path = "123456"
+# endpoint = "https://storage.googleapis.com"
+
+## The data storage options.
 [storage]
-# The working home directory.
+## The working home directory.
 data_home = "/tmp/greptimedb/"
-# Storage type.
+
+## The storage type used to store the data.
+## - `File`: the data is stored in the local file system.
+## - `S3`: the data is stored in the S3 object storage.
+## - `Gcs`: the data is stored in the Google Cloud Storage.
+## - `Azblob`: the data is stored in the Azure Blob Storage.
+## - `Oss`: the data is stored in the Aliyun OSS.
 type = "File"
-# TTL for all tables. Disabled by default.
+
-# global_ttl = "7d"
+## Cache configuration for object storage such as 'S3' etc.
-# Cache configuration for object storage such as 'S3' etc.
+## The local file cache directory.
-# cache_path = "/path/local_cache"
+## +toml2docs:none-default
-# The local file cache capacity in bytes.
+cache_path = "/path/local_cache"
-# cache_capacity = "256MB"
+
+## The local file cache capacity in bytes.
+## +toml2docs:none-default
+cache_capacity = "256MB"
+
+## The S3 bucket name.
+## **It's only used when the storage type is `S3`, `Oss` and `Gcs`**.
+## +toml2docs:none-default
+bucket = "greptimedb"
+
+## The S3 data will be stored in the specified prefix, for example, `s3://${bucket}/${root}`.
+## **It's only used when the storage type is `S3`, `Oss` and `Azblob`**.
+## +toml2docs:none-default
+root = "greptimedb"
+
+## The access key id of the aws account.
+## It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.
+## **It's only used when the storage type is `S3` and `Oss`**.
+## +toml2docs:none-default
+access_key_id = "test"
+
+## The secret access key of the aws account.
+## It's **highly recommended** to use AWS IAM roles instead of hardcoding the access key id and secret key.
+## **It's only used when the storage type is `S3`**.
+## +toml2docs:none-default
+secret_access_key = "test"
+
+## The secret access key of the aliyun account.
+## **It's only used when the storage type is `Oss`**.
+## +toml2docs:none-default
+access_key_secret = "test"
+
+## The account key of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+account_name = "test"
+
+## The account key of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+account_key = "test"
+
+## The scope of the google cloud storage.
+## **It's only used when the storage type is `Gcs`**.
+## +toml2docs:none-default
+scope = "test"
+
+## The credential path of the google cloud storage.
+## **It's only used when the storage type is `Gcs`**.
+## +toml2docs:none-default
+credential_path = "test"
+
+## The container of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+container = "greptimedb"
+
+## The sas token of the azure account.
+## **It's only used when the storage type is `Azblob`**.
+## +toml2docs:none-default
+sas_token = ""
+
+## The endpoint of the S3 service.
+## **It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**.
+## +toml2docs:none-default
+endpoint = "https://s3.amazonaws.com"
+
+## The region of the S3 service.
+## **It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**.
+## +toml2docs:none-default
+region = "us-west-2"
 
 # Custom storage options
 # [[storage.providers]]
@@ -177,69 +326,152 @@ type = "File"
 # [[storage.providers]]
 # type = "Gcs"
 
-# Mito engine options
+## The region engine options. You can configure multiple region engines.
 [[region_engine]]
+
+## The Mito engine options.
 [region_engine.mito]
-# Number of region workers
+
+## Number of region workers.
 num_workers = 8
-# Request channel size of each worker
+
+## Request channel size of each worker.
 worker_channel_size = 128
-# Max batch size for a worker to handle requests
+
+## Max batch size for a worker to handle requests.
 worker_request_batch_size = 64
-# Number of meta action updated to trigger a new checkpoint for the manifest
+
+## Number of meta action updated to trigger a new checkpoint for the manifest.
 manifest_checkpoint_distance = 10
-# Whether to compress manifest and checkpoint file by gzip (default false).
+
+## Whether to compress manifest and checkpoint file by gzip (default false).
 compress_manifest = false
-# Max number of running background jobs
+
+## Max number of running background jobs
 max_background_jobs = 4
-# Interval to auto flush a region if it has not flushed yet.
+
+## Interval to auto flush a region if it has not flushed yet.
 auto_flush_interval = "1h"
-# Global write buffer size for all regions.
+
+## Global write buffer size for all regions. If not set, it's default to 1/8 of OS memory with a max limitation of 1GB.
 global_write_buffer_size = "1GB"
-# Global write buffer size threshold to reject write requests (default 2G).
+
+## Global write buffer size threshold to reject write requests. If not set, it's default to 2 times of `global_write_buffer_size`
 global_write_buffer_reject_size = "2GB"
-# Cache size for SST metadata (default 128MB). Setting it to 0 to disable the cache.
+
+## Cache size for SST metadata. Setting it to 0 to disable the cache.
+## If not set, it's default to 1/32 of OS memory with a max limitation of 128MB.
 sst_meta_cache_size = "128MB"
-# Cache size for vectors and arrow arrays (default 512MB). Setting it to 0 to disable the cache.
+
+## Cache size for vectors and arrow arrays. Setting it to 0 to disable the cache.
+## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
 vector_cache_size = "512MB"
-# Cache size for pages of SST row groups (default 512MB). Setting it to 0 to disable the cache.
+
+## Cache size for pages of SST row groups. Setting it to 0 to disable the cache.
+## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
 page_cache_size = "512MB"
-# Buffer size for SST writing.
+
+## Buffer size for SST writing.
 sst_write_buffer_size = "8MB"
-# Parallelism to scan a region (default: 1/4 of cpu cores).
+
-# - 0: using the default value (1/4 of cpu cores).
+## Parallelism to scan a region (default: 1/4 of cpu cores).
-# - 1: scan in current thread.
+## - `0`: using the default value (1/4 of cpu cores).
-# - n: scan in parallelism n.
+## - `1`: scan in current thread.
+## - `n`: scan in parallelism n.
 scan_parallelism = 0
-# Capacity of the channel to send data from parallel scan tasks to the main task (default 32).
+
+## Capacity of the channel to send data from parallel scan tasks to the main task.
 parallel_scan_channel_size = 32
-# Whether to allow stale WAL entries read during replay.
+
+## Whether to allow stale WAL entries read during replay.
 allow_stale_entries = false
 
-# Log options
+## The options for inverted index in Mito engine.
-# [logging]
+[region_engine.mito.inverted_index]
-# Specify logs directory.
-# dir = "/tmp/greptimedb/logs"
-# Specify the log level [info | debug | error | warn]
-# level = "info"
-# whether enable tracing, default is false
-# enable_otlp_tracing = false
-# tracing exporter endpoint with format `ip:port`, we use grpc oltp as exporter, default endpoint is `localhost:4317`
-# otlp_endpoint = "localhost:4317"
-# The percentage of tracing will be sampled and exported. Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1. ratio > 1 are treated as 1. Fractions < 0 are treated as 0
-# tracing_sample_ratio = 1.0
-# Whether to append logs to stdout. Defaults to true.
-# append_stdout = true
 
-# Standalone export the metrics generated by itself
+## Whether to create the index on flush.
-# encoded to Prometheus remote-write format
+## - `auto`: automatically
-# and send to Prometheus remote-write compatible receiver (e.g. send to `greptimedb` itself)
+## - `disable`: never
-# This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+create_on_flush = "auto"
-# [export_metrics]
+
-# whether enable export metrics, default is false
+## Whether to create the index on compaction.
-# enable = false
+## - `auto`: automatically
-# The interval of export metrics
+## - `disable`: never
-# write_interval = "30s"
+create_on_compaction = "auto"
-# for `standalone`, `self_import` is recommend to collect metrics generated by itself
+
-# [export_metrics.self_import]
+## Whether to apply the index on query
-# db = "information_schema"
+## - `auto`: automatically
+## - `disable`: never
+apply_on_query = "auto"
+
+## Memory threshold for performing an external sort during index creation.
+## Setting to empty will disable external sorting, forcing all sorting operations to happen in memory.
+mem_threshold_on_create = "64M"
+
+## File system path to store intermediate files for external sorting (default `{data_home}/index_intermediate`).
+intermediate_path = ""
+
+[region_engine.mito.memtable]
+## Memtable type.
+## - `time_series`: time-series memtable
+## - `partition_tree`: partition tree memtable (experimental)
+type = "time_series"
+
+## The max number of keys in one shard.
+## Only available for `partition_tree` memtable.
+index_max_keys_per_shard = 8192
+
+## The max rows of data inside the actively writing buffer in one shard.
+## Only available for `partition_tree` memtable.
+data_freeze_threshold = 32768
+
+## Max dictionary bytes.
+## Only available for `partition_tree` memtable.
+fork_dictionary_bytes = "1GiB"
+
+## The logging options.
+[logging]
+## The directory to store the log files.
+dir = "/tmp/greptimedb/logs"
+
+## The log level. Can be `info`/`debug`/`warn`/`error`.
+## +toml2docs:none-default
+level = "info"
+
+## Enable OTLP tracing.
+enable_otlp_tracing = false
+
+## The OTLP tracing endpoint.
+## +toml2docs:none-default
+otlp_endpoint = ""
+
+## Whether to append logs to stdout.
+append_stdout = true
+
+## The percentage of tracing will be sampled and exported.
+## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
+## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
+[logging.tracing_sample_ratio]
+default_ratio = 1.0
+
+## The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.
+## This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape.
+[export_metrics]
+
+## whether enable export metrics.
+enable = false
+
+## The interval of export metrics.
+write_interval = "30s"
+
+## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+[export_metrics.self_import]
+## +toml2docs:none-default
+db = "information_schema"
+
+[export_metrics.remote_write]
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+url = ""
+
+## HTTP headers of Prometheus remote-write carry.
+headers = { }

docker/ci/ubuntu/Dockerfile

@@ -1,5 +1,11 @@
 FROM ubuntu:22.04
 
+# The root path under which contains all the dependencies to build this Dockerfile.
+ARG DOCKER_BUILD_ROOT=.
+# The binary name of GreptimeDB executable.
+# Defaults to "greptime", but sometimes in other projects it might be different.
+ARG TARGET_BIN=greptime
+
 RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
     ca-certificates \
     python3.10 \
@@ -7,14 +13,16 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
     python3-pip \
     curl
 
-COPY ./docker/python/requirements.txt /etc/greptime/requirements.txt
+COPY $DOCKER_BUILD_ROOT/docker/python/requirements.txt /etc/greptime/requirements.txt
 
 RUN python3 -m pip install -r /etc/greptime/requirements.txt
 
 ARG TARGETARCH
 
-ADD $TARGETARCH/greptime /greptime/bin/
+ADD $TARGETARCH/$TARGET_BIN /greptime/bin/
 
 ENV PATH /greptime/bin/:$PATH
 
-ENTRYPOINT ["greptime"]
+ENV TARGET_BIN=$TARGET_BIN
+
+ENTRYPOINT ["sh", "-c", "exec $TARGET_BIN \"$@\"", "--"]

docker/dev-builder/ubuntu/Dockerfile

@@ -1,5 +1,8 @@
 FROM ubuntu:20.04
 
+# The root path under which contains all the dependencies to build this Dockerfile.
+ARG DOCKER_BUILD_ROOT=.
+
 ENV LANG en_US.utf8
 WORKDIR /greptimedb
 
@@ -27,10 +30,20 @@ RUN apt-get -y purge python3.8 && \
     ln -s /usr/bin/python3.10 /usr/bin/python3 && \
     curl -sS https://bootstrap.pypa.io/get-pip.py | python3.10
 
-RUN git config --global --add safe.directory /greptimedb
+# Silence all `safe.directory` warnings, to avoid the "detect dubious repository" error when building with submodules.
+# Disabling the safe directory check here won't pose extra security issues, because in our usage for this dev build
+# image, we use it solely on our own environment (that github action's VM, or ECS created dynamically by ourselves),
+# and the repositories are pulled from trusted sources (still us, of course). Doing so does not violate the intention
+# of the Git's addition to the "safe.directory" at the first place (see the commit message here:
+# https://github.com/git/git/commit/8959555cee7ec045958f9b6dd62e541affb7e7d9).
+# There's also another solution to this, that we add the desired submodules to the safe directory, instead of using 
+# wildcard here. However, that requires the git's config files and the submodules all owned by the very same user.
+# It's troublesome to do this since the dev build runs in Docker, which is under user "root"; while outside the Docker,
+# it can be a different user that have prepared the submodules.
+RUN git config --global --add safe.directory *
 
 # Install Python dependencies.
-COPY ./docker/python/requirements.txt /etc/greptime/requirements.txt
+COPY $DOCKER_BUILD_ROOT/docker/python/requirements.txt /etc/greptime/requirements.txt
 RUN python3 -m pip install -r /etc/greptime/requirements.txt
 
 # Install Rust.

docs/benchmarks/tsbs/v0.7.0.md

@@ -0,0 +1,50 @@
+# TSBS benchmark - v0.7.0
+
+## Environment
+
+### Local
+|        |                                    |
+| ------ | ---------------------------------- |
+| CPU    | AMD Ryzen 7 7735HS (8 core 3.2GHz) |
+| Memory | 32GB                               |
+| Disk   | SOLIDIGM SSDPFKNU010TZ             |
+| OS     | Ubuntu 22.04.2 LTS                 |
+
+### Amazon EC2
+
+|         |                |
+| ------- | -------------- |
+| Machine | c5d.2xlarge    |
+| CPU     | 8 core         |
+| Memory  | 16GB           |
+| Disk    | 50GB (GP3)     |
+| OS      | Ubuntu 22.04.1 |
+
+
+## Write performance
+
+| Environment        | Ingest rate (rows/s)  |
+| ------------------ | --------------------- |
+| Local              | 3695814.64            |
+| EC2 c5d.2xlarge    | 2987166.64            |
+
+
+## Query performance
+
+| Query type            | Local (ms) | EC2 c5d.2xlarge (ms)   |
+| --------------------- | ---------- | ---------------------- |
+| cpu-max-all-1         | 30.56      | 54.74                  |
+| cpu-max-all-8         | 52.69      | 70.50                  |
+| double-groupby-1      | 664.30     | 1366.63                |
+| double-groupby-5      | 1391.26    | 2141.71                |
+| double-groupby-all    | 2828.94    | 3389.59                |
+| groupby-orderby-limit | 718.92     | 1213.90                |
+| high-cpu-1            | 29.21      | 52.98                  |
+| high-cpu-all          | 5514.12    | 7194.91                |
+| lastpoint             | 7571.40    | 9423.41                |
+| single-groupby-1-1-1  | 19.09      | 7.77                   |
+| single-groupby-1-1-12 | 27.28      | 51.64                  |
+| single-groupby-1-8-1  | 31.85      | 11.64                  |
+| single-groupby-5-1-1  | 16.14      | 9.67                   |
+| single-groupby-5-1-12 | 27.21      | 53.62                  |
+| single-groupby-5-8-1  | 39.62      | 14.96                  |

docs/rfcs/2023-05-09-distributed-planner.md

@@ -79,7 +79,7 @@ This RFC proposes to add a new expression node `MergeScan` to merge result from
 │               │    │                             │
 └─Frontend──────┘    └─Remote-Sources──────────────┘
 ```
-This merge operation simply chains all the the underlying remote data sources and return `RecordBatch`, just like a coalesce op. And each remote sources is a gRPC query to datanode via the substrait logical plan interface. The plan is transformed and divided from the original query that comes to frontend.
+This merge operation simply chains all the underlying remote data sources and return `RecordBatch`, just like a coalesce op. And each remote sources is a gRPC query to datanode via the substrait logical plan interface. The plan is transformed and divided from the original query that comes to frontend.
 
 ## Commutativity of MergeScan
 

docs/rfcs/2023-07-06-table-engine-refactor.md

@@ -27,8 +27,8 @@ subgraph Frontend["Frontend"]
     end
 end
 
-MyTable --> MetaSrv
+MyTable --> Metasrv
-MetaSrv --> ETCD
+Metasrv --> ETCD
 
 MyTable-->TableEngine0
 MyTable-->TableEngine1
@@ -95,8 +95,8 @@ subgraph Frontend["Frontend"]
     end
 end
 
-MyTable --> MetaSrv
+MyTable --> Metasrv
-MetaSrv --> ETCD
+Metasrv --> ETCD
 
 MyTable-->RegionEngine
 MyTable-->RegionEngine1

docs/rfcs/2023-11-03-inverted-index.md

@@ -1,6 +1,6 @@
 ---
 Feature Name: Inverted Index for SST File
-Tracking Issue: TBD
+Tracking Issue: https://github.com/GreptimeTeam/greptimedb/issues/2705
 Date: 2023-11-03
 Author: "Zhong Zhenchi <zhongzc_arch@outlook.com>"
 ---

docs/rfcs/2024-01-17-dataflow-framework.md

@@ -0,0 +1,97 @@
+---
+Feature Name: Dataflow Framework
+Tracking Issue: https://github.com/GreptimeTeam/greptimedb/issues/3187
+Date: 2024-01-17
+Author: "Discord9 <discord9@163.com>"
+---
+
+# Summary
+This RFC proposes a Lightweight Module for executing continuous aggregation queries on a stream of data.
+
+# Motivation
+Being able to do continuous aggregation is a very powerful tool. It allows you to do things like:
+1. downsample data from i.e. 1 milliseconds to 1 second
+2. calculate the average of a stream of data
+3. Keeping a sliding window of data in memory
+In order to do those things while maintaining a low memory footprint, you need to be able to manage the data in a smart way. Hence, we only store necessary data in memory, and send/recv data deltas to/from the client.
+
+# Details
+
+## System boundary / What it's and isn't
+- GreptimeFlow provides a way to perform continuous aggregation over time-series data.
+- It's not a complete streaming-processing system. Only a must subset functionalities are provided.
+- Flow can process a configured range of fresh data. Data exceeding this range will be dropped directly. Thus it cannot handle random datasets (random on timestamp).
+- Both sliding windows (e.g., latest 5m from present) and fixed windows (every 5m from some time) are supported. And these two are the major targeting scenarios.
+- Flow can handle most aggregate operators within one table(i.e. Sum, avg, min, max and comparison operators). But others (join, trigger, txn etc.) are not the target feature.
+
+## Framework
+- Greptime Flow's is built on top of [Hydroflow](https://github.com/hydro-project/hydroflow).
+- We have three choices for the Dataflow/Streaming process framework for our simple continuous aggregation feature:
+1. Based on the timely/differential dataflow crate that [materialize](https://github.com/MaterializeInc/materialize) based on. Later, it's proved too obscure for a simple usage, and is hard to customize memory usage control.
+2. Based on a simple dataflow framework that we write from ground up, like what [arroyo](https://www.arroyo.dev/) or [risingwave](https://www.risingwave.dev/) did, for example the core streaming logic of [arroyo](https://github.com/ArroyoSystems/arroyo/blob/master/arroyo-datastream/src/lib.rs) only takes up to 2000 line of codes. However, it means maintaining another layer of dataflow framework, which might seem easy in the beginning, but I fear it might be too burdensome to maintain once we need more features.
+3. Based on a simple and lower level dataflow framework that someone else write, like [hydroflow](https://github.com/hydro-project/hydroflow), this approach combines the best of both worlds. Firstly, it boasts ease of comprehension and customization. Secondly, the dataflow framework offers precisely the necessary features for crafting uncomplicated single-node dataflow programs while delivering decent performance.
+
+Hence, we choose the third option, and use a simple logical plan that's anagonistic to the underlying dataflow framework, as it only describe how the dataflow graph should be doing, not how it do that. And we built operator in hydroflow to execute the plan. And the result hydroflow graph is wrapped in a engine that only support data in/out and tick event to flush and compute the result. This provide a thin middle layer that's easy to maintain and allow switching to other dataflow framework if necessary.
+
+## Deploy mode and protocol
+- Greptime Flow is an independent streaming compute component. It can be used either within a standalone node or as a dedicated node at the same level as frontend in distributed mode.
+- It accepts insert request Rows, which is used between frontend and datanode.
+- New flow job is submitted in the format of modified SQL query like snowflake do, like: `CREATE TASK avg_over_5m WINDOW_SIZE = "5m" AS SELECT avg(value) FROM table WHERE time > now() - 5m GROUP BY time(1m)`. Flow job then got stored in Metasrv.
+- It also persists results in the format of Rows to frontend.
+- The query plan uses Substrait as codec format. It's the same with GreptimeDB's query engine.
+- Greptime Flow needs a WAL for recovering. It's possible to reuse datanode's.
+
+The workflow is shown in the following diagram
+```mermaid
+graph TB
+subgraph Flownode["Flownode"]
+    subgraph Dataflows
+        df1("Dataflow_1")
+        df2("Dataflow_2")
+    end
+end
+subgraph Frontend["Frontend"]
+    newLines["Mirror Insert
+Create Task From Query
+Write result from flow node"]
+end
+
+subgraph Datanode["Datanode"]
+end
+
+User --> Frontend
+Frontend -->|Register Task| Metasrv
+Metasrv -->|Read Task Metadata| Frontend
+Frontend -->|Create Task| Flownode
+
+Frontend -->|Mirror Insert| Flownode
+Flownode -->|Write back| Frontend
+
+Frontend --> Datanode
+Datanode --> Frontend
+
+```
+
+## Lifecycle of data
+- New data is inserted into frontend like before. Frontend will mirror insert request to Flow node if there is configured flow job.
+- Depending on the timestamp of incoming data, flow will either drop it (outdated data) or process it (fresh data).
+- Greptime Flow will periodically write results back to the result table through frontend.
+- Those result will then be written into a result table stored in datanode.
+- A small table of intermediate state is kept in memory, which is used to calculate the result.
+## Supported operations
+- Greptime Flow accepts a configurable "materialize window", data point exceeds that time window is discarded.
+- Data within that "materialize window" is queryable and updateable.
+- Greptime Flow can handle partitioning, if and only if the input query can be transformed to a fully partitioned plan according to the existing commutative rules. Otherwise the corresponding flow job has to be calculated in a single node.
+- Notice that Greptime Flow has to see all the data belongs to one partition.
+- Deletion and duplicate insertion are not supported at early stage.
+## Miscellaneous 
+- Greptime Flow can translate SQL to it's own plan, however only a selected few aggregate function is supported for now, like min/max/sum/count/avg
+- Greptime Flow's operator is configurable in terms of the size of the materialize window, whether to allow delay of incoming data etc., so simplest operator can choose to not tolerate any delay to save memory.
+
+# Future Work
+- Support UDF that can do one-to-one mapping. Preferably, we can reuse the UDF mechanism in GreptimeDB.
+- Support join operator.
+- Design syntax for config operator for different materialize window and delay tolerance.
+- Support cross partition merge operator that allows complex query plan that not necessary accord with partitioning rule to communicate between nodes and create final materialize result.
+- Duplicate insertion, which can be reverted easily within the current framework, so supporting it could be easy
+- Deletion within "materialize window", this requires operators like min/max to store all inputs within materialize window, which might require further optimization.

docs/rfcs/2024-02-21-multi-dimension-partition-rule/rfc.md

@@ -0,0 +1,101 @@
+---
+Feature Name:  Multi-dimension Partition Rule
+Tracking Issue: https://github.com/GreptimeTeam/greptimedb/issues/3351
+Date: 2024-02-21
+Author: "Ruihang Xia <waynestxia@gmail.com>"
+---
+
+# Summary
+
+A new region partition scheme that runs on multiple dimensions of the key space. The partition rule is defined by a set of simple expressions on the partition key columns.
+
+# Motivation
+
+The current partition rule is from MySQL's [`RANGE Partition`](https://dev.mysql.com/doc/refman/8.0/en/partitioning-range.html), which is based on a single dimension. It is sort of a [Hilbert Curve](https://en.wikipedia.org/wiki/Hilbert_curve) and pick several point on the curve to divide the space. It is neither easy to understand how the data get partitioned nor flexible enough to handle complex partitioning requirements.
+
+Considering the future requirements like region repartitioning or autonomous rebalancing, where both workload and partition may change frequently. Here proposes a new region partition scheme that uses a set of simple expressions on the partition key columns to divide the key space.
+
+# Details
+
+## Partition rule
+
+First, we define a simple expression that can be used to define the partition rule. The simple expression is a binary expression expression on the partition key columns that can be evaluated to a boolean value. The binary operator is limited to comparison operators only, like `=`, `!=`, `>`, `>=`, `<`, `<=`. And the operands are limited either literal value or partition column.
+
+Example of valid simple expressions are $`col_A = 10`$, $`col_A \gt 10 \& col_B \gt 20`$ or $`col_A \ne 10`$.
+
+Those expressions can be used as predicates to divide the key space into different regions. The following example have two partition columns `Col A` and `Col B`, and four partitioned regions.
+
+```math
+\left\{\begin{aligned}
+ 
+&col_A \le 10 &Region_1 \\
+&10 \lt col_A \& col_A \le 20 &Region_2 \\
+&20 \lt col_A \space \& \space col_B \lt 100 &Region_3 \\
+&20 \lt col_A \space \& \space col_B \ge 100 &Region_4
+
+\end{aligned}\right\}
+```
+
+An advantage of this scheme is that it is easy to understand how the data get partitioned. The above example can be visualized in a 2D space (two partition column is involved in the example).
+
+![example](2d-example.png)
+
+Here each expression draws a line in the 2D space. Managing data partitioning becomes a matter of drawing lines in the key space.
+
+To make it easy to use, there is a "default region" which catches all the data that doesn't match any of previous expressions. The default region exist by default and do not need to specify. It is also possible to remove this default region if the DB finds it is not necessary.
+
+## SQL interface
+
+The SQL interface is in response to two parts: specifying the partition columns and the partition rule. Thouth we are targeting an autonomous system, it's still allowed to give some bootstrap rules or hints on creating table.
+
+Partition column is specified by `PARTITION ON COLUMNS` sub-clause in `CREATE TABLE`:
+
+```sql
+CREATE TABLE t (...)
+PARTITION ON COLUMNS (...) ();
+```
+
+Two following brackets are for partition columns and partition rule respectively.
+
+Columns provided here are only used as an allow-list of how the partition rule can be defined. Which means (a) the sequence between columns doesn't matter, (b) the columns provided here are not necessarily being used in the partition rule.
+
+The partition rule part is a list of comma-separated simple expressions. Expressions here are not corresponding to region, as they might be changed by system to fit various workload.
+
+A full example of `CREATE TABLE` with partition rule is:
+
+```sql
+CREATE TABLE IF NOT EXISTS demo (
+  a STRING,
+  b STRING,
+  c STRING,
+  d STRING,
+  ts TIMESTAMP,
+  memory DOUBLE,
+  TIME INDEX (ts),
+  PRIMARY KEY (a, b, c, d)
+)
+PARTITION ON COLUMNS (c, b, a) (
+  a < 10,
+  10 >= a AND a < 20,
+  20 >= a AND b < 100,
+  20 >= a AND b > 100
+)
+```
+
+## Combine with storage
+
+Examining columns separately suits our columnar storage very well in two aspects.
+
+1. The simple expression can be pushed down to storage and file format, and is likely to hit existing index. Makes pruning operation very efficient.
+
+2. Columns in columnar storage are not tightly coupled like in the traditional row storages, which means we can easily add or remove columns from partition rule without much impact (like a global reshuffle) on data.
+
+The data file itself can be "projected" to the key space as a polyhedron, it is guaranteed that each plane is in parallel with some coordinate planes (in a 2D scenario, this is saying that all the files can be projected to a rectangle). Thus partition or repartition also only need to consider related columns.
+
+![sst-project](sst-project.png)
+
+An additional limitation is that considering how the index works and how we organize the primary keys at present, the partition columns are limited to be a subset of primary keys for better performance.
+
+# Drawbacks
+
+This is a breaking change.

docs/style-guide.md

@@ -0,0 +1,46 @@
+# GreptimeDB Style Guide
+
+This style guide is intended to help contributors to GreptimeDB write code that is consistent with the rest of the codebase. It is a living document and will be updated as the codebase evolves.
+
+It's mainly an complement to the [Rust Style Guide](https://pingcap.github.io/style-guide/rust/).
+
+## Table of Contents
+
+- Formatting
+- Modules
+- Comments
+
+## Formatting
+
+- Place all `mod` declaration before any `use`.
+- Use `unimplemented!()` instead of `todo!()` for things that aren't likely to be implemented.
+- Add an empty line before and after declaration blocks.
+- Place comment before attributes (`#[]`) and derive (`#[derive]`).
+
+## Modules
+
+- Use the file with same name instead of `mod.rs` to define a module. E.g.:
+
+```
+.
+├── cache
+│  ├── cache_size.rs
+│  └── write_cache.rs
+└── cache.rs
+```
+
+## Comments
+
+- Add comments for public functions and structs.
+- Prefer document comment (`///`) over normal comment (`//`) for structs, fields, functions etc.
+- Add link (`[]`) to struct, method, or any other reference. And make sure that link works.
+
+## Error handling
+
+- Define a custom error type for the module if needed.
+- Prefer `with_context()` over `context()` when allocation is needed to construct an error.
+- Use `error!()` or `warn!()` macros in the `common_telemetry` crate to log errors. E.g.:
+
+```rust
+error!(e; "Failed to do something");
+```

grafana/README.md

@@ -0,0 +1,10 @@
+Grafana dashboard for GreptimeDB
+--------------------------------
+
+GreptimeDB's official Grafana dashboard.
+
+Status notify: we are still working on this config. It's expected to change frequently in the recent days. Please feel free to submit your feedback and/or contribution to this dashboard 🤗
+
+# How to use
+
+Open Grafana Dashboard page, choose `New` -> `Import`. And upload `greptimedb.json` file.

licenserc.toml

@@ -19,6 +19,12 @@ includes = [
     "*.py",
 ]
 
+excludes = [
+    # copied sources
+    "src/common/base/src/readable_size.rs",
+    "src/servers/src/repeated_field.rs",
+]
+
 [properties]
 inceptionYear = 2023
 copyrightOwner = "Greptime Team"

rust-toolchain.toml

@@ -1,2 +1,2 @@
 [toolchain]
-channel = "nightly-2023-12-19"
+channel = "nightly-2024-04-18"

scripts/fetch-dashboard-assets.sh

@@ -1,8 +1,7 @@
 #!/usr/bin/env bash
 
 # This script is used to download built dashboard assets from the "GreptimeTeam/dashboard" repository.
-
+set -ex
-set -e -x
 
 declare -r SCRIPT_DIR=$(cd $(dirname ${0}) >/dev/null 2>&1 && pwd)
 declare -r ROOT_DIR=$(dirname ${SCRIPT_DIR})
@@ -13,13 +12,34 @@ RELEASE_VERSION="$(cat $STATIC_DIR/VERSION | tr -d '\t\r\n ')"
 
 echo "Downloading assets to dir: $OUT_DIR"
 cd $OUT_DIR
+
+if [[ -z "$GITHUB_PROXY_URL" ]]; then
+  GITHUB_URL="https://github.com"
+else
+  GITHUB_URL="${GITHUB_PROXY_URL%/}"
+fi
+
+function retry_fetch() {
+    local url=$1
+    local filename=$2
+
+    curl --connect-timeout 10 --retry 3 -fsSL $url --output $filename || {
+        echo "Failed to download $url"
+        echo "You may try to set http_proxy and https_proxy environment variables."
+        if [[ -z "$GITHUB_PROXY_URL" ]]; then
+          echo "You may try to set GITHUB_PROXY_URL=http://mirror.ghproxy.com/https://github.com/"
+        fi
+        exit 1
+     }
+}
+
 # Download the SHA256 checksum attached to the release. To verify the integrity
 # of the download, this checksum will be used to check the download tar file
 # containing the built dashboard assets.
-curl -Ls https://github.com/GreptimeTeam/dashboard/releases/download/$RELEASE_VERSION/sha256.txt --output sha256.txt
+retry_fetch "${GITHUB_URL}/GreptimeTeam/dashboard/releases/download/${RELEASE_VERSION}/sha256.txt" sha256.txt
 
 # Download the tar file containing the built dashboard assets.
-curl -L https://github.com/GreptimeTeam/dashboard/releases/download/$RELEASE_VERSION/build.tar.gz --output build.tar.gz
+retry_fetch "${GITHUB_URL}/GreptimeTeam/dashboard/releases/download/${RELEASE_VERSION}/build.tar.gz" build.tar.gz
 
 # Verify the checksums match; exit if they don't.
 case "$(uname -s)" in

src/api/Cargo.toml

@@ -4,6 +4,9 @@ version.workspace = true
 edition.workspace = true
 license.workspace = true
 
+[lints]
+workspace = true
+
 [dependencies]
 common-base.workspace = true
 common-decimal.workspace = true
@@ -15,7 +18,6 @@ greptime-proto.workspace = true
 paste = "1.0"
 prost.workspace = true
 snafu.workspace = true
-tonic.workspace = true
 
 [build-dependencies]
 tonic-build = "0.9"

src/api/src/helper.rs

@@ -707,7 +707,6 @@ pub fn pb_values_to_vector_ref(data_type: &ConcreteDataType, values: Values) ->
 }
 
 pub fn pb_values_to_values(data_type: &ConcreteDataType, values: Values) -> Vec<Value> {
-    // TODO(fys): use macros to optimize code
     match data_type {
         ConcreteDataType::Int64(_) => values
             .i64_values

src/api/src/lib.rs

@@ -21,6 +21,7 @@ pub mod prom_store {
     }
 }
 
+pub mod region;
 pub mod v1;
 
 pub use greptime_proto;

src/api/src/region.rs

@@ -0,0 +1,42 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::collections::HashMap;
+
+use common_base::AffectedRows;
+use greptime_proto::v1::region::RegionResponse as RegionResponseV1;
+
+/// This result struct is derived from [RegionResponseV1]
+#[derive(Debug)]
+pub struct RegionResponse {
+    pub affected_rows: AffectedRows,
+    pub extension: HashMap<String, Vec<u8>>,
+}
+
+impl RegionResponse {
+    pub fn from_region_response(region_response: RegionResponseV1) -> Self {
+        Self {
+            affected_rows: region_response.affected_rows as _,
+            extension: region_response.extension,
+        }
+    }
+
+    /// Creates one response without extension
+    pub fn new(affected_rows: AffectedRows) -> Self {
+        Self {
+            affected_rows,
+            extension: Default::default(),
+        }
+    }
+}

src/auth/Cargo.toml

@@ -8,13 +8,17 @@ license.workspace = true
 default = []
 testing = []
 
+[lints]
+workspace = true
+
 [dependencies]
 api.workspace = true
 async-trait.workspace = true
 common-error.workspace = true
 common-macro.workspace = true
+common-telemetry.workspace = true
 digest = "0.10"
-hex = { version = "0.4" }
+notify.workspace = true
 secrecy = { version = "0.8", features = ["serde", "alloc"] }
 sha1 = "0.10"
 snafu.workspace = true

src/auth/src/common.rs

@@ -22,6 +22,9 @@ use snafu::{ensure, OptionExt};
 use crate::error::{IllegalParamSnafu, InvalidConfigSnafu, Result, UserPasswordMismatchSnafu};
 use crate::user_info::DefaultUserInfo;
 use crate::user_provider::static_user_provider::{StaticUserProvider, STATIC_USER_PROVIDER};
+use crate::user_provider::watch_file_user_provider::{
+    WatchFileUserProvider, WATCH_FILE_USER_PROVIDER,
+};
 use crate::{UserInfoRef, UserProviderRef};
 
 pub(crate) const DEFAULT_USERNAME: &str = "greptime";
@@ -40,9 +43,12 @@ pub fn user_provider_from_option(opt: &String) -> Result<UserProviderRef> {
     match name {
         STATIC_USER_PROVIDER => {
             let provider =
-                StaticUserProvider::try_from(content).map(|p| Arc::new(p) as UserProviderRef)?;
+                StaticUserProvider::new(content).map(|p| Arc::new(p) as UserProviderRef)?;
             Ok(provider)
         }
+        WATCH_FILE_USER_PROVIDER => {
+            WatchFileUserProvider::new(content).map(|p| Arc::new(p) as UserProviderRef)
+        }
         _ => InvalidConfigSnafu {
             value: name.to_string(),
             msg: "Invalid UserProviderOption",

src/auth/src/error.rs

@@ -64,6 +64,13 @@ pub enum Error {
         username: String,
     },
 
+    #[snafu(display("Failed to initialize a watcher for file {}", path))]
+    FileWatch {
+        path: String,
+        #[snafu(source)]
+        error: notify::Error,
+    },
+
     #[snafu(display("User is not authorized to perform this action"))]
     PermissionDenied { location: Location },
 }
@@ -73,6 +80,7 @@ impl ErrorExt for Error {
         match self {
             Error::InvalidConfig { .. } => StatusCode::InvalidArguments,
             Error::IllegalParam { .. } => StatusCode::InvalidArguments,
+            Error::FileWatch { .. } => StatusCode::InvalidArguments,
             Error::InternalState { .. } => StatusCode::Unexpected,
             Error::Io { .. } => StatusCode::Internal,
             Error::AuthBackend { .. } => StatusCode::Internal,

src/auth/src/tests.rs

@@ -45,9 +45,9 @@ impl Default for MockUserProvider {
 
 impl MockUserProvider {
     pub fn set_authorization_info(&mut self, info: DatabaseAuthInfo) {
-        self.catalog = info.catalog.to_owned();
+        info.catalog.clone_into(&mut self.catalog);
-        self.schema = info.schema.to_owned();
+        info.schema.clone_into(&mut self.schema);
-        self.username = info.username.to_owned();
+        info.username.clone_into(&mut self.username);
     }
 }
 

src/auth/src/user_provider.rs

@@ -13,10 +13,24 @@
 // limitations under the License.
 
 pub(crate) mod static_user_provider;
+pub(crate) mod watch_file_user_provider;
+
+use std::collections::HashMap;
+use std::fs::File;
+use std::io;
+use std::io::BufRead;
+use std::path::Path;
+
+use secrecy::ExposeSecret;
+use snafu::{ensure, OptionExt, ResultExt};
 
 use crate::common::{Identity, Password};
-use crate::error::Result;
+use crate::error::{
-use crate::UserInfoRef;
+    IllegalParamSnafu, InvalidConfigSnafu, IoSnafu, Result, UnsupportedPasswordTypeSnafu,
+    UserNotFoundSnafu, UserPasswordMismatchSnafu,
+};
+use crate::user_info::DefaultUserInfo;
+use crate::{auth_mysql, UserInfoRef};
 
 #[async_trait::async_trait]
 pub trait UserProvider: Send + Sync {
@@ -44,3 +58,88 @@ pub trait UserProvider: Send + Sync {
         Ok(user_info)
     }
 }
+
+fn load_credential_from_file(filepath: &str) -> Result<Option<HashMap<String, Vec<u8>>>> {
+    // check valid path
+    let path = Path::new(filepath);
+    if !path.exists() {
+        return Ok(None);
+    }
+
+    ensure!(
+        path.is_file(),
+        InvalidConfigSnafu {
+            value: filepath,
+            msg: "UserProvider file must be a file",
+        }
+    );
+    let file = File::open(path).context(IoSnafu)?;
+    let credential = io::BufReader::new(file)
+        .lines()
+        .map_while(std::result::Result::ok)
+        .filter_map(|line| {
+            if let Some((k, v)) = line.split_once('=') {
+                Some((k.to_string(), v.as_bytes().to_vec()))
+            } else {
+                None
+            }
+        })
+        .collect::<HashMap<String, Vec<u8>>>();
+
+    ensure!(
+        !credential.is_empty(),
+        InvalidConfigSnafu {
+            value: filepath,
+            msg: "UserProvider's file must contains at least one valid credential",
+        }
+    );
+
+    Ok(Some(credential))
+}
+
+fn authenticate_with_credential(
+    users: &HashMap<String, Vec<u8>>,
+    input_id: Identity<'_>,
+    input_pwd: Password<'_>,
+) -> Result<UserInfoRef> {
+    match input_id {
+        Identity::UserId(username, _) => {
+            ensure!(
+                !username.is_empty(),
+                IllegalParamSnafu {
+                    msg: "blank username"
+                }
+            );
+            let save_pwd = users.get(username).context(UserNotFoundSnafu {
+                username: username.to_string(),
+            })?;
+
+            match input_pwd {
+                Password::PlainText(pwd) => {
+                    ensure!(
+                        !pwd.expose_secret().is_empty(),
+                        IllegalParamSnafu {
+                            msg: "blank password"
+                        }
+                    );
+                    if save_pwd == pwd.expose_secret().as_bytes() {
+                        Ok(DefaultUserInfo::with_name(username))
+                    } else {
+                        UserPasswordMismatchSnafu {
+                            username: username.to_string(),
+                        }
+                        .fail()
+                    }
+                }
+                Password::MysqlNativePassword(auth_data, salt) => {
+                    auth_mysql(auth_data, salt, username, save_pwd)
+                        .map(|_| DefaultUserInfo::with_name(username))
+                }
+                Password::PgMD5(_, _) => UnsupportedPasswordTypeSnafu {
+                    password_type: "pg_md5",
+                }
+                .fail(),
+            }
+        }
+    }
+}

src/auth/src/user_provider/static_user_provider.rs

@@ -13,60 +13,34 @@
 // limitations under the License.
 
 use std::collections::HashMap;
-use std::fs::File;
-use std::io;
-use std::io::BufRead;
-use std::path::Path;
 
 use async_trait::async_trait;
-use secrecy::ExposeSecret;
+use snafu::OptionExt;
-use snafu::{ensure, OptionExt, ResultExt};
 
-use crate::error::{
+use crate::error::{InvalidConfigSnafu, Result};
-    Error, IllegalParamSnafu, InvalidConfigSnafu, IoSnafu, Result, UnsupportedPasswordTypeSnafu,
+use crate::user_provider::{authenticate_with_credential, load_credential_from_file};
-    UserNotFoundSnafu, UserPasswordMismatchSnafu,
+use crate::{Identity, Password, UserInfoRef, UserProvider};
-};
-use crate::user_info::DefaultUserInfo;
-use crate::{auth_mysql, Identity, Password, UserInfoRef, UserProvider};
 
 pub(crate) const STATIC_USER_PROVIDER: &str = "static_user_provider";
 
-impl TryFrom<&str> for StaticUserProvider {
+pub(crate) struct StaticUserProvider {
-    type Error = Error;
+    users: HashMap<String, Vec<u8>>,
+}
 
-    fn try_from(value: &str) -> Result<Self> {
+impl StaticUserProvider {
+    pub(crate) fn new(value: &str) -> Result<Self> {
         let (mode, content) = value.split_once(':').context(InvalidConfigSnafu {
             value: value.to_string(),
             msg: "StaticUserProviderOption must be in format `<option>:<value>`",
         })?;
         return match mode {
             "file" => {
-                // check valid path
+                let users = load_credential_from_file(content)?
-                let path = Path::new(content);
+                    .context(InvalidConfigSnafu {
-                ensure!(path.exists() && path.is_file(), InvalidConfigSnafu {
                         value: content.to_string(),
-                    msg: "StaticUserProviderOption file must be a valid file path",
+                        msg: "StaticFileUserProvider must be a valid file path",
-                });
+                    })?;
-
+                Ok(StaticUserProvider { users })
-                let file = File::open(path).context(IoSnafu)?;
-                let credential = io::BufReader::new(file)
-                    .lines()
-                    .map_while(std::result::Result::ok)
-                    .filter_map(|line| {
-                        if let Some((k, v)) = line.split_once('=') {
-                            Some((k.to_string(), v.as_bytes().to_vec()))
-                        } else {
-                            None
-                        }
-                    })
-                    .collect::<HashMap<String, Vec<u8>>>();
-
-                ensure!(!credential.is_empty(), InvalidConfigSnafu {
-                    value: content.to_string(),
-                    msg: "StaticUserProviderOption file must contains at least one valid credential",
-                });
-
-                Ok(StaticUserProvider { users: credential, })
             }
             "cmd" => content
                 .split(',')
@@ -88,61 +62,14 @@ impl TryFrom<&str> for StaticUserProvider {
     }
 }
 
-pub(crate) struct StaticUserProvider {
-    users: HashMap<String, Vec<u8>>,
-}
-
 #[async_trait]
 impl UserProvider for StaticUserProvider {
     fn name(&self) -> &str {
         STATIC_USER_PROVIDER
     }
 
-    async fn authenticate(
+    async fn authenticate(&self, id: Identity<'_>, pwd: Password<'_>) -> Result<UserInfoRef> {
-        &self,
+        authenticate_with_credential(&self.users, id, pwd)
-        input_id: Identity<'_>,
-        input_pwd: Password<'_>,
-    ) -> Result<UserInfoRef> {
-        match input_id {
-            Identity::UserId(username, _) => {
-                ensure!(
-                    !username.is_empty(),
-                    IllegalParamSnafu {
-                        msg: "blank username"
-                    }
-                );
-                let save_pwd = self.users.get(username).context(UserNotFoundSnafu {
-                    username: username.to_string(),
-                })?;
-
-                match input_pwd {
-                    Password::PlainText(pwd) => {
-                        ensure!(
-                            !pwd.expose_secret().is_empty(),
-                            IllegalParamSnafu {
-                                msg: "blank password"
-                            }
-                        );
-                        return if save_pwd == pwd.expose_secret().as_bytes() {
-                            Ok(DefaultUserInfo::with_name(username))
-                        } else {
-                            UserPasswordMismatchSnafu {
-                                username: username.to_string(),
-                            }
-                            .fail()
-                        };
-                    }
-                    Password::MysqlNativePassword(auth_data, salt) => {
-                        auth_mysql(auth_data, salt, username, save_pwd)
-                            .map(|_| DefaultUserInfo::with_name(username))
-                    }
-                    Password::PgMD5(_, _) => UnsupportedPasswordTypeSnafu {
-                        password_type: "pg_md5",
-                    }
-                    .fail(),
-                }
-            }
-        }
     }
 
     async fn authorize(
@@ -181,7 +108,7 @@ pub mod test {
     #[tokio::test]
     async fn test_authorize() {
         let user_info = DefaultUserInfo::with_name("root");
-        let provider = StaticUserProvider::try_from("cmd:root=123456,admin=654321").unwrap();
+        let provider = StaticUserProvider::new("cmd:root=123456,admin=654321").unwrap();
         provider
             .authorize("catalog", "schema", &user_info)
             .await
@@ -190,7 +117,7 @@ pub mod test {
 
     #[tokio::test]
     async fn test_inline_provider() {
-        let provider = StaticUserProvider::try_from("cmd:root=123456,admin=654321").unwrap();
+        let provider = StaticUserProvider::new("cmd:root=123456,admin=654321").unwrap();
         test_authenticate(&provider, "root", "123456").await;
         test_authenticate(&provider, "admin", "654321").await;
     }
@@ -214,7 +141,7 @@ admin=654321",
         }
 
         let param = format!("file:{file_path}");
-        let provider = StaticUserProvider::try_from(param.as_str()).unwrap();
+        let provider = StaticUserProvider::new(param.as_str()).unwrap();
         test_authenticate(&provider, "root", "123456").await;
         test_authenticate(&provider, "admin", "654321").await;
     }

src/auth/src/user_provider/watch_file_user_provider.rs

@@ -0,0 +1,215 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::collections::HashMap;
+use std::path::Path;
+use std::sync::mpsc::channel;
+use std::sync::{Arc, Mutex};
+
+use async_trait::async_trait;
+use common_telemetry::{info, warn};
+use notify::{EventKind, RecursiveMode, Watcher};
+use snafu::{ensure, ResultExt};
+
+use crate::error::{FileWatchSnafu, InvalidConfigSnafu, Result};
+use crate::user_info::DefaultUserInfo;
+use crate::user_provider::{authenticate_with_credential, load_credential_from_file};
+use crate::{Identity, Password, UserInfoRef, UserProvider};
+
+pub(crate) const WATCH_FILE_USER_PROVIDER: &str = "watch_file_user_provider";
+
+type WatchedCredentialRef = Arc<Mutex<Option<HashMap<String, Vec<u8>>>>>;
+
+/// A user provider that reads user credential from a file and watches the file for changes.
+///
+/// Empty file is invalid; but file not exist means every user can be authenticated.
+pub(crate) struct WatchFileUserProvider {
+    users: WatchedCredentialRef,
+}
+
+impl WatchFileUserProvider {
+    pub fn new(filepath: &str) -> Result<Self> {
+        let credential = load_credential_from_file(filepath)?;
+        let users = Arc::new(Mutex::new(credential));
+        let this = WatchFileUserProvider {
+            users: users.clone(),
+        };
+
+        let (tx, rx) = channel::<notify::Result<notify::Event>>();
+        let mut debouncer =
+            notify::recommended_watcher(tx).context(FileWatchSnafu { path: "<none>" })?;
+        let mut dir = Path::new(filepath).to_path_buf();
+        ensure!(
+            dir.pop(),
+            InvalidConfigSnafu {
+                value: filepath,
+                msg: "UserProvider path must be a file path",
+            }
+        );
+        debouncer
+            .watch(&dir, RecursiveMode::NonRecursive)
+            .context(FileWatchSnafu { path: filepath })?;
+
+        let filepath = filepath.to_string();
+        std::thread::spawn(move || {
+            let filename = Path::new(&filepath).file_name();
+            let _hold = debouncer;
+            while let Ok(res) = rx.recv() {
+                if let Ok(event) = res {
+                    let is_this_file = event.paths.iter().any(|p| p.file_name() == filename);
+                    let is_relevant_event = matches!(
+                        event.kind,
+                        EventKind::Modify(_) | EventKind::Create(_) | EventKind::Remove(_)
+                    );
+                    if is_this_file && is_relevant_event {
+                        info!(?event.kind, "User provider file {} changed", &filepath);
+                        match load_credential_from_file(&filepath) {
+                            Ok(credential) => {
+                                let mut users =
+                                    users.lock().expect("users credential must be valid");
+                                #[cfg(not(test))]
+                                info!("User provider file {filepath} reloaded");
+                                #[cfg(test)]
+                                info!("User provider file {filepath} reloaded: {credential:?}");
+                                *users = credential;
+                            }
+                            Err(err) => {
+                                warn!(
+                                    ?err,
+                                    "Fail to load credential from file {filepath}; keep the old one",
+                                )
+                            }
+                        }
+                    }
+                }
+            }
+        });
+
+        Ok(this)
+    }
+}
+
+#[async_trait]
+impl UserProvider for WatchFileUserProvider {
+    fn name(&self) -> &str {
+        WATCH_FILE_USER_PROVIDER
+    }
+
+    async fn authenticate(&self, id: Identity<'_>, password: Password<'_>) -> Result<UserInfoRef> {
+        let users = self.users.lock().expect("users credential must be valid");
+        if let Some(users) = users.as_ref() {
+            authenticate_with_credential(users, id, password)
+        } else {
+            match id {
+                Identity::UserId(id, _) => {
+                    warn!(id, "User provider file not exist, allow all users");
+                    Ok(DefaultUserInfo::with_name(id))
+                }
+            }
+        }
+    }
+
+    async fn authorize(&self, _: &str, _: &str, _: &UserInfoRef) -> Result<()> {
+        // default allow all
+        Ok(())
+    }
+}
+
+#[cfg(test)]
+pub mod test {
+    use std::time::{Duration, Instant};
+
+    use common_test_util::temp_dir::create_temp_dir;
+    use tokio::time::sleep;
+
+    use crate::user_provider::watch_file_user_provider::WatchFileUserProvider;
+    use crate::user_provider::{Identity, Password};
+    use crate::UserProvider;
+
+    async fn test_authenticate(
+        provider: &dyn UserProvider,
+        username: &str,
+        password: &str,
+        ok: bool,
+        timeout: Option<Duration>,
+    ) {
+        if let Some(timeout) = timeout {
+            let deadline = Instant::now().checked_add(timeout).unwrap();
+            loop {
+                let re = provider
+                    .authenticate(
+                        Identity::UserId(username, None),
+                        Password::PlainText(password.to_string().into()),
+                    )
+                    .await;
+                if re.is_ok() == ok {
+                    break;
+                } else if Instant::now() < deadline {
+                    sleep(Duration::from_millis(100)).await;
+                } else {
+                    panic!("timeout (username: {username}, password: {password}, expected: {ok})");
+                }
+            }
+        } else {
+            let re = provider
+                .authenticate(
+                    Identity::UserId(username, None),
+                    Password::PlainText(password.to_string().into()),
+                )
+                .await;
+            assert_eq!(
+                re.is_ok(),
+                ok,
+                "username: {}, password: {}",
+                username,
+                password
+            );
+        }
+    }
+
+    #[tokio::test]
+    async fn test_file_provider() {
+        common_telemetry::init_default_ut_logging();
+
+        let dir = create_temp_dir("test_file_provider");
+        let file_path = format!("{}/test_file_provider", dir.path().to_str().unwrap());
+
+        // write a tmp file
+        assert!(std::fs::write(&file_path, "root=123456\nadmin=654321\n").is_ok());
+        let provider = WatchFileUserProvider::new(file_path.as_str()).unwrap();
+        let timeout = Duration::from_secs(60);
+
+        test_authenticate(&provider, "root", "123456", true, None).await;
+        test_authenticate(&provider, "admin", "654321", true, None).await;
+        test_authenticate(&provider, "root", "654321", false, None).await;
+
+        // update the tmp file
+        assert!(std::fs::write(&file_path, "root=654321\n").is_ok());
+        test_authenticate(&provider, "root", "123456", false, Some(timeout)).await;
+        test_authenticate(&provider, "root", "654321", true, Some(timeout)).await;
+        test_authenticate(&provider, "admin", "654321", false, Some(timeout)).await;
+
+        // remove the tmp file
+        assert!(std::fs::remove_file(&file_path).is_ok());
+        test_authenticate(&provider, "root", "123456", true, Some(timeout)).await;
+        test_authenticate(&provider, "root", "654321", true, Some(timeout)).await;
+        test_authenticate(&provider, "admin", "654321", true, Some(timeout)).await;
+
+        // recreate the tmp file
+        assert!(std::fs::write(&file_path, "root=123456\n").is_ok());
+        test_authenticate(&provider, "root", "123456", true, Some(timeout)).await;
+        test_authenticate(&provider, "root", "654321", false, Some(timeout)).await;
+        test_authenticate(&provider, "admin", "654321", false, Some(timeout)).await;
+    }
+}

src/catalog/Cargo.toml

@@ -7,24 +7,24 @@ license.workspace = true
 [features]
 testing = []
 
+[lints]
+workspace = true
+
 [dependencies]
 api.workspace = true
-arc-swap = "1.0"
-arrow-schema.workspace = true
 arrow.workspace = true
+arrow-schema.workspace = true
 async-stream.workspace = true
 async-trait = "0.1"
-build-data = "0.1"
 common-catalog.workspace = true
 common-error.workspace = true
-common-grpc.workspace = true
 common-macro.workspace = true
 common-meta.workspace = true
 common-query.workspace = true
 common-recordbatch.workspace = true
-common-runtime.workspace = true
 common-telemetry.workspace = true
 common-time.workspace = true
+common-version.workspace = true
 dashmap.workspace = true
 datafusion.workspace = true
 datatypes.workspace = true
@@ -33,16 +33,14 @@ futures-util.workspace = true
 itertools.workspace = true
 lazy_static.workspace = true
 meta-client.workspace = true
-moka = { workspace = true, features = ["future"] }
+moka = { workspace = true, features = ["future", "sync"] }
-parking_lot = "0.12"
 partition.workspace = true
 paste = "1.0"
 prometheus.workspace = true
-regex.workspace = true
-serde.workspace = true
 serde_json.workspace = true
 session.workspace = true
 snafu.workspace = true
+sql.workspace = true
 store-api.workspace = true
 table.workspace = true
 tokio.workspace = true

src/catalog/src/error.rs

@@ -41,6 +41,14 @@ pub enum Error {
         source: BoxedError,
     },
 
+    #[snafu(display("Failed to list {}.{}'s tables", catalog, schema))]
+    ListTables {
+        location: Location,
+        catalog: String,
+        schema: String,
+        source: BoxedError,
+    },
+
     #[snafu(display("Failed to re-compile script due to internal error"))]
     CompileScriptInternal {
         location: Location,
@@ -156,6 +164,12 @@ pub enum Error {
         location: Location,
     },
 
+    #[snafu(display("Failed to find table partitions"))]
+    FindPartitions { source: partition::error::Error },
+
+    #[snafu(display("Failed to find region routes"))]
+    FindRegionRoutes { source: partition::error::Error },
+
     #[snafu(display("Failed to read system catalog table records"))]
     ReadSystemCatalog {
         location: Location,
@@ -202,7 +216,7 @@ pub enum Error {
     },
 
     #[snafu(display("Failed to perform metasrv operation"))]
-    MetaSrv {
+    Metasrv {
         location: Location,
         source: meta_client::error::Error,
     },
@@ -237,6 +251,12 @@ pub enum Error {
         source: common_meta::error::Error,
         location: Location,
     },
+
+    #[snafu(display("Get null from table cache, key: {}", key))]
+    TableCacheNotGet { key: String, location: Location },
+
+    #[snafu(display("Failed to get table cache, err: {}", err_msg))]
+    GetTableCache { err_msg: String },
 }
 
 pub type Result<T> = std::result::Result<T, Error>;
@@ -246,11 +266,14 @@ impl ErrorExt for Error {
         match self {
             Error::InvalidKey { .. }
             | Error::SchemaNotFound { .. }
-            | Error::TableNotFound { .. }
             | Error::CatalogNotFound { .. }
+            | Error::FindPartitions { .. }
+            | Error::FindRegionRoutes { .. }
             | Error::InvalidEntryType { .. }
             | Error::ParallelOpenTable { .. } => StatusCode::Unexpected,
 
+            Error::TableNotFound { .. } => StatusCode::TableNotFound,
+
             Error::SystemCatalog { .. }
             | Error::EmptyValue { .. }
             | Error::ValueDeserialize { .. } => StatusCode::StorageUnavailable,
@@ -270,9 +293,9 @@ impl ErrorExt for Error {
                 StatusCode::InvalidArguments
             }
 
-            Error::ListCatalogs { source, .. } | Error::ListSchemas { source, .. } => {
+            Error::ListCatalogs { source, .. }
-                source.status_code()
+            | Error::ListSchemas { source, .. }
-            }
+            | Error::ListTables { source, .. } => source.status_code(),
 
             Error::OpenSystemCatalog { source, .. }
             | Error::CreateSystemCatalog { source, .. }
@@ -281,7 +304,7 @@ impl ErrorExt for Error {
             | Error::CreateTable { source, .. }
             | Error::TableSchemaMismatch { source, .. } => source.status_code(),
 
-            Error::MetaSrv { source, .. } => source.status_code(),
+            Error::Metasrv { source, .. } => source.status_code(),
             Error::SystemCatalogTableScan { source, .. } => source.status_code(),
             Error::SystemCatalogTableScanExec { source, .. } => source.status_code(),
             Error::InvalidTableInfoInCatalog { source, .. } => source.status_code(),
@@ -294,6 +317,7 @@ impl ErrorExt for Error {
             Error::QueryAccessDenied { .. } => StatusCode::AccessDenied,
             Error::Datafusion { .. } => StatusCode::EngineExecuteQuery,
             Error::TableMetadataManager { source, .. } => source.status_code(),
+            Error::TableCacheNotGet { .. } | Error::GetTableCache { .. } => StatusCode::Internal,
         }
     }
 
@@ -333,7 +357,7 @@ mod tests {
         assert_eq!(
             StatusCode::StorageUnavailable,
             Error::SystemCatalog {
-                msg: "".to_string(),
+                msg: String::default(),
                 location: Location::generate(),
             }
             .status_code()

src/catalog/src/information_schema.rs

@@ -12,14 +12,17 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-mod columns;
+pub mod columns;
-mod key_column_usage;
+pub mod key_column_usage;
 mod memory_table;
+mod partitions;
 mod predicate;
+mod region_peers;
 mod runtime_metrics;
-mod schemata;
+pub mod schemata;
+mod table_constraints;
 mod table_names;
-mod tables;
+pub mod tables;
 
 use std::collections::HashMap;
 use std::sync::{Arc, Weak};
@@ -39,16 +42,18 @@ use table::error::{SchemaConversionSnafu, TablesRecordBatchSnafu};
 use table::metadata::{
     FilterPushDownType, TableInfoBuilder, TableInfoRef, TableMetaBuilder, TableType,
 };
-use table::thin_table::{ThinTable, ThinTableAdapter};
+use table::{Table, TableRef};
-use table::TableRef;
 pub use table_names::*;
 
 use self::columns::InformationSchemaColumns;
 use crate::error::Result;
 use crate::information_schema::key_column_usage::InformationSchemaKeyColumnUsage;
 use crate::information_schema::memory_table::{get_schema_columns, MemoryTable};
+use crate::information_schema::partitions::InformationSchemaPartitions;
+use crate::information_schema::region_peers::InformationSchemaRegionPeers;
 use crate::information_schema::runtime_metrics::InformationSchemaMetrics;
 use crate::information_schema::schemata::InformationSchemaSchemata;
+use crate::information_schema::table_constraints::InformationSchemaTableConstraints;
 use crate::information_schema::tables::InformationSchemaTables;
 use crate::CatalogManager;
 
@@ -74,6 +79,7 @@ lazy_static! {
         TRIGGERS,
         GLOBAL_STATUS,
         SESSION_STATUS,
+        PARTITIONS,
     ];
 }
 
@@ -156,6 +162,10 @@ impl InformationSchemaProvider {
                 BUILD_INFO.to_string(),
                 self.build_table(BUILD_INFO).unwrap(),
             );
+            tables.insert(
+                REGION_PEERS.to_string(),
+                self.build_table(REGION_PEERS).unwrap(),
+            );
         }
 
         tables.insert(TABLES.to_string(), self.build_table(TABLES).unwrap());
@@ -165,6 +175,10 @@ impl InformationSchemaProvider {
             KEY_COLUMN_USAGE.to_string(),
             self.build_table(KEY_COLUMN_USAGE).unwrap(),
         );
+        tables.insert(
+            TABLE_CONSTRAINTS.to_string(),
+            self.build_table(TABLE_CONSTRAINTS).unwrap(),
+        );
 
         // Add memory tables
         for name in MEMORY_TABLES.iter() {
@@ -178,10 +192,9 @@ impl InformationSchemaProvider {
         self.information_table(name).map(|table| {
             let table_info = Self::table_info(self.catalog_name.clone(), &table);
             let filter_pushdown = FilterPushDownType::Inexact;
-            let thin_table = ThinTable::new(table_info, filter_pushdown);
-
             let data_source = Arc::new(InformationTableDataSource::new(table));
-            Arc::new(ThinTableAdapter::new(thin_table, data_source)) as _
+            let table = Table::new(table_info, filter_pushdown, data_source);
+            Arc::new(table)
         })
     }
 
@@ -226,6 +239,18 @@ impl InformationSchemaProvider {
                 self.catalog_manager.clone(),
             )) as _),
             RUNTIME_METRICS => Some(Arc::new(InformationSchemaMetrics::new())),
+            PARTITIONS => Some(Arc::new(InformationSchemaPartitions::new(
+                self.catalog_name.clone(),
+                self.catalog_manager.clone(),
+            )) as _),
+            REGION_PEERS => Some(Arc::new(InformationSchemaRegionPeers::new(
+                self.catalog_name.clone(),
+                self.catalog_manager.clone(),
+            )) as _),
+            TABLE_CONSTRAINTS => Some(Arc::new(InformationSchemaTableConstraints::new(
+                self.catalog_name.clone(),
+                self.catalog_manager.clone(),
+            )) as _),
             _ => None,
         }
     }
@@ -312,6 +337,7 @@ impl DataSource for InformationTableDataSource {
             schema: projected_schema,
             stream: Box::pin(stream),
             output_ordering: None,
+            metrics: Default::default(),
         };
 
         Ok(Box::pin(stream))

src/catalog/src/information_schema/columns.rs

@@ -26,12 +26,16 @@ use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
 use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatchStreamAdapter;
 use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
 use datafusion::physical_plan::SendableRecordBatchStream as DfSendableRecordBatchStream;
-use datatypes::prelude::{ConcreteDataType, DataType};
+use datatypes::prelude::{ConcreteDataType, DataType, MutableVector};
 use datatypes::scalars::ScalarVectorBuilder;
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::value::Value;
-use datatypes::vectors::{StringVectorBuilder, VectorRef};
+use datatypes::vectors::{
+    ConstantVector, Int64Vector, Int64VectorBuilder, StringVector, StringVectorBuilder, VectorRef,
+};
+use futures::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
+use sql::statements;
 use store_api::storage::{ScanRequest, TableId};
 
 use super::{InformationTable, COLUMNS};
@@ -47,16 +51,41 @@ pub(super) struct InformationSchemaColumns {
     catalog_manager: Weak<dyn CatalogManager>,
 }
 
-const TABLE_CATALOG: &str = "table_catalog";
+pub const TABLE_CATALOG: &str = "table_catalog";
-const TABLE_SCHEMA: &str = "table_schema";
+pub const TABLE_SCHEMA: &str = "table_schema";
-const TABLE_NAME: &str = "table_name";
+pub const TABLE_NAME: &str = "table_name";
-const COLUMN_NAME: &str = "column_name";
+pub const COLUMN_NAME: &str = "column_name";
-const DATA_TYPE: &str = "data_type";
+const ORDINAL_POSITION: &str = "ordinal_position";
-const SEMANTIC_TYPE: &str = "semantic_type";
+const CHARACTER_MAXIMUM_LENGTH: &str = "character_maximum_length";
-const COLUMN_DEFAULT: &str = "column_default";
+const CHARACTER_OCTET_LENGTH: &str = "character_octet_length";
-const IS_NULLABLE: &str = "is_nullable";
+const NUMERIC_PRECISION: &str = "numeric_precision";
+const NUMERIC_SCALE: &str = "numeric_scale";
+const DATETIME_PRECISION: &str = "datetime_precision";
+const CHARACTER_SET_NAME: &str = "character_set_name";
+pub const COLLATION_NAME: &str = "collation_name";
+pub const COLUMN_KEY: &str = "column_key";
+pub const EXTRA: &str = "extra";
+pub const PRIVILEGES: &str = "privileges";
+const GENERATION_EXPRESSION: &str = "generation_expression";
+// Extension field to keep greptime data type name
+pub const GREPTIME_DATA_TYPE: &str = "greptime_data_type";
+pub const DATA_TYPE: &str = "data_type";
+pub const SEMANTIC_TYPE: &str = "semantic_type";
+pub const COLUMN_DEFAULT: &str = "column_default";
+pub const IS_NULLABLE: &str = "is_nullable";
 const COLUMN_TYPE: &str = "column_type";
-const COLUMN_COMMENT: &str = "column_comment";
+pub const COLUMN_COMMENT: &str = "column_comment";
+const SRS_ID: &str = "srs_id";
+const INIT_CAPACITY: usize = 42;
+
+// The maximum length of string type
+const MAX_STRING_LENGTH: i64 = 2147483647;
+const UTF8_CHARSET_NAME: &str = "utf8";
+const UTF8_COLLATE_NAME: &str = "utf8_bin";
+const PRI_COLUMN_KEY: &str = "PRI";
+const TIME_INDEX_COLUMN_KEY: &str = "TIME INDEX";
+const DEFAULT_PRIVILEGES: &str = "select,insert";
+const EMPTY_STR: &str = "";
 
 impl InformationSchemaColumns {
     pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
@@ -73,12 +102,46 @@ impl InformationSchemaColumns {
             ColumnSchema::new(TABLE_SCHEMA, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(TABLE_NAME, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(COLUMN_NAME, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(ORDINAL_POSITION, ConcreteDataType::int64_datatype(), false),
+            ColumnSchema::new(
+                CHARACTER_MAXIMUM_LENGTH,
+                ConcreteDataType::int64_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                CHARACTER_OCTET_LENGTH,
+                ConcreteDataType::int64_datatype(),
+                true,
+            ),
+            ColumnSchema::new(NUMERIC_PRECISION, ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new(NUMERIC_SCALE, ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new(DATETIME_PRECISION, ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new(
+                CHARACTER_SET_NAME,
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new(COLLATION_NAME, ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new(COLUMN_KEY, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(EXTRA, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(PRIVILEGES, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(
+                GENERATION_EXPRESSION,
+                ConcreteDataType::string_datatype(),
+                false,
+            ),
+            ColumnSchema::new(
+                GREPTIME_DATA_TYPE,
+                ConcreteDataType::string_datatype(),
+                false,
+            ),
             ColumnSchema::new(DATA_TYPE, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(SEMANTIC_TYPE, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(COLUMN_DEFAULT, ConcreteDataType::string_datatype(), true),
             ColumnSchema::new(IS_NULLABLE, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(COLUMN_TYPE, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(COLUMN_COMMENT, ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new(SRS_ID, ConcreteDataType::int64_datatype(), true),
         ]))
     }
 
@@ -134,9 +197,18 @@ struct InformationSchemaColumnsBuilder {
     schema_names: StringVectorBuilder,
     table_names: StringVectorBuilder,
     column_names: StringVectorBuilder,
+    ordinal_positions: Int64VectorBuilder,
+    character_maximum_lengths: Int64VectorBuilder,
+    character_octet_lengths: Int64VectorBuilder,
+    numeric_precisions: Int64VectorBuilder,
+    numeric_scales: Int64VectorBuilder,
+    datetime_precisions: Int64VectorBuilder,
+    character_set_names: StringVectorBuilder,
+    collation_names: StringVectorBuilder,
+    column_keys: StringVectorBuilder,
+    greptime_data_types: StringVectorBuilder,
     data_types: StringVectorBuilder,
     semantic_types: StringVectorBuilder,
-
     column_defaults: StringVectorBuilder,
     is_nullables: StringVectorBuilder,
     column_types: StringVectorBuilder,
@@ -153,16 +225,26 @@ impl InformationSchemaColumnsBuilder {
             schema,
             catalog_name,
             catalog_manager,
-            catalog_names: StringVectorBuilder::with_capacity(42),
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            schema_names: StringVectorBuilder::with_capacity(42),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_names: StringVectorBuilder::with_capacity(42),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            column_names: StringVectorBuilder::with_capacity(42),
+            column_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            data_types: StringVectorBuilder::with_capacity(42),
+            ordinal_positions: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
-            semantic_types: StringVectorBuilder::with_capacity(42),
+            character_maximum_lengths: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
-            column_defaults: StringVectorBuilder::with_capacity(42),
+            character_octet_lengths: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
-            is_nullables: StringVectorBuilder::with_capacity(42),
+            numeric_precisions: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
-            column_types: StringVectorBuilder::with_capacity(42),
+            numeric_scales: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
-            column_comments: StringVectorBuilder::with_capacity(42),
+            datetime_precisions: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
+            character_set_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            collation_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            column_keys: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            greptime_data_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            data_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            semantic_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            column_defaults: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            is_nullables: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            column_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            column_comments: StringVectorBuilder::with_capacity(INIT_CAPACITY),
         }
     }
 
@@ -176,21 +258,9 @@ impl InformationSchemaColumnsBuilder {
         let predicates = Predicates::from_scan_request(&request);
 
         for schema_name in catalog_manager.schema_names(&catalog_name).await? {
-            if !catalog_manager
+            let mut stream = catalog_manager.tables(&catalog_name, &schema_name).await;
-                .schema_exists(&catalog_name, &schema_name)
-                .await?
-            {
-                continue;
-            }
 
-            for table_name in catalog_manager
+            while let Some(table) = stream.try_next().await? {
-                .table_names(&catalog_name, &schema_name)
-                .await?
-            {
-                if let Some(table) = catalog_manager
-                    .table(&catalog_name, &schema_name, &table_name)
-                    .await?
-                {
                 let keys = &table.table_info().meta.primary_key_indices;
                 let schema = table.schema();
 
@@ -205,32 +275,41 @@ impl InformationSchemaColumnsBuilder {
 
                     self.add_column(
                         &predicates,
+                        idx,
                         &catalog_name,
                         &schema_name,
-                            &table_name,
+                        &table.table_info().name,
                         semantic_type,
                         column,
                     );
                 }
-                } else {
-                    unreachable!();
-                }
             }
         }
 
         self.finish()
     }
 
+    #[allow(clippy::too_many_arguments)]
     fn add_column(
         &mut self,
         predicates: &Predicates,
+        index: usize,
         catalog_name: &str,
         schema_name: &str,
         table_name: &str,
         semantic_type: &str,
         column_schema: &ColumnSchema,
     ) {
-        let data_type = &column_schema.data_type.name();
+        // Use sql data type name
+        let data_type = statements::concrete_data_type_to_sql_data_type(&column_schema.data_type)
+            .map(|dt| dt.to_string().to_lowercase())
+            .unwrap_or_else(|_| column_schema.data_type.name());
+
+        let column_key = match semantic_type {
+            SEMANTIC_TYPE_PRIMARY_KEY => PRI_COLUMN_KEY,
+            SEMANTIC_TYPE_TIME_INDEX => TIME_INDEX_COLUMN_KEY,
+            _ => EMPTY_STR,
+        };
 
         let row = [
             (TABLE_CATALOG, &Value::from(catalog_name)),
@@ -239,6 +318,8 @@ impl InformationSchemaColumnsBuilder {
             (COLUMN_NAME, &Value::from(column_schema.name.as_str())),
             (DATA_TYPE, &Value::from(data_type.as_str())),
             (SEMANTIC_TYPE, &Value::from(semantic_type)),
+            (ORDINAL_POSITION, &Value::from((index + 1) as i64)),
+            (COLUMN_KEY, &Value::from(column_key)),
         ];
 
         if !predicates.eval(&row) {
@@ -249,7 +330,63 @@ impl InformationSchemaColumnsBuilder {
         self.schema_names.push(Some(schema_name));
         self.table_names.push(Some(table_name));
         self.column_names.push(Some(&column_schema.name));
-        self.data_types.push(Some(data_type));
+        // Starts from 1
+        self.ordinal_positions.push(Some((index + 1) as i64));
+
+        if column_schema.data_type.is_string() {
+            self.character_maximum_lengths.push(Some(MAX_STRING_LENGTH));
+            self.character_octet_lengths.push(Some(MAX_STRING_LENGTH));
+            self.numeric_precisions.push(None);
+            self.numeric_scales.push(None);
+            self.datetime_precisions.push(None);
+            self.character_set_names.push(Some(UTF8_CHARSET_NAME));
+            self.collation_names.push(Some(UTF8_COLLATE_NAME));
+        } else if column_schema.data_type.is_numeric() || column_schema.data_type.is_decimal() {
+            self.character_maximum_lengths.push(None);
+            self.character_octet_lengths.push(None);
+
+            self.numeric_precisions.push(
+                column_schema
+                    .data_type
+                    .numeric_precision()
+                    .map(|x| x as i64),
+            );
+            self.numeric_scales
+                .push(column_schema.data_type.numeric_scale().map(|x| x as i64));
+
+            self.datetime_precisions.push(None);
+            self.character_set_names.push(None);
+            self.collation_names.push(None);
+        } else {
+            self.character_maximum_lengths.push(None);
+            self.character_octet_lengths.push(None);
+            self.numeric_precisions.push(None);
+            self.numeric_scales.push(None);
+
+            match &column_schema.data_type {
+                ConcreteDataType::DateTime(datetime_type) => {
+                    self.datetime_precisions
+                        .push(Some(datetime_type.precision() as i64));
+                }
+                ConcreteDataType::Timestamp(ts_type) => {
+                    self.datetime_precisions
+                        .push(Some(ts_type.precision() as i64));
+                }
+                ConcreteDataType::Time(time_type) => {
+                    self.datetime_precisions
+                        .push(Some(time_type.precision() as i64));
+                }
+                _ => self.datetime_precisions.push(None),
+            }
+
+            self.character_set_names.push(None);
+            self.collation_names.push(None);
+        }
+
+        self.column_keys.push(Some(column_key));
+        self.greptime_data_types
+            .push(Some(&column_schema.data_type.name()));
+        self.data_types.push(Some(&data_type));
         self.semantic_types.push(Some(semantic_type));
         self.column_defaults.push(
             column_schema
@@ -262,23 +399,52 @@ impl InformationSchemaColumnsBuilder {
         } else {
             self.is_nullables.push(Some("No"));
         }
-        self.column_types.push(Some(data_type));
+        self.column_types.push(Some(&data_type));
         self.column_comments
             .push(column_schema.column_comment().map(|x| x.as_ref()));
     }
 
     fn finish(&mut self) -> Result<RecordBatch> {
+        let rows_num = self.collation_names.len();
+
+        let privileges = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec![DEFAULT_PRIVILEGES])),
+            rows_num,
+        ));
+        let empty_string = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec![EMPTY_STR])),
+            rows_num,
+        ));
+        let srs_ids = Arc::new(ConstantVector::new(
+            Arc::new(Int64Vector::from(vec![None])),
+            rows_num,
+        ));
+
         let columns: Vec<VectorRef> = vec![
             Arc::new(self.catalog_names.finish()),
             Arc::new(self.schema_names.finish()),
             Arc::new(self.table_names.finish()),
             Arc::new(self.column_names.finish()),
+            Arc::new(self.ordinal_positions.finish()),
+            Arc::new(self.character_maximum_lengths.finish()),
+            Arc::new(self.character_octet_lengths.finish()),
+            Arc::new(self.numeric_precisions.finish()),
+            Arc::new(self.numeric_scales.finish()),
+            Arc::new(self.datetime_precisions.finish()),
+            Arc::new(self.character_set_names.finish()),
+            Arc::new(self.collation_names.finish()),
+            Arc::new(self.column_keys.finish()),
+            empty_string.clone(),
+            privileges,
+            empty_string,
+            Arc::new(self.greptime_data_types.finish()),
             Arc::new(self.data_types.finish()),
             Arc::new(self.semantic_types.finish()),
             Arc::new(self.column_defaults.finish()),
             Arc::new(self.is_nullables.finish()),
             Arc::new(self.column_types.finish()),
             Arc::new(self.column_comments.finish()),
+            srs_ids,
         ];
 
         RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)

src/catalog/src/information_schema/key_column_usage.rs

@@ -23,10 +23,10 @@ use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
 use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatchStreamAdapter;
 use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
 use datafusion::physical_plan::SendableRecordBatchStream as DfSendableRecordBatchStream;
-use datatypes::prelude::{ConcreteDataType, ScalarVectorBuilder, VectorRef};
+use datatypes::prelude::{ConcreteDataType, MutableVector, ScalarVectorBuilder, VectorRef};
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::value::Value;
-use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder};
+use datatypes::vectors::{ConstantVector, StringVector, StringVectorBuilder, UInt32VectorBuilder};
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 
@@ -37,13 +37,22 @@ use crate::error::{
 use crate::information_schema::{InformationTable, Predicates};
 use crate::CatalogManager;
 
-const CONSTRAINT_SCHEMA: &str = "constraint_schema";
+pub const CONSTRAINT_SCHEMA: &str = "constraint_schema";
-const CONSTRAINT_NAME: &str = "constraint_name";
+pub const CONSTRAINT_NAME: &str = "constraint_name";
-const TABLE_CATALOG: &str = "table_catalog";
+// It's always `def` in MySQL
-const TABLE_SCHEMA: &str = "table_schema";
+pub const TABLE_CATALOG: &str = "table_catalog";
-const TABLE_NAME: &str = "table_name";
+// The real catalog name for this key column.
-const COLUMN_NAME: &str = "column_name";
+pub const REAL_TABLE_CATALOG: &str = "real_table_catalog";
-const ORDINAL_POSITION: &str = "ordinal_position";
+pub const TABLE_SCHEMA: &str = "table_schema";
+pub const TABLE_NAME: &str = "table_name";
+pub const COLUMN_NAME: &str = "column_name";
+pub const ORDINAL_POSITION: &str = "ordinal_position";
+const INIT_CAPACITY: usize = 42;
+
+/// Primary key constraint name
+pub(crate) const PRI_CONSTRAINT_NAME: &str = "PRIMARY";
+/// Time index constraint name
+pub(crate) const TIME_INDEX_CONSTRAINT_NAME: &str = "TIME INDEX";
 
 /// The virtual table implementation for `information_schema.KEY_COLUMN_USAGE`.
 pub(super) struct InformationSchemaKeyColumnUsage {
@@ -75,6 +84,11 @@ impl InformationSchemaKeyColumnUsage {
             ),
             ColumnSchema::new(CONSTRAINT_NAME, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(TABLE_CATALOG, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(
+                REAL_TABLE_CATALOG,
+                ConcreteDataType::string_datatype(),
+                false,
+            ),
             ColumnSchema::new(TABLE_SCHEMA, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(TABLE_NAME, ConcreteDataType::string_datatype(), false),
             ColumnSchema::new(COLUMN_NAME, ConcreteDataType::string_datatype(), false),
@@ -157,14 +171,12 @@ struct InformationSchemaKeyColumnUsageBuilder {
     constraint_schema: StringVectorBuilder,
     constraint_name: StringVectorBuilder,
     table_catalog: StringVectorBuilder,
+    real_table_catalog: StringVectorBuilder,
     table_schema: StringVectorBuilder,
     table_name: StringVectorBuilder,
     column_name: StringVectorBuilder,
     ordinal_position: UInt32VectorBuilder,
     position_in_unique_constraint: UInt32VectorBuilder,
-    referenced_table_schema: StringVectorBuilder,
-    referenced_table_name: StringVectorBuilder,
-    referenced_column_name: StringVectorBuilder,
 }
 
 impl InformationSchemaKeyColumnUsageBuilder {
@@ -177,18 +189,16 @@ impl InformationSchemaKeyColumnUsageBuilder {
             schema,
             catalog_name,
             catalog_manager,
-            constraint_catalog: StringVectorBuilder::with_capacity(42),
+            constraint_catalog: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            constraint_schema: StringVectorBuilder::with_capacity(42),
+            constraint_schema: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            constraint_name: StringVectorBuilder::with_capacity(42),
+            constraint_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_catalog: StringVectorBuilder::with_capacity(42),
+            table_catalog: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_schema: StringVectorBuilder::with_capacity(42),
+            real_table_catalog: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_name: StringVectorBuilder::with_capacity(42),
+            table_schema: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            column_name: StringVectorBuilder::with_capacity(42),
+            table_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            ordinal_position: UInt32VectorBuilder::with_capacity(42),
+            column_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            position_in_unique_constraint: UInt32VectorBuilder::with_capacity(42),
+            ordinal_position: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
-            referenced_table_schema: StringVectorBuilder::with_capacity(42),
+            position_in_unique_constraint: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
-            referenced_table_name: StringVectorBuilder::with_capacity(42),
-            referenced_column_name: StringVectorBuilder::with_capacity(42),
         }
     }
 
@@ -227,7 +237,8 @@ impl InformationSchemaKeyColumnUsageBuilder {
                             self.add_key_column_usage(
                                 &predicates,
                                 &schema_name,
-                                "TIME INDEX",
+                                TIME_INDEX_CONSTRAINT_NAME,
+                                &catalog_name,
                                 &schema_name,
                                 &table_name,
                                 &column.name,
@@ -236,6 +247,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
                         }
                         if keys.contains(&idx) {
                             primary_constraints.push((
+                                catalog_name.clone(),
                                 schema_name.clone(),
                                 table_name.clone(),
                                 column.name.clone(),
@@ -249,13 +261,14 @@ impl InformationSchemaKeyColumnUsageBuilder {
             }
         }
 
-        for (i, (schema_name, table_name, column_name)) in
+        for (i, (catalog_name, schema_name, table_name, column_name)) in
             primary_constraints.into_iter().enumerate()
         {
             self.add_key_column_usage(
                 &predicates,
                 &schema_name,
-                "PRIMARY",
+                PRI_CONSTRAINT_NAME,
+                &catalog_name,
                 &schema_name,
                 &table_name,
                 &column_name,
@@ -274,6 +287,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
         predicates: &Predicates,
         constraint_schema: &str,
         constraint_name: &str,
+        table_catalog: &str,
         table_schema: &str,
         table_name: &str,
         column_name: &str,
@@ -282,6 +296,7 @@ impl InformationSchemaKeyColumnUsageBuilder {
         let row = [
             (CONSTRAINT_SCHEMA, &Value::from(constraint_schema)),
             (CONSTRAINT_NAME, &Value::from(constraint_name)),
+            (REAL_TABLE_CATALOG, &Value::from(table_catalog)),
             (TABLE_SCHEMA, &Value::from(table_schema)),
             (TABLE_NAME, &Value::from(table_name)),
             (COLUMN_NAME, &Value::from(column_name)),
@@ -296,30 +311,35 @@ impl InformationSchemaKeyColumnUsageBuilder {
         self.constraint_schema.push(Some(constraint_schema));
         self.constraint_name.push(Some(constraint_name));
         self.table_catalog.push(Some("def"));
+        self.real_table_catalog.push(Some(table_catalog));
         self.table_schema.push(Some(table_schema));
         self.table_name.push(Some(table_name));
         self.column_name.push(Some(column_name));
         self.ordinal_position.push(Some(ordinal_position));
         self.position_in_unique_constraint.push(None);
-        self.referenced_table_schema.push(None);
-        self.referenced_table_name.push(None);
-        self.referenced_column_name.push(None);
     }
 
     fn finish(&mut self) -> Result<RecordBatch> {
+        let rows_num = self.table_catalog.len();
+
+        let null_string_vector = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec![None as Option<&str>])),
+            rows_num,
+        ));
         let columns: Vec<VectorRef> = vec![
             Arc::new(self.constraint_catalog.finish()),
             Arc::new(self.constraint_schema.finish()),
             Arc::new(self.constraint_name.finish()),
             Arc::new(self.table_catalog.finish()),
+            Arc::new(self.real_table_catalog.finish()),
             Arc::new(self.table_schema.finish()),
             Arc::new(self.table_name.finish()),
             Arc::new(self.column_name.finish()),
             Arc::new(self.ordinal_position.finish()),
             Arc::new(self.position_in_unique_constraint.finish()),
-            Arc::new(self.referenced_table_schema.finish()),
+            null_string_vector.clone(),
-            Arc::new(self.referenced_table_name.finish()),
+            null_string_vector.clone(),
-            Arc::new(self.referenced_column_name.finish()),
+            null_string_vector,
         ];
         RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
     }

src/catalog/src/information_schema/memory_table/tables.rs

@@ -14,14 +14,14 @@
 
 use std::sync::Arc;
 
-use common_catalog::consts::MITO_ENGINE;
+use common_catalog::consts::{METRIC_ENGINE, MITO_ENGINE};
 use datatypes::prelude::{ConcreteDataType, VectorRef};
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::vectors::{Int64Vector, StringVector};
 
 use crate::information_schema::table_names::*;
 
-const UNKNOWN: &str = "unknown";
+const NO_VALUE: &str = "NO";
 
 /// Find the schema and columns by the table_name, only valid for memory tables.
 /// Safety: the user MUST ensure the table schema exists, panic otherwise.
@@ -61,18 +61,21 @@ pub fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>) {
                 "SAVEPOINTS",
             ]),
             vec![
-                Arc::new(StringVector::from(vec![MITO_ENGINE])),
+                Arc::new(StringVector::from(vec![MITO_ENGINE, METRIC_ENGINE])),
-                Arc::new(StringVector::from(vec!["DEFAULT"])),
+                Arc::new(StringVector::from(vec!["DEFAULT", "YES"])),
                 Arc::new(StringVector::from(vec![
                     "Storage engine for time-series data",
+                    "Storage engine for observability scenarios, which is adept at handling a large number of small tables, making it particularly suitable for cloud-native monitoring",
                 ])),
-                Arc::new(StringVector::from(vec!["NO"])),
+                Arc::new(StringVector::from(vec![NO_VALUE, NO_VALUE])),
-                Arc::new(StringVector::from(vec!["NO"])),
+                Arc::new(StringVector::from(vec![NO_VALUE, NO_VALUE])),
-                Arc::new(StringVector::from(vec!["NO"])),
+                Arc::new(StringVector::from(vec![NO_VALUE, NO_VALUE])),
             ],
         ),
 
-        BUILD_INFO => (
+        BUILD_INFO => {
+            let build_info = common_version::build_info();
+            (
                 string_columns(&[
                     "GIT_BRANCH",
                     "GIT_COMMIT",
@@ -81,21 +84,16 @@ pub fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>) {
                     "PKG_VERSION",
                 ]),
                 vec![
-                Arc::new(StringVector::from(vec![
+                    Arc::new(StringVector::from(vec![build_info.branch.to_string()])),
-                    build_data::get_git_branch().unwrap_or_else(|_| UNKNOWN.to_string())
+                    Arc::new(StringVector::from(vec![build_info.commit.to_string()])),
-                ])),
+                    Arc::new(StringVector::from(vec![build_info
-                Arc::new(StringVector::from(vec![
+                        .commit_short
-                    build_data::get_git_commit().unwrap_or_else(|_| UNKNOWN.to_string())
+                        .to_string()])),
-                ])),
+                    Arc::new(StringVector::from(vec![build_info.dirty.to_string()])),
-                Arc::new(StringVector::from(vec![
+                    Arc::new(StringVector::from(vec![build_info.version.to_string()])),
-                    build_data::get_git_commit_short().unwrap_or_else(|_| UNKNOWN.to_string())
-                ])),
-                Arc::new(StringVector::from(vec![
-                    build_data::get_git_dirty().map_or(UNKNOWN.to_string(), |v| v.to_string())
-                ])),
-                Arc::new(StringVector::from(vec![option_env!("CARGO_PKG_VERSION")])),
                 ],
-        ),
+            )
+        }
 
         CHARACTER_SETS => (
             vec![

src/catalog/src/information_schema/partitions.rs

@@ -0,0 +1,424 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use core::pin::pin;
+use std::sync::{Arc, Weak};
+
+use arrow_schema::SchemaRef as ArrowSchemaRef;
+use common_catalog::consts::INFORMATION_SCHEMA_PARTITIONS_TABLE_ID;
+use common_error::ext::BoxedError;
+use common_query::physical_plan::TaskContext;
+use common_recordbatch::adapter::RecordBatchStreamAdapter;
+use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
+use common_time::datetime::DateTime;
+use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatchStreamAdapter;
+use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
+use datafusion::physical_plan::SendableRecordBatchStream as DfSendableRecordBatchStream;
+use datatypes::prelude::{ConcreteDataType, ScalarVectorBuilder, VectorRef};
+use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
+use datatypes::value::Value;
+use datatypes::vectors::{
+    ConstantVector, DateTimeVector, DateTimeVectorBuilder, Int64Vector, Int64VectorBuilder,
+    MutableVector, StringVector, StringVectorBuilder, UInt64VectorBuilder,
+};
+use futures::{StreamExt, TryStreamExt};
+use partition::manager::PartitionInfo;
+use partition::partition::PartitionDef;
+use snafu::{OptionExt, ResultExt};
+use store_api::storage::{RegionId, ScanRequest, TableId};
+use table::metadata::{TableInfo, TableType};
+
+use super::PARTITIONS;
+use crate::error::{
+    CreateRecordBatchSnafu, FindPartitionsSnafu, InternalSnafu, Result,
+    UpgradeWeakCatalogManagerRefSnafu,
+};
+use crate::information_schema::{InformationTable, Predicates};
+use crate::kvbackend::KvBackendCatalogManager;
+use crate::CatalogManager;
+
+const TABLE_CATALOG: &str = "table_catalog";
+const TABLE_SCHEMA: &str = "table_schema";
+const TABLE_NAME: &str = "table_name";
+const PARTITION_NAME: &str = "partition_name";
+const PARTITION_EXPRESSION: &str = "partition_expression";
+/// The region id
+const GREPTIME_PARTITION_ID: &str = "greptime_partition_id";
+const INIT_CAPACITY: usize = 42;
+
+/// The `PARTITIONS` table provides information about partitioned tables.
+/// See https://dev.mysql.com/doc/refman/8.0/en/information-schema-partitions-table.html
+/// We provide an extral column `greptime_partition_id` for GreptimeDB region id.
+pub(super) struct InformationSchemaPartitions {
+    schema: SchemaRef,
+    catalog_name: String,
+    catalog_manager: Weak<dyn CatalogManager>,
+}
+
+impl InformationSchemaPartitions {
+    pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
+        Self {
+            schema: Self::schema(),
+            catalog_name,
+            catalog_manager,
+        }
+    }
+
+    pub(crate) fn schema() -> SchemaRef {
+        Arc::new(Schema::new(vec![
+            ColumnSchema::new(TABLE_CATALOG, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(TABLE_SCHEMA, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(TABLE_NAME, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(PARTITION_NAME, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(
+                "subpartition_name",
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                "partition_ordinal_position",
+                ConcreteDataType::int64_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                "subpartition_ordinal_position",
+                ConcreteDataType::int64_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                "partition_method",
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                "subpartition_method",
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                PARTITION_EXPRESSION,
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                "subpartition_expression",
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new(
+                "partition_description",
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new("table_rows", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new("avg_row_length", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new("data_length", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new("max_data_length", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new("index_length", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new("data_free", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new("create_time", ConcreteDataType::datetime_datatype(), true),
+            ColumnSchema::new("update_time", ConcreteDataType::datetime_datatype(), true),
+            ColumnSchema::new("check_time", ConcreteDataType::datetime_datatype(), true),
+            ColumnSchema::new("checksum", ConcreteDataType::int64_datatype(), true),
+            ColumnSchema::new(
+                "partition_comment",
+                ConcreteDataType::string_datatype(),
+                true,
+            ),
+            ColumnSchema::new("nodegroup", ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new("tablespace_name", ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new(
+                GREPTIME_PARTITION_ID,
+                ConcreteDataType::uint64_datatype(),
+                true,
+            ),
+        ]))
+    }
+
+    fn builder(&self) -> InformationSchemaPartitionsBuilder {
+        InformationSchemaPartitionsBuilder::new(
+            self.schema.clone(),
+            self.catalog_name.clone(),
+            self.catalog_manager.clone(),
+        )
+    }
+}
+
+impl InformationTable for InformationSchemaPartitions {
+    fn table_id(&self) -> TableId {
+        INFORMATION_SCHEMA_PARTITIONS_TABLE_ID
+    }
+
+    fn table_name(&self) -> &'static str {
+        PARTITIONS
+    }
+
+    fn schema(&self) -> SchemaRef {
+        self.schema.clone()
+    }
+
+    fn to_stream(&self, request: ScanRequest) -> Result<SendableRecordBatchStream> {
+        let schema = self.schema.arrow_schema().clone();
+        let mut builder = self.builder();
+        let stream = Box::pin(DfRecordBatchStreamAdapter::new(
+            schema,
+            futures::stream::once(async move {
+                builder
+                    .make_partitions(Some(request))
+                    .await
+                    .map(|x| x.into_df_record_batch())
+                    .map_err(Into::into)
+            }),
+        ));
+        Ok(Box::pin(
+            RecordBatchStreamAdapter::try_new(stream)
+                .map_err(BoxedError::new)
+                .context(InternalSnafu)?,
+        ))
+    }
+}
+
+struct InformationSchemaPartitionsBuilder {
+    schema: SchemaRef,
+    catalog_name: String,
+    catalog_manager: Weak<dyn CatalogManager>,
+
+    catalog_names: StringVectorBuilder,
+    schema_names: StringVectorBuilder,
+    table_names: StringVectorBuilder,
+    partition_names: StringVectorBuilder,
+    partition_ordinal_positions: Int64VectorBuilder,
+    partition_expressions: StringVectorBuilder,
+    create_times: DateTimeVectorBuilder,
+    partition_ids: UInt64VectorBuilder,
+}
+
+impl InformationSchemaPartitionsBuilder {
+    fn new(
+        schema: SchemaRef,
+        catalog_name: String,
+        catalog_manager: Weak<dyn CatalogManager>,
+    ) -> Self {
+        Self {
+            schema,
+            catalog_name,
+            catalog_manager,
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            partition_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            partition_ordinal_positions: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
+            partition_expressions: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            create_times: DateTimeVectorBuilder::with_capacity(INIT_CAPACITY),
+            partition_ids: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
+        }
+    }
+
+    /// Construct the `information_schema.partitions` virtual table
+    async fn make_partitions(&mut self, request: Option<ScanRequest>) -> Result<RecordBatch> {
+        let catalog_name = self.catalog_name.clone();
+        let catalog_manager = self
+            .catalog_manager
+            .upgrade()
+            .context(UpgradeWeakCatalogManagerRefSnafu)?;
+
+        let partition_manager = catalog_manager
+            .as_any()
+            .downcast_ref::<KvBackendCatalogManager>()
+            .map(|catalog_manager| catalog_manager.partition_manager());
+
+        let predicates = Predicates::from_scan_request(&request);
+
+        for schema_name in catalog_manager.schema_names(&catalog_name).await? {
+            let table_info_stream = catalog_manager
+                .tables(&catalog_name, &schema_name)
+                .await
+                .try_filter_map(|t| async move {
+                    let table_info = t.table_info();
+                    if table_info.table_type == TableType::Temporary {
+                        Ok(None)
+                    } else {
+                        Ok(Some(table_info))
+                    }
+                });
+
+            const BATCH_SIZE: usize = 128;
+
+            // Split table infos into chunks
+            let mut table_info_chunks = pin!(table_info_stream.ready_chunks(BATCH_SIZE));
+
+            while let Some(table_infos) = table_info_chunks.next().await {
+                let table_infos = table_infos.into_iter().collect::<Result<Vec<_>>>()?;
+                let table_ids: Vec<TableId> =
+                    table_infos.iter().map(|info| info.ident.table_id).collect();
+
+                let mut table_partitions = if let Some(partition_manager) = &partition_manager {
+                    partition_manager
+                        .batch_find_table_partitions(&table_ids)
+                        .await
+                        .context(FindPartitionsSnafu)?
+                } else {
+                    // Current node must be a standalone instance, contains only one partition by default.
+                    // TODO(dennis): change it when we support multi-regions for standalone.
+                    table_ids
+                        .into_iter()
+                        .map(|table_id| {
+                            (
+                                table_id,
+                                vec![PartitionInfo {
+                                    id: RegionId::new(table_id, 0),
+                                    partition: PartitionDef::new(vec![], vec![]),
+                                }],
+                            )
+                        })
+                        .collect()
+                };
+
+                for table_info in table_infos {
+                    let partitions = table_partitions
+                        .remove(&table_info.ident.table_id)
+                        .unwrap_or(vec![]);
+
+                    self.add_partitions(
+                        &predicates,
+                        &table_info,
+                        &catalog_name,
+                        &schema_name,
+                        &table_info.name,
+                        &partitions,
+                    );
+                }
+            }
+        }
+
+        self.finish()
+    }
+
+    #[allow(clippy::too_many_arguments)]
+    fn add_partitions(
+        &mut self,
+        predicates: &Predicates,
+        table_info: &TableInfo,
+        catalog_name: &str,
+        schema_name: &str,
+        table_name: &str,
+        partitions: &[PartitionInfo],
+    ) {
+        let row = [
+            (TABLE_CATALOG, &Value::from(catalog_name)),
+            (TABLE_SCHEMA, &Value::from(schema_name)),
+            (TABLE_NAME, &Value::from(table_name)),
+        ];
+
+        if !predicates.eval(&row) {
+            return;
+        }
+
+        for (index, partition) in partitions.iter().enumerate() {
+            let partition_name = format!("p{index}");
+
+            self.catalog_names.push(Some(catalog_name));
+            self.schema_names.push(Some(schema_name));
+            self.table_names.push(Some(table_name));
+            self.partition_names.push(Some(&partition_name));
+            self.partition_ordinal_positions
+                .push(Some((index + 1) as i64));
+            let expressions = if partition.partition.partition_columns().is_empty() {
+                None
+            } else {
+                Some(partition.partition.to_string())
+            };
+
+            self.partition_expressions.push(expressions.as_deref());
+            self.create_times.push(Some(DateTime::from(
+                table_info.meta.created_on.timestamp_millis(),
+            )));
+            self.partition_ids.push(Some(partition.id.as_u64()));
+        }
+    }
+
+    fn finish(&mut self) -> Result<RecordBatch> {
+        let rows_num = self.catalog_names.len();
+
+        let null_string_vector = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec![None as Option<&str>])),
+            rows_num,
+        ));
+        let null_i64_vector = Arc::new(ConstantVector::new(
+            Arc::new(Int64Vector::from(vec![None])),
+            rows_num,
+        ));
+        let null_datetime_vector = Arc::new(ConstantVector::new(
+            Arc::new(DateTimeVector::from(vec![None])),
+            rows_num,
+        ));
+        let partition_methods = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec![Some("RANGE")])),
+            rows_num,
+        ));
+
+        let columns: Vec<VectorRef> = vec![
+            Arc::new(self.catalog_names.finish()),
+            Arc::new(self.schema_names.finish()),
+            Arc::new(self.table_names.finish()),
+            Arc::new(self.partition_names.finish()),
+            null_string_vector.clone(),
+            Arc::new(self.partition_ordinal_positions.finish()),
+            null_i64_vector.clone(),
+            partition_methods,
+            null_string_vector.clone(),
+            Arc::new(self.partition_expressions.finish()),
+            null_string_vector.clone(),
+            null_string_vector.clone(),
+            // TODO(dennis): rows and index statistics info
+            null_i64_vector.clone(),
+            null_i64_vector.clone(),
+            null_i64_vector.clone(),
+            null_i64_vector.clone(),
+            null_i64_vector.clone(),
+            null_i64_vector.clone(),
+            Arc::new(self.create_times.finish()),
+            // TODO(dennis): supports update_time
+            null_datetime_vector.clone(),
+            null_datetime_vector,
+            null_i64_vector,
+            null_string_vector.clone(),
+            null_string_vector.clone(),
+            null_string_vector,
+            Arc::new(self.partition_ids.finish()),
+        ];
+        RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
+    }
+}
+
+impl DfPartitionStream for InformationSchemaPartitions {
+    fn schema(&self) -> &ArrowSchemaRef {
+        self.schema.arrow_schema()
+    }
+
+    fn execute(&self, _: Arc<TaskContext>) -> DfSendableRecordBatchStream {
+        let schema = self.schema.arrow_schema().clone();
+        let mut builder = self.builder();
+        Box::pin(DfRecordBatchStreamAdapter::new(
+            schema,
+            futures::stream::once(async move {
+                builder
+                    .make_partitions(None)
+                    .await
+                    .map(|x| x.into_df_record_batch())
+                    .map_err(Into::into)
+            }),
+        ))
+    }
+}

src/catalog/src/information_schema/predicate.rs

@@ -109,11 +109,7 @@ impl Predicate {
                 };
             }
             Predicate::Not(p) => {
-                let Some(b) = p.eval(row) else {
+                return Some(!p.eval(row)?);
-                    return None;
-                };
-
-                return Some(!b);
             }
         }
 
@@ -125,13 +121,7 @@ impl Predicate {
     fn from_expr(expr: DfExpr) -> Option<Predicate> {
         match expr {
             // NOT expr
-            DfExpr::Not(expr) => {
+            DfExpr::Not(expr) => Some(Predicate::Not(Box::new(Self::from_expr(*expr)?))),
-                let Some(p) = Self::from_expr(*expr) else {
-                    return None;
-                };
-
-                Some(Predicate::Not(Box::new(p)))
-            }
             // expr LIKE pattern
             DfExpr::Like(Like {
                 negated,
@@ -178,25 +168,15 @@ impl Predicate {
                 }
                 // left AND right
                 (left, Operator::And, right) => {
-                    let Some(left) = Self::from_expr(left) else {
+                    let left = Self::from_expr(left)?;
-                        return None;
+                    let right = Self::from_expr(right)?;
-                    };
-
-                    let Some(right) = Self::from_expr(right) else {
-                        return None;
-                    };
 
                     Some(Predicate::And(Box::new(left), Box::new(right)))
                 }
                 // left OR right
                 (left, Operator::Or, right) => {
-                    let Some(left) = Self::from_expr(left) else {
+                    let left = Self::from_expr(left)?;
-                        return None;
+                    let right = Self::from_expr(right)?;
-                    };
-
-                    let Some(right) = Self::from_expr(right) else {
-                        return None;
-                    };
 
                     Some(Predicate::Or(Box::new(left), Box::new(right)))
                 }

src/catalog/src/information_schema/region_peers.rs

@@ -0,0 +1,279 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use core::pin::pin;
+use std::sync::{Arc, Weak};
+
+use arrow_schema::SchemaRef as ArrowSchemaRef;
+use common_catalog::consts::INFORMATION_SCHEMA_REGION_PEERS_TABLE_ID;
+use common_error::ext::BoxedError;
+use common_meta::rpc::router::RegionRoute;
+use common_query::physical_plan::TaskContext;
+use common_recordbatch::adapter::RecordBatchStreamAdapter;
+use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
+use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatchStreamAdapter;
+use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
+use datafusion::physical_plan::SendableRecordBatchStream as DfSendableRecordBatchStream;
+use datatypes::prelude::{ConcreteDataType, ScalarVectorBuilder, VectorRef};
+use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
+use datatypes::value::Value;
+use datatypes::vectors::{Int64VectorBuilder, StringVectorBuilder, UInt64VectorBuilder};
+use futures::{StreamExt, TryStreamExt};
+use snafu::{OptionExt, ResultExt};
+use store_api::storage::{ScanRequest, TableId};
+use table::metadata::TableType;
+
+use super::REGION_PEERS;
+use crate::error::{
+    CreateRecordBatchSnafu, FindRegionRoutesSnafu, InternalSnafu, Result,
+    UpgradeWeakCatalogManagerRefSnafu,
+};
+use crate::information_schema::{InformationTable, Predicates};
+use crate::kvbackend::KvBackendCatalogManager;
+use crate::CatalogManager;
+
+const REGION_ID: &str = "region_id";
+const PEER_ID: &str = "peer_id";
+const PEER_ADDR: &str = "peer_addr";
+const IS_LEADER: &str = "is_leader";
+const STATUS: &str = "status";
+const DOWN_SECONDS: &str = "down_seconds";
+const INIT_CAPACITY: usize = 42;
+
+/// The `REGION_PEERS` table provides information about the region distribution and routes. Including fields:
+///
+/// - `region_id`: the region id
+/// - `peer_id`: the region storage datanode peer id
+/// - `peer_addr`: the region storage datanode peer address
+/// - `is_leader`: whether the peer is the leader
+/// - `status`: the region status, `ALIVE` or `DOWNGRADED`.
+/// - `down_seconds`: the duration of being offline, in seconds.
+///
+pub(super) struct InformationSchemaRegionPeers {
+    schema: SchemaRef,
+    catalog_name: String,
+    catalog_manager: Weak<dyn CatalogManager>,
+}
+
+impl InformationSchemaRegionPeers {
+    pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
+        Self {
+            schema: Self::schema(),
+            catalog_name,
+            catalog_manager,
+        }
+    }
+
+    pub(crate) fn schema() -> SchemaRef {
+        Arc::new(Schema::new(vec![
+            ColumnSchema::new(REGION_ID, ConcreteDataType::uint64_datatype(), false),
+            ColumnSchema::new(PEER_ID, ConcreteDataType::uint64_datatype(), true),
+            ColumnSchema::new(PEER_ADDR, ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new(IS_LEADER, ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new(STATUS, ConcreteDataType::string_datatype(), true),
+            ColumnSchema::new(DOWN_SECONDS, ConcreteDataType::int64_datatype(), true),
+        ]))
+    }
+
+    fn builder(&self) -> InformationSchemaRegionPeersBuilder {
+        InformationSchemaRegionPeersBuilder::new(
+            self.schema.clone(),
+            self.catalog_name.clone(),
+            self.catalog_manager.clone(),
+        )
+    }
+}
+
+impl InformationTable for InformationSchemaRegionPeers {
+    fn table_id(&self) -> TableId {
+        INFORMATION_SCHEMA_REGION_PEERS_TABLE_ID
+    }
+
+    fn table_name(&self) -> &'static str {
+        REGION_PEERS
+    }
+
+    fn schema(&self) -> SchemaRef {
+        self.schema.clone()
+    }
+
+    fn to_stream(&self, request: ScanRequest) -> Result<SendableRecordBatchStream> {
+        let schema = self.schema.arrow_schema().clone();
+        let mut builder = self.builder();
+        let stream = Box::pin(DfRecordBatchStreamAdapter::new(
+            schema,
+            futures::stream::once(async move {
+                builder
+                    .make_region_peers(Some(request))
+                    .await
+                    .map(|x| x.into_df_record_batch())
+                    .map_err(Into::into)
+            }),
+        ));
+        Ok(Box::pin(
+            RecordBatchStreamAdapter::try_new(stream)
+                .map_err(BoxedError::new)
+                .context(InternalSnafu)?,
+        ))
+    }
+}
+
+struct InformationSchemaRegionPeersBuilder {
+    schema: SchemaRef,
+    catalog_name: String,
+    catalog_manager: Weak<dyn CatalogManager>,
+
+    region_ids: UInt64VectorBuilder,
+    peer_ids: UInt64VectorBuilder,
+    peer_addrs: StringVectorBuilder,
+    is_leaders: StringVectorBuilder,
+    statuses: StringVectorBuilder,
+    down_seconds: Int64VectorBuilder,
+}
+
+impl InformationSchemaRegionPeersBuilder {
+    fn new(
+        schema: SchemaRef,
+        catalog_name: String,
+        catalog_manager: Weak<dyn CatalogManager>,
+    ) -> Self {
+        Self {
+            schema,
+            catalog_name,
+            catalog_manager,
+            region_ids: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
+            peer_ids: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
+            peer_addrs: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            is_leaders: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            statuses: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            down_seconds: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
+        }
+    }
+
+    /// Construct the `information_schema.region_peers` virtual table
+    async fn make_region_peers(&mut self, request: Option<ScanRequest>) -> Result<RecordBatch> {
+        let catalog_name = self.catalog_name.clone();
+        let catalog_manager = self
+            .catalog_manager
+            .upgrade()
+            .context(UpgradeWeakCatalogManagerRefSnafu)?;
+
+        let partition_manager = catalog_manager
+            .as_any()
+            .downcast_ref::<KvBackendCatalogManager>()
+            .map(|catalog_manager| catalog_manager.partition_manager());
+
+        let predicates = Predicates::from_scan_request(&request);
+
+        for schema_name in catalog_manager.schema_names(&catalog_name).await? {
+            let table_id_stream = catalog_manager
+                .tables(&catalog_name, &schema_name)
+                .await
+                .try_filter_map(|t| async move {
+                    let table_info = t.table_info();
+                    if table_info.table_type == TableType::Temporary {
+                        Ok(None)
+                    } else {
+                        Ok(Some(table_info.ident.table_id))
+                    }
+                });
+
+            const BATCH_SIZE: usize = 128;
+
+            // Split table ids into chunks
+            let mut table_id_chunks = pin!(table_id_stream.ready_chunks(BATCH_SIZE));
+
+            while let Some(table_ids) = table_id_chunks.next().await {
+                let table_ids = table_ids.into_iter().collect::<Result<Vec<_>>>()?;
+
+                let table_routes = if let Some(partition_manager) = &partition_manager {
+                    partition_manager
+                        .batch_find_region_routes(&table_ids)
+                        .await
+                        .context(FindRegionRoutesSnafu)?
+                } else {
+                    table_ids.into_iter().map(|id| (id, vec![])).collect()
+                };
+
+                for routes in table_routes.values() {
+                    self.add_region_peers(&predicates, routes);
+                }
+            }
+        }
+
+        self.finish()
+    }
+
+    fn add_region_peers(&mut self, predicates: &Predicates, routes: &[RegionRoute]) {
+        for route in routes {
+            let region_id = route.region.id.as_u64();
+            let peer_id = route.leader_peer.clone().map(|p| p.id);
+            let peer_addr = route.leader_peer.clone().map(|p| p.addr);
+            let status = if let Some(status) = route.leader_status {
+                Some(status.as_ref().to_string())
+            } else {
+                // Alive by default
+                Some("ALIVE".to_string())
+            };
+
+            let row = [(REGION_ID, &Value::from(region_id))];
+
+            if !predicates.eval(&row) {
+                return;
+            }
+
+            // TODO(dennis): adds followers.
+            self.region_ids.push(Some(region_id));
+            self.peer_ids.push(peer_id);
+            self.peer_addrs.push(peer_addr.as_deref());
+            self.is_leaders.push(Some("Yes"));
+            self.statuses.push(status.as_deref());
+            self.down_seconds
+                .push(route.leader_down_millis().map(|m| m / 1000));
+        }
+    }
+
+    fn finish(&mut self) -> Result<RecordBatch> {
+        let columns: Vec<VectorRef> = vec![
+            Arc::new(self.region_ids.finish()),
+            Arc::new(self.peer_ids.finish()),
+            Arc::new(self.peer_addrs.finish()),
+            Arc::new(self.is_leaders.finish()),
+            Arc::new(self.statuses.finish()),
+            Arc::new(self.down_seconds.finish()),
+        ];
+        RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
+    }
+}
+
+impl DfPartitionStream for InformationSchemaRegionPeers {
+    fn schema(&self) -> &ArrowSchemaRef {
+        self.schema.arrow_schema()
+    }
+
+    fn execute(&self, _: Arc<TaskContext>) -> DfSendableRecordBatchStream {
+        let schema = self.schema.arrow_schema().clone();
+        let mut builder = self.builder();
+        Box::pin(DfRecordBatchStreamAdapter::new(
+            schema,
+            futures::stream::once(async move {
+                builder
+                    .make_region_peers(None)
+                    .await
+                    .map(|x| x.into_df_record_batch())
+                    .map_err(Into::into)
+            }),
+        ))
+    }
+}

src/catalog/src/information_schema/schemata.rs

@@ -37,10 +37,11 @@ use crate::error::{
 use crate::information_schema::{InformationTable, Predicates};
 use crate::CatalogManager;
 
-const CATALOG_NAME: &str = "catalog_name";
+pub const CATALOG_NAME: &str = "catalog_name";
-const SCHEMA_NAME: &str = "schema_name";
+pub const SCHEMA_NAME: &str = "schema_name";
 const DEFAULT_CHARACTER_SET_NAME: &str = "default_character_set_name";
 const DEFAULT_COLLATION_NAME: &str = "default_collation_name";
+const INIT_CAPACITY: usize = 42;
 
 /// The `information_schema.schemata` table implementation.
 pub(super) struct InformationSchemaSchemata {
@@ -144,11 +145,11 @@ impl InformationSchemaSchemataBuilder {
             schema,
             catalog_name,
             catalog_manager,
-            catalog_names: StringVectorBuilder::with_capacity(42),
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            schema_names: StringVectorBuilder::with_capacity(42),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            charset_names: StringVectorBuilder::with_capacity(42),
+            charset_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            collation_names: StringVectorBuilder::with_capacity(42),
+            collation_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            sql_paths: StringVectorBuilder::with_capacity(42),
+            sql_paths: StringVectorBuilder::with_capacity(INIT_CAPACITY),
         }
     }
 
@@ -162,13 +163,6 @@ impl InformationSchemaSchemataBuilder {
         let predicates = Predicates::from_scan_request(&request);
 
         for schema_name in catalog_manager.schema_names(&catalog_name).await? {
-            if !catalog_manager
-                .schema_exists(&catalog_name, &schema_name)
-                .await?
-            {
-                continue;
-            }
-
             self.add_schema(&predicates, &catalog_name, &schema_name);
         }
 

src/catalog/src/information_schema/table_constraints.rs

@@ -0,0 +1,286 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::{Arc, Weak};
+
+use arrow_schema::SchemaRef as ArrowSchemaRef;
+use common_catalog::consts::INFORMATION_SCHEMA_TABLE_CONSTRAINTS_TABLE_ID;
+use common_error::ext::BoxedError;
+use common_query::physical_plan::TaskContext;
+use common_recordbatch::adapter::RecordBatchStreamAdapter;
+use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
+use datafusion::physical_plan::stream::RecordBatchStreamAdapter as DfRecordBatchStreamAdapter;
+use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
+use datafusion::physical_plan::SendableRecordBatchStream as DfSendableRecordBatchStream;
+use datatypes::prelude::{ConcreteDataType, MutableVector};
+use datatypes::scalars::ScalarVectorBuilder;
+use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
+use datatypes::value::Value;
+use datatypes::vectors::{ConstantVector, StringVector, StringVectorBuilder, VectorRef};
+use futures::TryStreamExt;
+use snafu::{OptionExt, ResultExt};
+use store_api::storage::{ScanRequest, TableId};
+
+use super::{InformationTable, TABLE_CONSTRAINTS};
+use crate::error::{
+    CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
+};
+use crate::information_schema::key_column_usage::{
+    PRI_CONSTRAINT_NAME, TIME_INDEX_CONSTRAINT_NAME,
+};
+use crate::information_schema::Predicates;
+use crate::CatalogManager;
+
+/// The `TABLE_CONSTRAINTS` table describes which tables have constraints.
+pub(super) struct InformationSchemaTableConstraints {
+    schema: SchemaRef,
+    catalog_name: String,
+    catalog_manager: Weak<dyn CatalogManager>,
+}
+
+const CONSTRAINT_CATALOG: &str = "constraint_catalog";
+const CONSTRAINT_SCHEMA: &str = "constraint_schema";
+const CONSTRAINT_NAME: &str = "constraint_name";
+const TABLE_SCHEMA: &str = "table_schema";
+const TABLE_NAME: &str = "table_name";
+const CONSTRAINT_TYPE: &str = "constraint_type";
+const ENFORCED: &str = "enforced";
+
+const INIT_CAPACITY: usize = 42;
+
+const TIME_INDEX_CONSTRAINT_TYPE: &str = "TIME INDEX";
+const PRI_KEY_CONSTRAINT_TYPE: &str = "PRIMARY KEY";
+
+impl InformationSchemaTableConstraints {
+    pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
+        Self {
+            schema: Self::schema(),
+            catalog_name,
+            catalog_manager,
+        }
+    }
+
+    fn schema() -> SchemaRef {
+        Arc::new(Schema::new(vec![
+            ColumnSchema::new(
+                CONSTRAINT_CATALOG,
+                ConcreteDataType::string_datatype(),
+                false,
+            ),
+            ColumnSchema::new(
+                CONSTRAINT_SCHEMA,
+                ConcreteDataType::string_datatype(),
+                false,
+            ),
+            ColumnSchema::new(CONSTRAINT_NAME, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(TABLE_SCHEMA, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(TABLE_NAME, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(CONSTRAINT_TYPE, ConcreteDataType::string_datatype(), false),
+            ColumnSchema::new(ENFORCED, ConcreteDataType::string_datatype(), false),
+        ]))
+    }
+
+    fn builder(&self) -> InformationSchemaTableConstraintsBuilder {
+        InformationSchemaTableConstraintsBuilder::new(
+            self.schema.clone(),
+            self.catalog_name.clone(),
+            self.catalog_manager.clone(),
+        )
+    }
+}
+
+impl InformationTable for InformationSchemaTableConstraints {
+    fn table_id(&self) -> TableId {
+        INFORMATION_SCHEMA_TABLE_CONSTRAINTS_TABLE_ID
+    }
+
+    fn table_name(&self) -> &'static str {
+        TABLE_CONSTRAINTS
+    }
+
+    fn schema(&self) -> SchemaRef {
+        self.schema.clone()
+    }
+
+    fn to_stream(&self, request: ScanRequest) -> Result<SendableRecordBatchStream> {
+        let schema = self.schema.arrow_schema().clone();
+        let mut builder = self.builder();
+        let stream = Box::pin(DfRecordBatchStreamAdapter::new(
+            schema,
+            futures::stream::once(async move {
+                builder
+                    .make_table_constraints(Some(request))
+                    .await
+                    .map(|x| x.into_df_record_batch())
+                    .map_err(Into::into)
+            }),
+        ));
+        Ok(Box::pin(
+            RecordBatchStreamAdapter::try_new(stream)
+                .map_err(BoxedError::new)
+                .context(InternalSnafu)?,
+        ))
+    }
+}
+
+struct InformationSchemaTableConstraintsBuilder {
+    schema: SchemaRef,
+    catalog_name: String,
+    catalog_manager: Weak<dyn CatalogManager>,
+
+    constraint_schemas: StringVectorBuilder,
+    constraint_names: StringVectorBuilder,
+    table_schemas: StringVectorBuilder,
+    table_names: StringVectorBuilder,
+    constraint_types: StringVectorBuilder,
+}
+
+impl InformationSchemaTableConstraintsBuilder {
+    fn new(
+        schema: SchemaRef,
+        catalog_name: String,
+        catalog_manager: Weak<dyn CatalogManager>,
+    ) -> Self {
+        Self {
+            schema,
+            catalog_name,
+            catalog_manager,
+            constraint_schemas: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            constraint_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_schemas: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            constraint_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+        }
+    }
+
+    /// Construct the `information_schema.table_constraints` virtual table
+    async fn make_table_constraints(
+        &mut self,
+        request: Option<ScanRequest>,
+    ) -> Result<RecordBatch> {
+        let catalog_name = self.catalog_name.clone();
+        let catalog_manager = self
+            .catalog_manager
+            .upgrade()
+            .context(UpgradeWeakCatalogManagerRefSnafu)?;
+        let predicates = Predicates::from_scan_request(&request);
+
+        for schema_name in catalog_manager.schema_names(&catalog_name).await? {
+            let mut stream = catalog_manager.tables(&catalog_name, &schema_name).await;
+
+            while let Some(table) = stream.try_next().await? {
+                let keys = &table.table_info().meta.primary_key_indices;
+                let schema = table.schema();
+
+                if schema.timestamp_index().is_some() {
+                    self.add_table_constraint(
+                        &predicates,
+                        &schema_name,
+                        TIME_INDEX_CONSTRAINT_NAME,
+                        &schema_name,
+                        &table.table_info().name,
+                        TIME_INDEX_CONSTRAINT_TYPE,
+                    );
+                }
+
+                if !keys.is_empty() {
+                    self.add_table_constraint(
+                        &predicates,
+                        &schema_name,
+                        PRI_CONSTRAINT_NAME,
+                        &schema_name,
+                        &table.table_info().name,
+                        PRI_KEY_CONSTRAINT_TYPE,
+                    );
+                }
+            }
+        }
+
+        self.finish()
+    }
+
+    fn add_table_constraint(
+        &mut self,
+        predicates: &Predicates,
+        constraint_schema: &str,
+        constraint_name: &str,
+        table_schema: &str,
+        table_name: &str,
+        constraint_type: &str,
+    ) {
+        let row = [
+            (CONSTRAINT_SCHEMA, &Value::from(constraint_schema)),
+            (CONSTRAINT_NAME, &Value::from(constraint_name)),
+            (TABLE_SCHEMA, &Value::from(table_schema)),
+            (TABLE_NAME, &Value::from(table_name)),
+            (CONSTRAINT_TYPE, &Value::from(constraint_type)),
+        ];
+
+        if !predicates.eval(&row) {
+            return;
+        }
+
+        self.constraint_schemas.push(Some(constraint_schema));
+        self.constraint_names.push(Some(constraint_name));
+        self.table_schemas.push(Some(table_schema));
+        self.table_names.push(Some(table_name));
+        self.constraint_types.push(Some(constraint_type));
+    }
+
+    fn finish(&mut self) -> Result<RecordBatch> {
+        let rows_num = self.constraint_names.len();
+
+        let constraint_catalogs = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec!["def"])),
+            rows_num,
+        ));
+        let enforceds = Arc::new(ConstantVector::new(
+            Arc::new(StringVector::from(vec!["YES"])),
+            rows_num,
+        ));
+
+        let columns: Vec<VectorRef> = vec![
+            constraint_catalogs,
+            Arc::new(self.constraint_schemas.finish()),
+            Arc::new(self.constraint_names.finish()),
+            Arc::new(self.table_schemas.finish()),
+            Arc::new(self.table_names.finish()),
+            Arc::new(self.constraint_types.finish()),
+            enforceds,
+        ];
+
+        RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
+    }
+}
+
+impl DfPartitionStream for InformationSchemaTableConstraints {
+    fn schema(&self) -> &ArrowSchemaRef {
+        self.schema.arrow_schema()
+    }
+
+    fn execute(&self, _: Arc<TaskContext>) -> DfSendableRecordBatchStream {
+        let schema = self.schema.arrow_schema().clone();
+        let mut builder = self.builder();
+        Box::pin(DfRecordBatchStreamAdapter::new(
+            schema,
+            futures::stream::once(async move {
+                builder
+                    .make_table_constraints(None)
+                    .await
+                    .map(|x| x.into_df_record_batch())
+                    .map_err(Into::into)
+            }),
+        ))
+    }
+}

src/catalog/src/information_schema/table_names.rs

@@ -39,3 +39,6 @@ pub const TRIGGERS: &str = "triggers";
 pub const GLOBAL_STATUS: &str = "global_status";
 pub const SESSION_STATUS: &str = "session_status";
 pub const RUNTIME_METRICS: &str = "runtime_metrics";
+pub const PARTITIONS: &str = "partitions";
+pub const REGION_PEERS: &str = "greptime_region_peers";
+pub const TABLE_CONSTRAINTS: &str = "table_constraints";

src/catalog/src/information_schema/tables.rs

@@ -27,6 +27,7 @@ use datatypes::prelude::{ConcreteDataType, ScalarVectorBuilder, VectorRef};
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::value::Value;
 use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder};
+use futures::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 use table::metadata::TableType;
@@ -38,12 +39,13 @@ use crate::error::{
 use crate::information_schema::{InformationTable, Predicates};
 use crate::CatalogManager;
 
-const TABLE_CATALOG: &str = "table_catalog";
+pub const TABLE_CATALOG: &str = "table_catalog";
-const TABLE_SCHEMA: &str = "table_schema";
+pub const TABLE_SCHEMA: &str = "table_schema";
-const TABLE_NAME: &str = "table_name";
+pub const TABLE_NAME: &str = "table_name";
-const TABLE_TYPE: &str = "table_type";
+pub const TABLE_TYPE: &str = "table_type";
 const TABLE_ID: &str = "table_id";
 const ENGINE: &str = "engine";
+const INIT_CAPACITY: usize = 42;
 
 pub(super) struct InformationSchemaTables {
     schema: SchemaRef,
@@ -140,12 +142,12 @@ impl InformationSchemaTablesBuilder {
             schema,
             catalog_name,
             catalog_manager,
-            catalog_names: StringVectorBuilder::with_capacity(42),
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            schema_names: StringVectorBuilder::with_capacity(42),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_names: StringVectorBuilder::with_capacity(42),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_types: StringVectorBuilder::with_capacity(42),
+            table_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            table_ids: UInt32VectorBuilder::with_capacity(42),
+            table_ids: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
-            engines: StringVectorBuilder::with_capacity(42),
+            engines: StringVectorBuilder::with_capacity(INIT_CAPACITY),
         }
     }
 
@@ -159,34 +161,19 @@ impl InformationSchemaTablesBuilder {
         let predicates = Predicates::from_scan_request(&request);
 
         for schema_name in catalog_manager.schema_names(&catalog_name).await? {
-            if !catalog_manager
+            let mut stream = catalog_manager.tables(&catalog_name, &schema_name).await;
-                .schema_exists(&catalog_name, &schema_name)
-                .await?
-            {
-                continue;
-            }
 
-            for table_name in catalog_manager
+            while let Some(table) = stream.try_next().await? {
-                .table_names(&catalog_name, &schema_name)
-                .await?
-            {
-                if let Some(table) = catalog_manager
-                    .table(&catalog_name, &schema_name, &table_name)
-                    .await?
-                {
                 let table_info = table.table_info();
                 self.add_table(
                     &predicates,
                     &catalog_name,
                     &schema_name,
-                        &table_name,
+                    &table_info.name,
                     table.table_type(),
                     Some(table_info.ident.table_id),
                     Some(&table_info.meta.engine),
                 );
-                } else {
-                    unreachable!();
-                }
             }
         }
 

src/catalog/src/kvbackend.rs

@@ -12,11 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-pub use client::{CachedMetaKvBackend, MetaKvBackend};
+pub use client::{CachedMetaKvBackend, CachedMetaKvBackendBuilder, MetaKvBackend};
 
 mod client;
 mod manager;
 
-#[cfg(feature = "testing")]
-pub mod mock;
 pub use manager::KvBackendCatalogManager;

src/catalog/src/kvbackend/client.rs

@@ -14,7 +14,8 @@
 
 use std::any::Any;
 use std::fmt::Debug;
-use std::sync::Arc;
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::{Arc, Mutex};
 use std::time::Duration;
 
 use common_error::ext::BoxedError;
@@ -33,18 +34,89 @@ use meta_client::client::MetaClient;
 use moka::future::{Cache, CacheBuilder};
 use snafu::{OptionExt, ResultExt};
 
-use crate::metrics::{METRIC_CATALOG_KV_GET, METRIC_CATALOG_KV_REMOTE_GET};
+use crate::metrics::{
+    METRIC_CATALOG_KV_BATCH_GET, METRIC_CATALOG_KV_GET, METRIC_CATALOG_KV_REMOTE_GET,
+};
 
-const CACHE_MAX_CAPACITY: u64 = 10000;
+const DEFAULT_CACHE_MAX_CAPACITY: u64 = 10000;
-const CACHE_TTL_SECOND: u64 = 10 * 60;
+const DEFAULT_CACHE_TTL: Duration = Duration::from_secs(10 * 60);
-const CACHE_TTI_SECOND: u64 = 5 * 60;
+const DEFAULT_CACHE_TTI: Duration = Duration::from_secs(5 * 60);
 
-pub type CacheBackendRef = Arc<Cache<Vec<u8>, KeyValue>>;
+pub struct CachedMetaKvBackendBuilder {
+    cache_max_capacity: Option<u64>,
+    cache_ttl: Option<Duration>,
+    cache_tti: Option<Duration>,
+    meta_client: Arc<MetaClient>,
+}
 
+impl CachedMetaKvBackendBuilder {
+    pub fn new(meta_client: Arc<MetaClient>) -> Self {
+        Self {
+            cache_max_capacity: None,
+            cache_ttl: None,
+            cache_tti: None,
+            meta_client,
+        }
+    }
+
+    pub fn cache_max_capacity(mut self, cache_max_capacity: u64) -> Self {
+        self.cache_max_capacity.replace(cache_max_capacity);
+        self
+    }
+
+    pub fn cache_ttl(mut self, cache_ttl: Duration) -> Self {
+        self.cache_ttl.replace(cache_ttl);
+        self
+    }
+
+    pub fn cache_tti(mut self, cache_tti: Duration) -> Self {
+        self.cache_tti.replace(cache_tti);
+        self
+    }
+
+    pub fn build(self) -> CachedMetaKvBackend {
+        let cache_max_capacity = self
+            .cache_max_capacity
+            .unwrap_or(DEFAULT_CACHE_MAX_CAPACITY);
+        let cache_ttl = self.cache_ttl.unwrap_or(DEFAULT_CACHE_TTL);
+        let cache_tti = self.cache_tti.unwrap_or(DEFAULT_CACHE_TTI);
+
+        let cache = CacheBuilder::new(cache_max_capacity)
+            .time_to_live(cache_ttl)
+            .time_to_idle(cache_tti)
+            .build();
+
+        let kv_backend = Arc::new(MetaKvBackend {
+            client: self.meta_client,
+        });
+        let name = format!("CachedKvBackend({})", kv_backend.name());
+        let version = AtomicUsize::new(0);
+
+        CachedMetaKvBackend {
+            kv_backend,
+            cache,
+            name,
+            version,
+        }
+    }
+}
+
+pub type CacheBackend = Cache<Vec<u8>, KeyValue>;
+
+/// A wrapper of `MetaKvBackend` with cache support.
+///
+/// CachedMetaKvBackend is mainly used to read metadata information from Metasrv, and provides
+/// cache for get and batch_get. One way to trigger cache invalidation of CachedMetaKvBackend:
+/// when metadata information changes, Metasrv will broadcast a metadata invalidation request.
+///
+/// Therefore, it is recommended to use CachedMetaKvBackend to only read metadata related
+/// information. Note: If you read other information, you may read expired data, which depends on
+/// TTL and TTI for cache.
 pub struct CachedMetaKvBackend {
     kv_backend: KvBackendRef,
-    cache: CacheBackendRef,
+    cache: CacheBackend,
     name: String,
+    version: AtomicUsize,
 }
 
 impl TxnService for CachedMetaKvBackend {
@@ -96,7 +168,38 @@ impl KvBackend for CachedMetaKvBackend {
     }
 
     async fn batch_get(&self, req: BatchGetRequest) -> Result<BatchGetResponse> {
-        self.kv_backend.batch_get(req).await
+        let _timer = METRIC_CATALOG_KV_BATCH_GET.start_timer();
+
+        let mut kvs = Vec::with_capacity(req.keys.len());
+        let mut miss_keys = Vec::with_capacity(req.keys.len());
+
+        for key in req.keys {
+            if let Some(val) = self.cache.get(&key).await {
+                kvs.push(val);
+            } else {
+                miss_keys.push(key);
+            }
+        }
+
+        let batch_get_req = BatchGetRequest::new().with_keys(miss_keys.clone());
+
+        let pre_version = self.version();
+
+        let unhit_kvs = self.kv_backend.batch_get(batch_get_req).await?.kvs;
+
+        for kv in unhit_kvs.iter() {
+            self.cache.insert(kv.key().to_vec(), kv.clone()).await;
+        }
+
+        if !self.validate_version(pre_version) {
+            for key in miss_keys.iter() {
+                self.cache.invalidate(key).await;
+            }
+        }
+
+        kvs.extend(unhit_kvs);
+
+        Ok(BatchGetResponse { kvs })
     }
 
     async fn compare_and_put(&self, req: CompareAndPutRequest) -> Result<CompareAndPutResponse> {
@@ -154,8 +257,14 @@ impl KvBackend for CachedMetaKvBackend {
     async fn get(&self, key: &[u8]) -> Result<Option<KeyValue>> {
         let _timer = METRIC_CATALOG_KV_GET.start_timer();
 
+        let pre_version = Arc::new(Mutex::new(None));
+
         let init = async {
+            let version_clone = pre_version.clone();
             let _timer = METRIC_CATALOG_KV_REMOTE_GET.start_timer();
+
+            version_clone.lock().unwrap().replace(self.version());
+
             self.kv_backend.get(key).await.map(|val| {
                 val.with_context(|| CacheNotGetSnafu {
                     key: String::from_utf8_lossy(key),
@@ -166,7 +275,7 @@ impl KvBackend for CachedMetaKvBackend {
         // currently moka doesn't have `optionally_try_get_with_by_ref`
         // TODO(fys): change to moka method when available
         // https://github.com/moka-rs/moka/issues/254
-        match self.cache.try_get_with_by_ref(key, init).await {
+        let ret = match self.cache.try_get_with_by_ref(key, init).await {
             Ok(val) => Ok(Some(val)),
             Err(e) => match e.as_ref() {
                 CacheNotGet { .. } => Ok(None),
@@ -175,43 +284,65 @@ impl KvBackend for CachedMetaKvBackend {
         }
         .map_err(|e| GetKvCache {
             err_msg: e.to_string(),
-        })
+        });
+
+        // "cache.invalidate_key" and "cache.try_get_with_by_ref" are not mutually exclusive. So we need
+        // to use the version mechanism to prevent expired data from being put into the cache.
+        if pre_version
+            .lock()
+            .unwrap()
+            .as_ref()
+            .map_or(false, |v| !self.validate_version(*v))
+        {
+            self.cache.invalidate(key).await;
+        }
+
+        ret
     }
 }
 
 #[async_trait::async_trait]
 impl KvCacheInvalidator for CachedMetaKvBackend {
     async fn invalidate_key(&self, key: &[u8]) {
+        self.create_new_version();
         self.cache.invalidate(key).await;
         debug!("invalidated cache key: {}", String::from_utf8_lossy(key));
     }
 }
 
 impl CachedMetaKvBackend {
-    pub fn new(client: Arc<MetaClient>) -> Self {
+    // only for test
-        let kv_backend = Arc::new(MetaKvBackend { client });
+    #[cfg(test)]
-        Self::wrap(kv_backend)
+    fn wrap(kv_backend: KvBackendRef) -> Self {
-    }
+        let cache = CacheBuilder::new(DEFAULT_CACHE_MAX_CAPACITY)
-
+            .time_to_live(DEFAULT_CACHE_TTL)
-    pub fn wrap(kv_backend: KvBackendRef) -> Self {
+            .time_to_idle(DEFAULT_CACHE_TTI)
-        let cache = Arc::new(
+            .build();
-            CacheBuilder::new(CACHE_MAX_CAPACITY)
-                .time_to_live(Duration::from_secs(CACHE_TTL_SECOND))
-                .time_to_idle(Duration::from_secs(CACHE_TTI_SECOND))
-                .build(),
-        );
 
         let name = format!("CachedKvBackend({})", kv_backend.name());
         Self {
             kv_backend,
             cache,
             name,
+            version: AtomicUsize::new(0),
         }
     }
 
-    pub fn cache(&self) -> &CacheBackendRef {
+    pub fn cache(&self) -> &CacheBackend {
         &self.cache
     }
+
+    fn version(&self) -> usize {
+        self.version.load(Ordering::Relaxed)
+    }
+
+    fn validate_version(&self, pre_version: usize) -> bool {
+        self.version() == pre_version
+    }
+
+    fn create_new_version(&self) -> usize {
+        self.version.fetch_add(1, Ordering::Relaxed) + 1
+    }
 }
 
 #[derive(Debug)]
@@ -232,6 +363,10 @@ impl KvBackend for MetaKvBackend {
         "MetaKvBackend"
     }
 
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
     async fn range(&self, req: RangeRequest) -> Result<RangeResponse> {
         self.client
             .range(req)
@@ -240,27 +375,6 @@ impl KvBackend for MetaKvBackend {
             .context(ExternalSnafu)
     }
 
-    async fn get(&self, key: &[u8]) -> Result<Option<KeyValue>> {
-        let mut response = self
-            .client
-            .range(RangeRequest::new().with_key(key))
-            .await
-            .map_err(BoxedError::new)
-            .context(ExternalSnafu)?;
-        Ok(response.take_kvs().get_mut(0).map(|kv| KeyValue {
-            key: kv.take_key(),
-            value: kv.take_value(),
-        }))
-    }
-
-    async fn batch_put(&self, req: BatchPutRequest) -> Result<BatchPutResponse> {
-        self.client
-            .batch_put(req)
-            .await
-            .map_err(BoxedError::new)
-            .context(ExternalSnafu)
-    }
-
     async fn put(&self, req: PutRequest) -> Result<PutResponse> {
         self.client
             .put(req)
@@ -269,17 +383,9 @@ impl KvBackend for MetaKvBackend {
             .context(ExternalSnafu)
     }
 
-    async fn delete_range(&self, req: DeleteRangeRequest) -> Result<DeleteRangeResponse> {
+    async fn batch_put(&self, req: BatchPutRequest) -> Result<BatchPutResponse> {
         self.client
-            .delete_range(req)
+            .batch_put(req)
-            .await
-            .map_err(BoxedError::new)
-            .context(ExternalSnafu)
-    }
-
-    async fn batch_delete(&self, req: BatchDeleteRequest) -> Result<BatchDeleteResponse> {
-        self.client
-            .batch_delete(req)
             .await
             .map_err(BoxedError::new)
             .context(ExternalSnafu)
@@ -304,7 +410,191 @@ impl KvBackend for MetaKvBackend {
             .context(ExternalSnafu)
     }
 
+    async fn delete_range(&self, req: DeleteRangeRequest) -> Result<DeleteRangeResponse> {
+        self.client
+            .delete_range(req)
+            .await
+            .map_err(BoxedError::new)
+            .context(ExternalSnafu)
+    }
+
+    async fn batch_delete(&self, req: BatchDeleteRequest) -> Result<BatchDeleteResponse> {
+        self.client
+            .batch_delete(req)
+            .await
+            .map_err(BoxedError::new)
+            .context(ExternalSnafu)
+    }
+
+    async fn get(&self, key: &[u8]) -> Result<Option<KeyValue>> {
+        let mut response = self
+            .client
+            .range(RangeRequest::new().with_key(key))
+            .await
+            .map_err(BoxedError::new)
+            .context(ExternalSnafu)?;
+        Ok(response.take_kvs().get_mut(0).map(|kv| KeyValue {
+            key: kv.take_key(),
+            value: kv.take_value(),
+        }))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use std::any::Any;
+    use std::sync::atomic::{AtomicU32, Ordering};
+    use std::sync::Arc;
+
+    use async_trait::async_trait;
+    use common_meta::kv_backend::{KvBackend, TxnService};
+    use common_meta::rpc::store::{
+        BatchDeleteRequest, BatchDeleteResponse, BatchGetRequest, BatchGetResponse,
+        BatchPutRequest, BatchPutResponse, CompareAndPutRequest, CompareAndPutResponse,
+        DeleteRangeRequest, DeleteRangeResponse, PutRequest, PutResponse, RangeRequest,
+        RangeResponse,
+    };
+    use common_meta::rpc::KeyValue;
+    use dashmap::DashMap;
+
+    use super::CachedMetaKvBackend;
+
+    #[derive(Default)]
+    pub struct SimpleKvBackend {
+        inner_map: DashMap<Vec<u8>, Vec<u8>>,
+        get_execute_times: Arc<AtomicU32>,
+    }
+
+    impl TxnService for SimpleKvBackend {
+        type Error = common_meta::error::Error;
+    }
+
+    #[async_trait]
+    impl KvBackend for SimpleKvBackend {
+        fn name(&self) -> &str {
+            "SimpleKvBackend"
+        }
+
         fn as_any(&self) -> &dyn Any {
             self
         }
+
+        async fn batch_get(&self, req: BatchGetRequest) -> Result<BatchGetResponse, Self::Error> {
+            let mut kvs = Vec::with_capacity(req.keys.len());
+            for key in req.keys.iter() {
+                if let Some(kv) = self.get(key).await? {
+                    kvs.push(kv);
+                }
+            }
+            Ok(BatchGetResponse { kvs })
+        }
+
+        async fn put(&self, req: PutRequest) -> Result<PutResponse, Self::Error> {
+            self.inner_map.insert(req.key, req.value);
+            // always return None as prev_kv, since we don't use it in this test.
+            Ok(PutResponse { prev_kv: None })
+        }
+
+        async fn get(&self, key: &[u8]) -> Result<Option<KeyValue>, Self::Error> {
+            self.get_execute_times
+                .fetch_add(1, std::sync::atomic::Ordering::SeqCst);
+            Ok(self.inner_map.get(key).map(|v| KeyValue {
+                key: key.to_vec(),
+                value: v.value().clone(),
+            }))
+        }
+
+        async fn range(&self, _req: RangeRequest) -> Result<RangeResponse, Self::Error> {
+            unimplemented!()
+        }
+
+        async fn batch_put(&self, _req: BatchPutRequest) -> Result<BatchPutResponse, Self::Error> {
+            unimplemented!()
+        }
+
+        async fn compare_and_put(
+            &self,
+            _req: CompareAndPutRequest,
+        ) -> Result<CompareAndPutResponse, Self::Error> {
+            unimplemented!()
+        }
+
+        async fn delete_range(
+            &self,
+            _req: DeleteRangeRequest,
+        ) -> Result<DeleteRangeResponse, Self::Error> {
+            unimplemented!()
+        }
+
+        async fn batch_delete(
+            &self,
+            _req: BatchDeleteRequest,
+        ) -> Result<BatchDeleteResponse, Self::Error> {
+            unimplemented!()
+        }
+    }
+
+    #[tokio::test]
+    async fn test_cached_kv_backend() {
+        let simple_kv = Arc::new(SimpleKvBackend::default());
+        let get_execute_times = simple_kv.get_execute_times.clone();
+        let cached_kv = CachedMetaKvBackend::wrap(simple_kv);
+
+        add_some_vals(&cached_kv).await;
+
+        let batch_get_req = BatchGetRequest {
+            keys: vec![b"k1".to_vec(), b"k2".to_vec()],
+        };
+
+        assert_eq!(get_execute_times.load(Ordering::SeqCst), 0);
+
+        for _ in 0..10 {
+            let _batch_get_resp = cached_kv.batch_get(batch_get_req.clone()).await.unwrap();
+
+            assert_eq!(get_execute_times.load(Ordering::SeqCst), 2);
+        }
+
+        let batch_get_req = BatchGetRequest {
+            keys: vec![b"k1".to_vec(), b"k2".to_vec(), b"k3".to_vec()],
+        };
+
+        let _batch_get_resp = cached_kv.batch_get(batch_get_req.clone()).await.unwrap();
+
+        assert_eq!(get_execute_times.load(Ordering::SeqCst), 3);
+
+        for _ in 0..10 {
+            let _batch_get_resp = cached_kv.batch_get(batch_get_req.clone()).await.unwrap();
+
+            assert_eq!(get_execute_times.load(Ordering::SeqCst), 3);
+        }
+    }
+
+    async fn add_some_vals(kv_backend: &impl KvBackend) {
+        kv_backend
+            .put(PutRequest {
+                key: b"k1".to_vec(),
+                value: b"v1".to_vec(),
+                prev_kv: false,
+            })
+            .await
+            .unwrap();
+
+        kv_backend
+            .put(PutRequest {
+                key: b"k2".to_vec(),
+                value: b"v2".to_vec(),
+                prev_kv: false,
+            })
+            .await
+            .unwrap();
+
+        kv_backend
+            .put(PutRequest {
+                key: b"k3".to_vec(),
+                value: b"v3".to_vec(),
+                prev_kv: false,
+            })
+            .await
+            .unwrap();
+    }
 }

src/catalog/src/kvbackend/manager.rs

@@ -15,28 +15,36 @@
 use std::any::Any;
 use std::collections::BTreeSet;
 use std::sync::{Arc, Weak};
+use std::time::Duration;
 
-use common_catalog::consts::{DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, NUMBERS_TABLE_ID};
+use async_stream::try_stream;
+use common_catalog::consts::{
+    DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, NUMBERS_TABLE_ID,
+};
+use common_catalog::format_full_table_name;
 use common_error::ext::BoxedError;
-use common_meta::cache_invalidator::{CacheInvalidator, CacheInvalidatorRef, Context};
+use common_meta::cache_invalidator::{CacheInvalidator, Context, MultiCacheInvalidator};
-use common_meta::error::Result as MetaResult;
+use common_meta::instruction::CacheIdent;
 use common_meta::key::catalog_name::CatalogNameKey;
 use common_meta::key::schema_name::SchemaNameKey;
+use common_meta::key::table_info::TableInfoValue;
 use common_meta::key::table_name::TableNameKey;
 use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
 use common_meta::kv_backend::KvBackendRef;
-use common_meta::table_name::TableName;
+use futures_util::stream::BoxStream;
-use futures_util::TryStreamExt;
+use futures_util::{StreamExt, TryStreamExt};
+use moka::future::{Cache as AsyncCache, CacheBuilder};
+use moka::sync::Cache;
 use partition::manager::{PartitionRuleManager, PartitionRuleManagerRef};
 use snafu::prelude::*;
 use table::dist_table::DistTable;
-use table::metadata::TableId;
 use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
 use table::TableRef;
 
+use crate::error::Error::{GetTableCache, TableCacheNotGet};
 use crate::error::{
-    self as catalog_err, ListCatalogsSnafu, ListSchemasSnafu, Result as CatalogResult,
+    InvalidTableInfoInCatalogSnafu, ListCatalogsSnafu, ListSchemasSnafu, ListTablesSnafu, Result,
-    TableMetadataManagerSnafu,
+    TableCacheNotGetSnafu, TableMetadataManagerSnafu,
 };
 use crate::information_schema::InformationSchemaProvider;
 use crate::CatalogManager;
@@ -48,45 +56,74 @@ use crate::CatalogManager;
 /// comes from `SystemCatalog`, which is static and read-only.
 #[derive(Clone)]
 pub struct KvBackendCatalogManager {
-    // TODO(LFC): Maybe use a real implementation for Standalone mode.
-    // Now we use `NoopKvCacheInvalidator` for Standalone mode. In Standalone mode, the KV backend
-    // is implemented by RaftEngine. Maybe we need a cache for it?
-    cache_invalidator: CacheInvalidatorRef,
     partition_manager: PartitionRuleManagerRef,
     table_metadata_manager: TableMetadataManagerRef,
     /// A sub-CatalogManager that handles system tables
     system_catalog: SystemCatalog,
+    table_cache: AsyncCache<String, TableRef>,
+}
+
+struct TableCacheInvalidator {
+    table_cache: AsyncCache<String, TableRef>,
+}
+
+impl TableCacheInvalidator {
+    pub fn new(table_cache: AsyncCache<String, TableRef>) -> Self {
+        Self { table_cache }
+    }
 }
 
 #[async_trait::async_trait]
-impl CacheInvalidator for KvBackendCatalogManager {
+impl CacheInvalidator for TableCacheInvalidator {
-    async fn invalidate_table_name(&self, ctx: &Context, table_name: TableName) -> MetaResult<()> {
+    async fn invalidate(
-        self.cache_invalidator
+        &self,
-            .invalidate_table_name(ctx, table_name)
+        _ctx: &Context,
-            .await
+        caches: Vec<CacheIdent>,
+    ) -> common_meta::error::Result<()> {
+        for cache in caches {
+            if let CacheIdent::TableName(table_name) = cache {
+                let table_cache_key = format_full_table_name(
+                    &table_name.catalog_name,
+                    &table_name.schema_name,
+                    &table_name.table_name,
+                );
+                self.table_cache.invalidate(&table_cache_key).await;
+            }
+        }
+        Ok(())
+    }
 }
 
-    async fn invalidate_table_id(&self, ctx: &Context, table_id: TableId) -> MetaResult<()> {
+const CATALOG_CACHE_MAX_CAPACITY: u64 = 128;
-        self.cache_invalidator
+const TABLE_CACHE_MAX_CAPACITY: u64 = 65536;
-            .invalidate_table_id(ctx, table_id)
+const TABLE_CACHE_TTL: Duration = Duration::from_secs(10 * 60);
-            .await
+const TABLE_CACHE_TTI: Duration = Duration::from_secs(5 * 60);
-    }
-}
 
 impl KvBackendCatalogManager {
-    pub fn new(backend: KvBackendRef, cache_invalidator: CacheInvalidatorRef) -> Arc<Self> {
+    pub async fn new(
+        backend: KvBackendRef,
+        multi_cache_invalidator: Arc<MultiCacheInvalidator>,
+    ) -> Arc<Self> {
+        let table_cache: AsyncCache<String, TableRef> = CacheBuilder::new(TABLE_CACHE_MAX_CAPACITY)
+            .time_to_live(TABLE_CACHE_TTL)
+            .time_to_idle(TABLE_CACHE_TTI)
+            .build();
+        multi_cache_invalidator
+            .add_invalidator(Arc::new(TableCacheInvalidator::new(table_cache.clone())))
+            .await;
+
         Arc::new_cyclic(|me| Self {
             partition_manager: Arc::new(PartitionRuleManager::new(backend.clone())),
             table_metadata_manager: Arc::new(TableMetadataManager::new(backend)),
-            cache_invalidator,
             system_catalog: SystemCatalog {
                 catalog_manager: me.clone(),
+                catalog_cache: Cache::new(CATALOG_CACHE_MAX_CAPACITY),
                 information_schema_provider: Arc::new(InformationSchemaProvider::new(
-                    // The catalog name is not used in system_catalog, so let it empty
+                    DEFAULT_CATALOG_NAME.to_string(),
-                    "".to_string(),
                     me.clone(),
                 )),
             },
+            table_cache,
         })
     }
 
@@ -101,12 +138,15 @@ impl KvBackendCatalogManager {
 
 #[async_trait::async_trait]
 impl CatalogManager for KvBackendCatalogManager {
-    async fn catalog_names(&self) -> CatalogResult<Vec<String>> {
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
+    async fn catalog_names(&self) -> Result<Vec<String>> {
         let stream = self
             .table_metadata_manager
             .catalog_manager()
-            .catalog_names()
+            .catalog_names();
-            .await;
 
         let keys = stream
             .try_collect::<Vec<_>>()
@@ -117,12 +157,11 @@ impl CatalogManager for KvBackendCatalogManager {
         Ok(keys)
     }
 
-    async fn schema_names(&self, catalog: &str) -> CatalogResult<Vec<String>> {
+    async fn schema_names(&self, catalog: &str) -> Result<Vec<String>> {
         let stream = self
             .table_metadata_manager
             .schema_manager()
-            .schema_names(catalog)
+            .schema_names(catalog);
-            .await;
         let mut keys = stream
             .try_collect::<BTreeSet<_>>()
             .await
@@ -134,22 +173,25 @@ impl CatalogManager for KvBackendCatalogManager {
         Ok(keys.into_iter().collect())
     }
 
-    async fn table_names(&self, catalog: &str, schema: &str) -> CatalogResult<Vec<String>> {
+    async fn table_names(&self, catalog: &str, schema: &str) -> Result<Vec<String>> {
-        let mut tables = self
+        let stream = self
             .table_metadata_manager
             .table_name_manager()
-            .tables(catalog, schema)
+            .tables(catalog, schema);
+        let mut tables = stream
+            .try_collect::<Vec<_>>()
             .await
-            .context(TableMetadataManagerSnafu)?
+            .map_err(BoxedError::new)
+            .context(ListTablesSnafu { catalog, schema })?
             .into_iter()
             .map(|(k, _)| k)
-            .collect::<Vec<String>>();
+            .collect::<Vec<_>>();
         tables.extend_from_slice(&self.system_catalog.table_names(schema));
 
-        Ok(tables)
+        Ok(tables.into_iter().collect())
     }
 
-    async fn catalog_exists(&self, catalog: &str) -> CatalogResult<bool> {
+    async fn catalog_exists(&self, catalog: &str) -> Result<bool> {
         self.table_metadata_manager
             .catalog_manager()
             .exists(CatalogNameKey::new(catalog))
@@ -157,7 +199,7 @@ impl CatalogManager for KvBackendCatalogManager {
             .context(TableMetadataManagerSnafu)
     }
 
-    async fn schema_exists(&self, catalog: &str, schema: &str) -> CatalogResult<bool> {
+    async fn schema_exists(&self, catalog: &str, schema: &str) -> Result<bool> {
         if self.system_catalog.schema_exist(schema) {
             return Ok(true);
         }
@@ -169,7 +211,7 @@ impl CatalogManager for KvBackendCatalogManager {
             .context(TableMetadataManagerSnafu)
     }
 
-    async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> CatalogResult<bool> {
+    async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
         if self.system_catalog.table_exist(schema, table) {
             return Ok(true);
         }
@@ -188,20 +230,24 @@ impl CatalogManager for KvBackendCatalogManager {
         catalog: &str,
         schema: &str,
         table_name: &str,
-    ) -> CatalogResult<Option<TableRef>> {
+    ) -> Result<Option<TableRef>> {
         if let Some(table) = self.system_catalog.table(catalog, schema, table_name) {
             return Ok(Some(table));
         }
 
-        let key = TableNameKey::new(catalog, schema, table_name);
+        let init = async {
+            let table_name_key = TableNameKey::new(catalog, schema, table_name);
             let Some(table_name_value) = self
                 .table_metadata_manager
                 .table_name_manager()
-            .get(key)
+                .get(table_name_key)
                 .await
                 .context(TableMetadataManagerSnafu)?
             else {
-            return Ok(None);
+                return TableCacheNotGetSnafu {
+                    key: table_name_key.to_string(),
+                }
+                .fail();
             };
             let table_id = table_name_value.table_id();
 
@@ -213,21 +259,86 @@ impl CatalogManager for KvBackendCatalogManager {
                 .context(TableMetadataManagerSnafu)?
                 .map(|v| v.into_inner())
             else {
-            return Ok(None);
+                return TableCacheNotGetSnafu {
+                    key: table_name_key.to_string(),
+                }
+                .fail();
             };
-        let table_info = Arc::new(
+            build_table(table_info_value)
-            table_info_value
+        };
-                .table_info
+
-                .try_into()
+        match self
-                .context(catalog_err::InvalidTableInfoInCatalogSnafu)?,
+            .table_cache
-        );
+            .try_get_with_by_ref(&format_full_table_name(catalog, schema, table_name), init)
-        Ok(Some(DistTable::table(table_info)))
+            .await
+        {
+            Ok(table) => Ok(Some(table)),
+            Err(err) => match err.as_ref() {
+                TableCacheNotGet { .. } => Ok(None),
+                _ => Err(err),
+            },
+        }
+        .map_err(|err| GetTableCache {
+            err_msg: err.to_string(),
+        })
     }
 
-    fn as_any(&self) -> &dyn Any {
+    async fn tables<'a>(
-        self
+        &'a self,
+        catalog: &'a str,
+        schema: &'a str,
+    ) -> BoxStream<'a, Result<TableRef>> {
+        let sys_tables = try_stream!({
+            // System tables
+            let sys_table_names = self.system_catalog.table_names(schema);
+            for table_name in sys_table_names {
+                if let Some(table) = self.system_catalog.table(catalog, schema, &table_name) {
+                    yield table;
                 }
             }
+        });
+
+        let table_id_stream = self
+            .table_metadata_manager
+            .table_name_manager()
+            .tables(catalog, schema)
+            .map_ok(|(_, v)| v.table_id());
+        const BATCH_SIZE: usize = 128;
+        let user_tables = try_stream!({
+            // Split table ids into chunks
+            let mut table_id_chunks = table_id_stream.ready_chunks(BATCH_SIZE);
+
+            while let Some(table_ids) = table_id_chunks.next().await {
+                let table_ids = table_ids
+                    .into_iter()
+                    .collect::<std::result::Result<Vec<_>, _>>()
+                    .map_err(BoxedError::new)
+                    .context(ListTablesSnafu { catalog, schema })?;
+
+                let table_info_values = self
+                    .table_metadata_manager
+                    .table_info_manager()
+                    .batch_get(&table_ids)
+                    .await
+                    .context(TableMetadataManagerSnafu)?;
+
+                for table_info_value in table_info_values.into_values() {
+                    yield build_table(table_info_value)?;
+                }
+            }
+        });
+
+        Box::pin(sys_tables.chain(user_tables))
+    }
+}
+
+fn build_table(table_info_value: TableInfoValue) -> Result<TableRef> {
+    let table_info = table_info_value
+        .table_info
+        .try_into()
+        .context(InvalidTableInfoInCatalogSnafu)?;
+    Ok(DistTable::table(Arc::new(table_info)))
+}
 
 // TODO: This struct can hold a static map of all system tables when
 // the upper layer (e.g., procedure) can inform the catalog manager
@@ -238,6 +349,7 @@ impl CatalogManager for KvBackendCatalogManager {
 #[derive(Clone)]
 struct SystemCatalog {
     catalog_manager: Weak<KvBackendCatalogManager>,
+    catalog_cache: Cache<String, Arc<InformationSchemaProvider>>,
     information_schema_provider: Arc<InformationSchemaProvider>,
 }
 
@@ -273,7 +385,12 @@ impl SystemCatalog {
     fn table(&self, catalog: &str, schema: &str, table_name: &str) -> Option<TableRef> {
         if schema == INFORMATION_SCHEMA_NAME {
             let information_schema_provider =
-                InformationSchemaProvider::new(catalog.to_string(), self.catalog_manager.clone());
+                self.catalog_cache.get_with_by_ref(catalog, move || {
+                    Arc::new(InformationSchemaProvider::new(
+                        catalog.to_string(),
+                        self.catalog_manager.clone(),
+                    ))
+                });
             information_schema_provider.table(table_name)
         } else if schema == DEFAULT_SCHEMA_NAME && table_name == NUMBERS_TABLE_NAME {
             Some(NumbersTable::table(NUMBERS_TABLE_ID))

src/catalog/src/kvbackend/mock.rs

@@ -1,128 +0,0 @@
-// Copyright 2023 Greptime Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::collections::HashMap;
-use std::sync::{Arc, RwLock as StdRwLock};
-
-use common_recordbatch::RecordBatch;
-use datatypes::data_type::ConcreteDataType;
-use datatypes::schema::{ColumnSchema, Schema};
-use datatypes::vectors::StringVector;
-use table::engine::{CloseTableResult, EngineContext, TableEngine};
-use table::metadata::TableId;
-use table::requests::{
-    AlterTableRequest, CloseTableRequest, CreateTableRequest, DropTableRequest, OpenTableRequest,
-    TruncateTableRequest,
-};
-use table::test_util::MemTable;
-use table::TableRef;
-
-#[derive(Default)]
-pub struct MockTableEngine {
-    tables: StdRwLock<HashMap<TableId, TableRef>>,
-}
-
-#[async_trait::async_trait]
-impl TableEngine for MockTableEngine {
-    fn name(&self) -> &str {
-        "MockTableEngine"
-    }
-
-    /// Create a table with only one column
-    async fn create_table(
-        &self,
-        _ctx: &EngineContext,
-        request: CreateTableRequest,
-    ) -> table::Result<TableRef> {
-        let table_id = request.id;
-
-        let schema = Arc::new(Schema::new(vec![ColumnSchema::new(
-            "name",
-            ConcreteDataType::string_datatype(),
-            true,
-        )]));
-
-        let data = vec![Arc::new(StringVector::from(vec!["a", "b", "c"])) as _];
-        let record_batch = RecordBatch::new(schema, data).unwrap();
-        let table = MemTable::new_with_catalog(
-            &request.table_name,
-            record_batch,
-            table_id,
-            request.catalog_name,
-            request.schema_name,
-            vec![0],
-        );
-
-        let mut tables = self.tables.write().unwrap();
-        let _ = tables.insert(table_id, table.clone() as TableRef);
-        Ok(table)
-    }
-
-    async fn open_table(
-        &self,
-        _ctx: &EngineContext,
-        request: OpenTableRequest,
-    ) -> table::Result<Option<TableRef>> {
-        Ok(self.tables.read().unwrap().get(&request.table_id).cloned())
-    }
-
-    async fn alter_table(
-        &self,
-        _ctx: &EngineContext,
-        _request: AlterTableRequest,
-    ) -> table::Result<TableRef> {
-        unimplemented!()
-    }
-
-    fn get_table(
-        &self,
-        _ctx: &EngineContext,
-        table_id: TableId,
-    ) -> table::Result<Option<TableRef>> {
-        Ok(self.tables.read().unwrap().get(&table_id).cloned())
-    }
-
-    fn table_exists(&self, _ctx: &EngineContext, table_id: TableId) -> bool {
-        self.tables.read().unwrap().contains_key(&table_id)
-    }
-
-    async fn drop_table(
-        &self,
-        _ctx: &EngineContext,
-        _request: DropTableRequest,
-    ) -> table::Result<bool> {
-        unimplemented!()
-    }
-
-    async fn close_table(
-        &self,
-        _ctx: &EngineContext,
-        request: CloseTableRequest,
-    ) -> table::Result<CloseTableResult> {
-        let _ = self.tables.write().unwrap().remove(&request.table_id);
-        Ok(CloseTableResult::Released(vec![]))
-    }
-
-    async fn close(&self) -> table::Result<()> {
-        Ok(())
-    }
-
-    async fn truncate_table(
-        &self,
-        _ctx: &EngineContext,
-        _request: TruncateTableRequest,
-    ) -> table::Result<bool> {
-        Ok(true)
-    }
-}

src/catalog/src/lib.rs

@@ -19,9 +19,10 @@ use std::any::Any;
 use std::fmt::{Debug, Formatter};
 use std::sync::Arc;
 
+use api::v1::CreateTableExpr;
 use futures::future::BoxFuture;
+use futures_util::stream::BoxStream;
 use table::metadata::TableId;
-use table::requests::CreateTableRequest;
 use table::TableRef;
 
 use crate::error::Result;
@@ -56,6 +57,13 @@ pub trait CatalogManager: Send + Sync {
         schema: &str,
         table_name: &str,
     ) -> Result<Option<TableRef>>;
+
+    /// Returns all tables with a stream by catalog and schema.
+    async fn tables<'a>(
+        &'a self,
+        catalog: &'a str,
+        schema: &'a str,
+    ) -> BoxStream<'a, Result<TableRef>>;
 }
 
 pub type CatalogManagerRef = Arc<dyn CatalogManager>;
@@ -67,9 +75,9 @@ pub type OpenSystemTableHook =
 /// Register system table request:
 /// - When system table is already created and registered, the hook will be called
 ///     with table ref after opening the system table
-/// - When system table is not exists, create and register the table by create_table_request and calls open_hook with the created table.
+/// - When system table is not exists, create and register the table by `create_table_expr` and calls `open_hook` with the created table.
 pub struct RegisterSystemTableRequest {
-    pub create_table_request: CreateTableRequest,
+    pub create_table_expr: CreateTableExpr,
     pub open_hook: Option<OpenSystemTableHook>,
 }
 

src/catalog/src/memory/manager.rs

@@ -17,10 +17,12 @@ use std::collections::hash_map::Entry;
 use std::collections::HashMap;
 use std::sync::{Arc, RwLock, Weak};
 
+use async_stream::{stream, try_stream};
 use common_catalog::build_db_string;
 use common_catalog::consts::{
     DEFAULT_CATALOG_NAME, DEFAULT_PRIVATE_SCHEMA_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME,
 };
+use futures_util::stream::BoxStream;
 use snafu::OptionExt;
 use table::TableRef;
 
@@ -39,10 +41,64 @@ pub struct MemoryCatalogManager {
 
 #[async_trait::async_trait]
 impl CatalogManager for MemoryCatalogManager {
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
+    async fn catalog_names(&self) -> Result<Vec<String>> {
+        Ok(self.catalogs.read().unwrap().keys().cloned().collect())
+    }
+
+    async fn schema_names(&self, catalog: &str) -> Result<Vec<String>> {
+        Ok(self
+            .catalogs
+            .read()
+            .unwrap()
+            .get(catalog)
+            .with_context(|| CatalogNotFoundSnafu {
+                catalog_name: catalog,
+            })?
+            .keys()
+            .cloned()
+            .collect())
+    }
+
+    async fn table_names(&self, catalog: &str, schema: &str) -> Result<Vec<String>> {
+        Ok(self
+            .catalogs
+            .read()
+            .unwrap()
+            .get(catalog)
+            .with_context(|| CatalogNotFoundSnafu {
+                catalog_name: catalog,
+            })?
+            .get(schema)
+            .with_context(|| SchemaNotFoundSnafu { catalog, schema })?
+            .keys()
+            .cloned()
+            .collect())
+    }
+
+    async fn catalog_exists(&self, catalog: &str) -> Result<bool> {
+        self.catalog_exist_sync(catalog)
+    }
+
     async fn schema_exists(&self, catalog: &str, schema: &str) -> Result<bool> {
         self.schema_exist_sync(catalog, schema)
     }
 
+    async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
+        let catalogs = self.catalogs.read().unwrap();
+        Ok(catalogs
+            .get(catalog)
+            .with_context(|| CatalogNotFoundSnafu {
+                catalog_name: catalog,
+            })?
+            .get(schema)
+            .with_context(|| SchemaNotFoundSnafu { catalog, schema })?
+            .contains_key(table))
+    }
+
     async fn table(
         &self,
         catalog: &str,
@@ -61,57 +117,35 @@ impl CatalogManager for MemoryCatalogManager {
         Ok(result)
     }
 
-    async fn catalog_exists(&self, catalog: &str) -> Result<bool> {
+    async fn tables<'a>(
-        self.catalog_exist_sync(catalog)
+        &'a self,
-    }
+        catalog: &'a str,
-
+        schema: &'a str,
-    async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
+    ) -> BoxStream<'a, Result<TableRef>> {
         let catalogs = self.catalogs.read().unwrap();
-        Ok(catalogs
+
-            .get(catalog)
+        let Some(schemas) = catalogs.get(catalog) else {
-            .with_context(|| CatalogNotFoundSnafu {
+            return Box::pin(stream!({
+                yield CatalogNotFoundSnafu {
                     catalog_name: catalog,
-            })?
-            .get(schema)
-            .with_context(|| SchemaNotFoundSnafu { catalog, schema })?
-            .contains_key(table))
                 }
+                .fail();
+            }));
+        };
 
-    async fn catalog_names(&self) -> Result<Vec<String>> {
+        let Some(tables) = schemas.get(schema) else {
-        Ok(self.catalogs.read().unwrap().keys().cloned().collect())
+            return Box::pin(stream!({
+                yield SchemaNotFoundSnafu { catalog, schema }.fail();
+            }));
+        };
+
+        let tables = tables.values().cloned().collect::<Vec<_>>();
+
+        return Box::pin(try_stream!({
+            for table in tables {
+                yield table;
             }
-
+        }));
-    async fn schema_names(&self, catalog_name: &str) -> Result<Vec<String>> {
-        Ok(self
-            .catalogs
-            .read()
-            .unwrap()
-            .get(catalog_name)
-            .with_context(|| CatalogNotFoundSnafu { catalog_name })?
-            .keys()
-            .cloned()
-            .collect())
-    }
-
-    async fn table_names(&self, catalog_name: &str, schema_name: &str) -> Result<Vec<String>> {
-        Ok(self
-            .catalogs
-            .read()
-            .unwrap()
-            .get(catalog_name)
-            .with_context(|| CatalogNotFoundSnafu { catalog_name })?
-            .get(schema_name)
-            .with_context(|| SchemaNotFoundSnafu {
-                catalog: catalog_name,
-                schema: schema_name,
-            })?
-            .keys()
-            .cloned()
-            .collect())
-    }
-
-    fn as_any(&self) -> &dyn Any {
-        self
     }
 }
 
@@ -307,6 +341,7 @@ pub fn new_memory_catalog_manager() -> Result<Arc<MemoryCatalogManager>> {
 #[cfg(test)]
 mod tests {
     use common_catalog::consts::*;
+    use futures_util::TryStreamExt;
     use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
 
     use super::*;
@@ -331,8 +366,18 @@ mod tests {
                 NUMBERS_TABLE_NAME,
             )
             .await
+            .unwrap()
             .unwrap();
-        let _ = table.unwrap();
+        let stream = catalog_list
+            .tables(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME)
+            .await;
+        let tables = stream.try_collect::<Vec<_>>().await.unwrap();
+        assert_eq!(tables.len(), 1);
+        assert_eq!(
+            table.table_info().table_id(),
+            tables[0].table_info().table_id()
+        );
+
         assert!(catalog_list
             .table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "not_exists")
             .await

src/catalog/src/metrics.rs

@@ -32,4 +32,6 @@ lazy_static! {
         register_histogram!("greptime_catalog_kv_get_remote", "catalog kv get remote").unwrap();
     pub static ref METRIC_CATALOG_KV_GET: Histogram =
         register_histogram!("greptime_catalog_kv_get", "catalog kv get").unwrap();
+    pub static ref METRIC_CATALOG_KV_BATCH_GET: Histogram =
+        register_histogram!("greptime_catalog_kv_batch_get", "catalog kv batch get").unwrap();
 }