revert toml format

Signed-off-by: Ruihang Xia <waynestxia@gmail.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/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 +25,7 @@ inputs:
 runs:
   using: composite
   steps:
-    - uses: arduino/setup-protoc@v1
+    - uses: arduino/setup-protoc@v3
 
     - name: Install rust toolchain
       uses: dtolnay/rust-toolchain@master
@@ -38,7 +38,7 @@ runs:
       uses: Swatinem/rust-cache@v2
 
     - name: Install Python
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: '3.10'
 
@@ -62,10 +62,10 @@ 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

.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

@@ -19,8 +19,8 @@ jobs:
   apidoc:
     runs-on: ubuntu-20.04
     steps:
-    - uses: actions/checkout@v3
-    - uses: arduino/setup-protoc@v1
+    - uses: actions/checkout@v4
+    - uses: arduino/setup-protoc@v3
       with:
         repo-token: ${{ secrets.GITHUB_TOKEN }}
     - uses: dtolnay/rust-toolchain@master

.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

@@ -1,7 +1,7 @@
 on:
   merge_group:
   pull_request:
-    types: [opened, synchronize, reopened, ready_for_review]
+    types: [ opened, synchronize, reopened, ready_for_review ]
     paths-ignore:
       - 'docs/**'
       - 'config/**'
@@ -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
@@ -35,20 +37,19 @@ 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: ${{ 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: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -56,93 +57,170 @@ 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:
-    name: Sqlness Test
-    if: github.event.pull_request.draft == false
+  build:
+    name: Build GreptimeDB binaries
     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: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
+      - uses: dtolnay/rust-toolchain@master
         with:
-          repo-token: ${{ secrets.GITHUB_TOKEN }}
+          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" ]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
       - uses: dtolnay/rust-toolchain@master
         with:
           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: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - name: Download pre-built binaries
+        uses: actions/download-artifact@v4
         with:
-          repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
-      - name: Rust Cache
-        uses: Swatinem/rust-cache@v2
+          name: bins
+          path: .
+      - name: Unzip binaries
+        run: tar -xvf ./bins.tar.gz
       - 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
-          path: ${{ runner.temp }}/greptime-*.log
+          name: sqlness-logs-with-kafka-wal
+          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: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -151,17 +229,19 @@ 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: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: dtolnay/rust-toolchain@master
@@ -170,6 +250,10 @@ 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 +262,8 @@ jobs:
     runs-on: ubuntu-20.04-8-cores
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v3
-      - uses: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - uses: KyleMayes/install-llvm-action@v1
@@ -192,12 +276,15 @@ jobs:
           components: llvm-tools-preview
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
+        with:
+          # Shares cross multiple jobs
+          shared-key: "coverage-test"
       - 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
@@ -222,10 +309,28 @@ jobs:
           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: ubuntu-20.04
+    runs-on: ${{ matrix.os }}
+    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@v4

.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

@@ -24,8 +24,8 @@ jobs:
         os: [ windows-latest-8-cores ]
     timeout-minutes: 60
     steps:
-      - uses: actions/checkout@v4.1.0
-      - uses: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - 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: arduino/setup-protoc@v1
+      - uses: actions/checkout@v4
+      - 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

.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

@@ -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/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/

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",
@@ -52,31 +55,41 @@ members = [
     "src/store-api",
     "src/table",
     "src/index",
+    "tests-fuzz",
     "tests-integration",
     "tests/runner",
 ]
 resolver = "2"
 
 [workspace.package]
-version = "0.6.0"
+version = "0.7.0"
 edition = "2021"
 license = "Apache-2.0"
 
+[workspace.lints]
+clippy.print_stdout = "warn"
+clippy.print_stderr = "warn"
+clippy.implicit_clone = "warn"
+rust.unknown_lints = "deny"
+
 [workspace.dependencies]
 ahash = { version = "0.8", features = ["compile-time-rng"] }
 aquamarine = "0.3"
 arrow = { version = "47.0" }
 arrow-array = "47.0"
 arrow-flight = "47.0"
+arrow-ipc = { version = "47.0", features = ["lz4"] }
 arrow-schema = { version = "47.0", 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-common = { git = "https://github.com/apache/arrow-datafusion.git", rev = "26e43acac3a96cec8dd4c8365f22dfb1a84306e9" }
@@ -90,13 +103,14 @@ etcd-client = "0.12"
 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 = "96f1f0404f421ee560a4310c73c5071e49168168" }
 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"
+num_cpus = "1.16"
 once_cell = "1.18"
 opentelemetry-proto = { git = "https://github.com/waynexia/opentelemetry-rust.git", rev = "33841b38dda79b15f2024952be5f32533325ca02", features = [
     "gen-tonic",
@@ -108,7 +122,7 @@ 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"] }
@@ -120,9 +134,11 @@ reqwest = { version = "0.11", default-features = false, features = [
 rskafka = "0.5"
 rust_decimal = "1.33"
 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"
+sysinfo = "0.30"
 # on branch v0.38.x
 sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "6a93567ae38d42be5c8d08b13c8ff4dde26502ef", features = [
     "visitor",
@@ -155,7 +171,7 @@ 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 +181,7 @@ 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 +209,7 @@ 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.
-nextest:
+.PHONY: nextest
+nextest: ## Install nextest tools.
 	cargo --list | grep nextest || cargo install cargo-nextest --locked
 
 .PHONY: sqlness-test

README.md

@@ -127,12 +127,16 @@ To write and query data, GreptimeDB is compatible with multiple [protocols and c
 
 - [GreptimeDB C++ Client](https://github.com/GreptimeTeam/greptimedb-client-cpp)
 - [GreptimeDB Erlang Client](https://github.com/GreptimeTeam/greptimedb-client-erl)
-- [GreptimeDB Go Client](https://github.com/GreptimeTeam/greptimedb-client-go)
-- [GreptimeDB Java Client](https://github.com/GreptimeTeam/greptimedb-client-java)
+- [GreptimeDB Go Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-go)
+- [GreptimeDB Java Ingester](https://github.com/GreptimeTeam/greptimedb-ingester-java)
 - [GreptimeDB Python Client](https://github.com/GreptimeTeam/greptimedb-client-py) (WIP)
 - [GreptimeDB Rust Client](https://github.com/GreptimeTeam/greptimedb-client-rust)
 - [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
@@ -161,18 +165,17 @@ In addition, you may:
 
 ## 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.
 
 ## 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).
-- [Apache OpenDAL (incubating)](https://opendal.apache.org) gives GreptimeDB a very general and elegant data access abstraction layer.
+
+- 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://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

@@ -4,10 +4,13 @@ version.workspace = true
 edition.workspace = true
 license.workspace = true
 
+[lints]
+workspace = true
+
 [dependencies]
 arrow.workspace = true
 chrono.workspace = true
-clap = { version = "4.0", features = ["derive"] }
+clap.workspace = true
 client.workspace = true
 futures-util.workspace = true
 indicatif = "0.17.1"

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;
@@ -258,7 +258,7 @@ 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 +502,9 @@ 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 {
-                Output::AffectedRows(_) | Output::RecordBatches(_) => (),
-                Output::Stream(stream) => {
+            match res.data {
+                OutputData::AffectedRows(_) | OutputData::RecordBatches(_) => (),
+                OutputData::Stream(stream) => {
                     stream.try_collect::<Vec<_>>().await.unwrap();
                 }
             }

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/datanode.example.toml

@@ -14,7 +14,7 @@ require_lease_before_startup = false
 
 # Initialize all regions in the background during the startup.
 # By default, it provides services after all regions have been initialized.
-initialize_region_in_background = false
+init_regions_in_background = false
 
 [heartbeat]
 # Interval for sending heartbeat messages to the Metasrv, 3 seconds by default.
@@ -94,15 +94,18 @@ compress_manifest = false
 max_background_jobs = 4
 # 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.
 sst_write_buffer_size = "8MB"
@@ -116,6 +119,37 @@ parallel_scan_channel_size = 32
 # Whether to allow stale WAL entries read during replay.
 allow_stale_entries = false
 
+[region_engine.mito.inverted_index]
+# Whether to create the index on flush.
+# - "auto": automatically
+# - "disable": never
+create_on_flush = "auto"
+# Whether to create the index on compaction.
+# - "auto": automatically
+# - "disable": never
+create_on_compaction = "auto"
+# Whether to apply the index on query
+# - "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.
+# - "experimental": experimental memtable
+# - "time_series": time-series memtable (deprecated)
+type = "experimental"
+# The max number of keys in one shard.
+index_max_keys_per_shard = 8192
+# The max rows of data inside the actively writing buffer in one shard.
+data_freeze_threshold = 32768
+# Max dictionary bytes.
+fork_dictionary_bytes = "1GiB"
+
 # Log options, see `standalone.example.toml`
 # [logging]
 # dir = "/tmp/greptimedb/logs"

config/frontend.example.toml

@@ -31,6 +31,7 @@ runtime_size = 2
 mode = "disable"
 cert_path = ""
 key_path = ""
+watch = false
 
 # PostgresSQL server options, see `standalone.example.toml`.
 [postgres]
@@ -43,6 +44,7 @@ runtime_size = 2
 mode = "disable"
 cert_path = ""
 key_path = ""
+watch = false
 
 # OpenTSDB protocol options, see `standalone.example.toml`.
 [opentsdb]
@@ -57,6 +59,9 @@ enable = true
 # Prometheus remote storage options, see `standalone.example.toml`.
 [prom_store]
 enable = true
+# Whether to store the data from Prometheus remote write in metric engine.
+# true by default
+with_metric_engine = true
 
 # Metasrv client options, see `datanode.example.toml`.
 [meta_client]
@@ -66,6 +71,13 @@ timeout = "3s"
 ddl_timeout = "10s"
 connect_timeout = "1s"
 tcp_nodelay = true
+# The configuration about the cache of the Metadata.
+# default: 100000
+metadata_cache_max_capacity = 100000
+# default: 10m
+metadata_cache_ttl = "10m"
+# default: 5m
+metadata_cache_tti = "5m"
 
 # Log options, see `standalone.example.toml`
 # [logging]

config/standalone.example.toml

@@ -44,6 +44,8 @@ mode = "disable"
 cert_path = ""
 # Private key file path.
 key_path = ""
+# Watch for Certificate and key file change and auto reload
+watch = false
 
 # PostgresSQL server options.
 [postgres]
@@ -62,6 +64,8 @@ mode = "disable"
 cert_path = ""
 # private key file path.
 key_path = ""
+# Watch for Certificate and key file change and auto reload
+watch = false
 
 # OpenTSDB protocol options.
 [opentsdb]
@@ -81,6 +85,9 @@ enable = true
 [prom_store]
 # Whether to enable Prometheus remote write and read in HTTP API, true by default.
 enable = true
+# Whether to store the data from Prometheus remote write in metric engine.
+# true by default
+with_metric_engine = true
 
 [wal]
 # Available wal providers:
@@ -88,7 +95,25 @@ enable = true
 # - "kafka"
 provider = "raft_engine"
 
-# There're none raft-engine wal config since meta srv only involves in remote wal currently.
+# Raft-engine wal options.
+# WAL data directory
+# dir = "/tmp/greptimedb/wal"
+# WAL file size in bytes.
+file_size = "256MB"
+# WAL purge threshold.
+purge_threshold = "4GB"
+# WAL purge interval in seconds.
+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"
 
 # Kafka wal options.
 # The broker endpoints of the Kafka cluster. ["127.0.0.1:9092"] by default.
@@ -97,7 +122,7 @@ provider = "raft_engine"
 # Number of topics to be created upon start.
 # num_topics = 64
 # Topic selector type.
-# Available selector types: 
+# Available selector types:
 # - "round_robin" (default)
 # selector_type = "round_robin"
 # The prefix of topic name.
@@ -125,25 +150,6 @@ provider = "raft_engine"
 # 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.
-purge_threshold = "4GB"
-# WAL purge interval in seconds.
-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.
 [metadata_store]
 # Kv file size in bytes.
@@ -194,15 +200,18 @@ compress_manifest = false
 max_background_jobs = 4
 # 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.
 sst_write_buffer_size = "8MB"
@@ -216,6 +225,37 @@ parallel_scan_channel_size = 32
 # Whether to allow stale WAL entries read during replay.
 allow_stale_entries = false
 
+[region_engine.mito.inverted_index]
+# Whether to create the index on flush.
+# - "auto": automatically
+# - "disable": never
+create_on_flush = "auto"
+# Whether to create the index on compaction.
+# - "auto": automatically
+# - "disable": never
+create_on_compaction = "auto"
+# Whether to apply the index on query
+# - "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.
+# - "experimental": experimental memtable
+# - "time_series": time-series memtable (deprecated)
+type = "experimental"
+# The max number of keys in one shard.
+index_max_keys_per_shard = 8192
+# The max rows of data inside the actively writing buffer in one shard.
+data_freeze_threshold = 32768
+# Max dictionary bytes.
+fork_dictionary_bytes = "1GiB"
+
 # Log options
 # [logging]
 # Specify logs directory.
@@ -226,10 +266,11 @@ allow_stale_entries = 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
+# 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 = 0.0
 
 # Standalone export the metrics generated by itself
 # encoded to Prometheus remote-write format

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/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/2023-11-03-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.

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.

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 -e -x
+set -ex
 
 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/"
+        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

src/auth/Cargo.toml

@@ -8,6 +8,9 @@ license.workspace = true
 default = []
 testing = []
 
+[lints]
+workspace = true
+
 [dependencies]
 api.workspace = true
 async-trait.workspace = true

src/catalog/Cargo.toml

@@ -7,14 +7,16 @@ 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
@@ -25,6 +27,7 @@ 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
@@ -33,7 +36,7 @@ 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"

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,
@@ -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, .. } => {
-                source.status_code()
-            }
+            Error::ListCatalogs { source, .. }
+            | Error::ListSchemas { source, .. }
+            | Error::ListTables { source, .. } => source.status_code(),
 
             Error::OpenSystemCatalog { source, .. }
             | Error::CreateSystemCatalog { source, .. }
@@ -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

@@ -15,11 +15,13 @@
 mod columns;
 mod key_column_usage;
 mod memory_table;
+mod partitions;
 mod predicate;
+mod region_peers;
 mod runtime_metrics;
-mod schemata;
+pub mod schemata;
 mod table_names;
-mod tables;
+pub mod tables;
 
 use std::collections::HashMap;
 use std::sync::{Arc, Weak};
@@ -47,6 +49,8 @@ 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::tables::InformationSchemaTables;
@@ -74,6 +78,7 @@ lazy_static! {
         TRIGGERS,
         GLOBAL_STATUS,
         SESSION_STATUS,
+        PARTITIONS,
     ];
 }
 
@@ -156,6 +161,10 @@ 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());
@@ -226,6 +235,14 @@ 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 _),
             _ => None,
         }
     }
@@ -312,6 +329,7 @@ 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

@@ -31,6 +31,7 @@ use datatypes::scalars::ScalarVectorBuilder;
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::value::Value;
 use datatypes::vectors::{StringVectorBuilder, VectorRef};
+use futures::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};
 
@@ -57,6 +58,7 @@ const COLUMN_DEFAULT: &str = "column_default";
 const IS_NULLABLE: &str = "is_nullable";
 const COLUMN_TYPE: &str = "column_type";
 const COLUMN_COMMENT: &str = "column_comment";
+const INIT_CAPACITY: usize = 42;
 
 impl InformationSchemaColumns {
     pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
@@ -153,16 +155,16 @@ impl InformationSchemaColumnsBuilder {
             schema,
             catalog_name,
             catalog_manager,
-            catalog_names: StringVectorBuilder::with_capacity(42),
-            schema_names: StringVectorBuilder::with_capacity(42),
-            table_names: StringVectorBuilder::with_capacity(42),
-            column_names: StringVectorBuilder::with_capacity(42),
-            data_types: StringVectorBuilder::with_capacity(42),
-            semantic_types: StringVectorBuilder::with_capacity(42),
-            column_defaults: StringVectorBuilder::with_capacity(42),
-            is_nullables: StringVectorBuilder::with_capacity(42),
-            column_types: StringVectorBuilder::with_capacity(42),
-            column_comments: StringVectorBuilder::with_capacity(42),
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            column_names: 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),
         }
     }
 
@@ -176,44 +178,29 @@ impl InformationSchemaColumnsBuilder {
         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;
-            }
+            let mut stream = catalog_manager.tables(&catalog_name, &schema_name).await;
 
-            for table_name in catalog_manager
-                .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();
+            while let Some(table) = stream.try_next().await? {
+                let keys = &table.table_info().meta.primary_key_indices;
+                let schema = table.schema();
 
-                    for (idx, column) in schema.column_schemas().iter().enumerate() {
-                        let semantic_type = if column.is_time_index() {
-                            SEMANTIC_TYPE_TIME_INDEX
-                        } else if keys.contains(&idx) {
-                            SEMANTIC_TYPE_PRIMARY_KEY
-                        } else {
-                            SEMANTIC_TYPE_FIELD
-                        };
+                for (idx, column) in schema.column_schemas().iter().enumerate() {
+                    let semantic_type = if column.is_time_index() {
+                        SEMANTIC_TYPE_TIME_INDEX
+                    } else if keys.contains(&idx) {
+                        SEMANTIC_TYPE_PRIMARY_KEY
+                    } else {
+                        SEMANTIC_TYPE_FIELD
+                    };
 
-                        self.add_column(
-                            &predicates,
-                            &catalog_name,
-                            &schema_name,
-                            &table_name,
-                            semantic_type,
-                            column,
-                        );
-                    }
-                } else {
-                    unreachable!();
+                    self.add_column(
+                        &predicates,
+                        &catalog_name,
+                        &schema_name,
+                        &table.table_info().name,
+                        semantic_type,
+                        column,
+                    );
                 }
             }
         }

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};
 
@@ -44,6 +44,7 @@ const TABLE_SCHEMA: &str = "table_schema";
 const TABLE_NAME: &str = "table_name";
 const COLUMN_NAME: &str = "column_name";
 const ORDINAL_POSITION: &str = "ordinal_position";
+const INIT_CAPACITY: usize = 42;
 
 /// The virtual table implementation for `information_schema.KEY_COLUMN_USAGE`.
 pub(super) struct InformationSchemaKeyColumnUsage {
@@ -162,9 +163,6 @@ struct InformationSchemaKeyColumnUsageBuilder {
     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 +175,15 @@ impl InformationSchemaKeyColumnUsageBuilder {
             schema,
             catalog_name,
             catalog_manager,
-            constraint_catalog: StringVectorBuilder::with_capacity(42),
-            constraint_schema: StringVectorBuilder::with_capacity(42),
-            constraint_name: StringVectorBuilder::with_capacity(42),
-            table_catalog: StringVectorBuilder::with_capacity(42),
-            table_schema: StringVectorBuilder::with_capacity(42),
-            table_name: StringVectorBuilder::with_capacity(42),
-            column_name: StringVectorBuilder::with_capacity(42),
-            ordinal_position: UInt32VectorBuilder::with_capacity(42),
-            position_in_unique_constraint: UInt32VectorBuilder::with_capacity(42),
-            referenced_table_schema: StringVectorBuilder::with_capacity(42),
-            referenced_table_name: StringVectorBuilder::with_capacity(42),
-            referenced_column_name: StringVectorBuilder::with_capacity(42),
+            constraint_catalog: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            constraint_schema: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            constraint_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_catalog: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_schema: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            column_name: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            ordinal_position: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
+            position_in_unique_constraint: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
         }
     }
 
@@ -301,12 +296,15 @@ impl InformationSchemaKeyColumnUsageBuilder {
         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()),
@@ -317,9 +315,9 @@ impl InformationSchemaKeyColumnUsageBuilder {
             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()),
-            Arc::new(self.referenced_table_name.finish()),
-            Arc::new(self.referenced_column_name.finish()),
+            null_string_vector.clone(),
+            null_string_vector.clone(),
+            null_string_vector,
         ];
         RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
     }

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

@@ -21,8 +21,6 @@ use datatypes::vectors::{Int64Vector, StringVector};
 
 use crate::information_schema::table_names::*;
 
-const UNKNOWN: &str = "unknown";
-
 /// 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.
 pub fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>) {
@@ -72,30 +70,27 @@ pub fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>) {
             ],
         ),
 
-        BUILD_INFO => (
-            string_columns(&[
-                "GIT_BRANCH",
-                "GIT_COMMIT",
-                "GIT_COMMIT_SHORT",
-                "GIT_DIRTY",
-                "PKG_VERSION",
-            ]),
-            vec![
-                Arc::new(StringVector::from(vec![
-                    build_data::get_git_branch().unwrap_or_else(|_| UNKNOWN.to_string())
-                ])),
-                Arc::new(StringVector::from(vec![
-                    build_data::get_git_commit().unwrap_or_else(|_| UNKNOWN.to_string())
-                ])),
-                Arc::new(StringVector::from(vec![
-                    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")])),
-            ],
-        ),
+        BUILD_INFO => {
+            let build_info = common_version::build_info();
+            (
+                string_columns(&[
+                    "GIT_BRANCH",
+                    "GIT_COMMIT",
+                    "GIT_COMMIT_SHORT",
+                    "GIT_DIRTY",
+                    "PKG_VERSION",
+                ]),
+                vec![
+                    Arc::new(StringVector::from(vec![build_info.branch.to_string()])),
+                    Arc::new(StringVector::from(vec![build_info.commit.to_string()])),
+                    Arc::new(StringVector::from(vec![build_info
+                        .commit_short
+                        .to_string()])),
+                    Arc::new(StringVector::from(vec![build_info.dirty.to_string()])),
+                    Arc::new(StringVector::from(vec![build_info.version.to_string()])),
+                ],
+            )
+        }
 
         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/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";
-const SCHEMA_NAME: &str = "schema_name";
+pub const CATALOG_NAME: &str = "catalog_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),
-            schema_names: StringVectorBuilder::with_capacity(42),
-            charset_names: StringVectorBuilder::with_capacity(42),
-            collation_names: StringVectorBuilder::with_capacity(42),
-            sql_paths: StringVectorBuilder::with_capacity(42),
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            charset_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            collation_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            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_names.rs

@@ -39,3 +39,5 @@ 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";

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";
-const TABLE_SCHEMA: &str = "table_schema";
-const TABLE_NAME: &str = "table_name";
-const TABLE_TYPE: &str = "table_type";
+pub const TABLE_CATALOG: &str = "table_catalog";
+pub const TABLE_SCHEMA: &str = "table_schema";
+pub const TABLE_NAME: &str = "table_name";
+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),
-            schema_names: StringVectorBuilder::with_capacity(42),
-            table_names: StringVectorBuilder::with_capacity(42),
-            table_types: StringVectorBuilder::with_capacity(42),
-            table_ids: UInt32VectorBuilder::with_capacity(42),
-            engines: StringVectorBuilder::with_capacity(42),
+            catalog_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            schema_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_types: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            table_ids: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
+            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
-                .schema_exists(&catalog_name, &schema_name)
-                .await?
-            {
-                continue;
-            }
+            let mut stream = catalog_manager.tables(&catalog_name, &schema_name).await;
 
-            for table_name in catalog_manager
-                .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.table_type(),
-                        Some(table_info.ident.table_id),
-                        Some(&table_info.meta.engine),
-                    );
-                } else {
-                    unreachable!();
-                }
+            while let Some(table) = stream.try_next().await? {
+                let table_info = table.table_info();
+                self.add_table(
+                    &predicates,
+                    &catalog_name,
+                    &schema_name,
+                    &table_info.name,
+                    table.table_type(),
+                    Some(table_info.ident.table_id),
+                    Some(&table_info.meta.engine),
+                );
             }
         }
 

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

@@ -14,8 +14,10 @@
 
 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 std::usize;
 
 use common_error::ext::BoxedError;
 use common_meta::cache_invalidator::KvCacheInvalidator;
@@ -33,18 +35,89 @@ 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 CACHE_TTL_SECOND: u64 = 10 * 60;
-const CACHE_TTI_SECOND: u64 = 5 * 60;
+const DEFAULT_CACHE_MAX_CAPACITY: u64 = 10000;
+const DEFAULT_CACHE_TTL: Duration = Duration::from_secs(10 * 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 +169,38 @@ 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 +258,14 @@ 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 +276,7 @@ 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 +285,65 @@ 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 {
-        let kv_backend = Arc::new(MetaKvBackend { client });
-        Self::wrap(kv_backend)
-    }
-
-    pub fn wrap(kv_backend: KvBackendRef) -> Self {
-        let cache = Arc::new(
-            CacheBuilder::new(CACHE_MAX_CAPACITY)
-                .time_to_live(Duration::from_secs(CACHE_TTL_SECOND))
-                .time_to_idle(Duration::from_secs(CACHE_TTI_SECOND))
-                .build(),
-        );
+    // only for test
+    #[cfg(test)]
+    fn wrap(kv_backend: KvBackendRef) -> Self {
+        let cache = CacheBuilder::new(DEFAULT_CACHE_MAX_CAPACITY)
+            .time_to_live(DEFAULT_CACHE_TTL)
+            .time_to_idle(DEFAULT_CACHE_TTI)
+            .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)]
@@ -308,3 +440,162 @@ impl KvBackend for MetaKvBackend {
         self
     }
 }
+
+#[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> {
+            todo!()
+        }
+
+        async fn batch_put(&self, _req: BatchPutRequest) -> Result<BatchPutResponse, Self::Error> {
+            todo!()
+        }
+
+        async fn compare_and_put(
+            &self,
+            _req: CompareAndPutRequest,
+        ) -> Result<CompareAndPutResponse, Self::Error> {
+            todo!()
+        }
+
+        async fn delete_range(
+            &self,
+            _req: DeleteRangeRequest,
+        ) -> Result<DeleteRangeResponse, Self::Error> {
+            todo!()
+        }
+
+        async fn batch_delete(
+            &self,
+            _req: BatchDeleteRequest,
+        ) -> Result<BatchDeleteResponse, Self::Error> {
+            todo!()
+        }
+    }
+
+    #[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

@@ -15,18 +15,27 @@
 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::error::Result as MetaResult;
 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::TryStreamExt;
+use futures_util::stream::BoxStream;
+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;
@@ -34,9 +43,10 @@ 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,
-    TableMetadataManagerSnafu,
+    self as catalog_err, ListCatalogsSnafu, ListSchemasSnafu, ListTablesSnafu,
+    Result as CatalogResult, TableCacheNotGetSnafu, TableMetadataManagerSnafu,
 };
 use crate::information_schema::InformationSchemaProvider;
 use crate::CatalogManager;
@@ -56,23 +66,45 @@ pub struct KvBackendCatalogManager {
     table_metadata_manager: TableMetadataManagerRef,
     /// A sub-CatalogManager that handles system tables
     system_catalog: SystemCatalog,
+    table_cache: AsyncCache<String, TableRef>,
+}
+
+fn make_table(table_info_value: TableInfoValue) -> CatalogResult<TableRef> {
+    let table_info = table_info_value
+        .table_info
+        .try_into()
+        .context(catalog_err::InvalidTableInfoInCatalogSnafu)?;
+    Ok(DistTable::table(Arc::new(table_info)))
 }
 
 #[async_trait::async_trait]
 impl CacheInvalidator for KvBackendCatalogManager {
-    async fn invalidate_table_name(&self, ctx: &Context, table_name: TableName) -> MetaResult<()> {
-        self.cache_invalidator
-            .invalidate_table_name(ctx, table_name)
-            .await
-    }
-
     async fn invalidate_table_id(&self, ctx: &Context, table_id: TableId) -> MetaResult<()> {
         self.cache_invalidator
             .invalidate_table_id(ctx, table_id)
             .await
     }
+
+    async fn invalidate_table_name(&self, ctx: &Context, table_name: TableName) -> MetaResult<()> {
+        let table_cache_key = format_full_table_name(
+            &table_name.catalog_name,
+            &table_name.schema_name,
+            &table_name.table_name,
+        );
+        self.cache_invalidator
+            .invalidate_table_name(ctx, table_name)
+            .await?;
+        self.table_cache.invalidate(&table_cache_key).await;
+
+        Ok(())
+    }
 }
 
+const CATALOG_CACHE_MAX_CAPACITY: u64 = 128;
+const TABLE_CACHE_MAX_CAPACITY: u64 = 65536;
+const TABLE_CACHE_TTL: Duration = Duration::from_secs(10 * 60);
+const TABLE_CACHE_TTI: Duration = Duration::from_secs(5 * 60);
+
 impl KvBackendCatalogManager {
     pub fn new(backend: KvBackendRef, cache_invalidator: CacheInvalidatorRef) -> Arc<Self> {
         Arc::new_cyclic(|me| Self {
@@ -81,12 +113,16 @@ impl KvBackendCatalogManager {
             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
-                    "".to_string(),
+                    DEFAULT_CATALOG_NAME.to_string(),
                     me.clone(),
                 )),
             },
+            table_cache: CacheBuilder::new(TABLE_CACHE_MAX_CAPACITY)
+                .time_to_live(TABLE_CACHE_TTL)
+                .time_to_idle(TABLE_CACHE_TTI)
+                .build(),
         })
     }
 
@@ -101,6 +137,10 @@ impl KvBackendCatalogManager {
 
 #[async_trait::async_trait]
 impl CatalogManager for KvBackendCatalogManager {
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
     async fn catalog_names(&self) -> CatalogResult<Vec<String>> {
         let stream = self
             .table_metadata_manager
@@ -135,18 +175,22 @@ impl CatalogManager for KvBackendCatalogManager {
     }
 
     async fn table_names(&self, catalog: &str, schema: &str) -> CatalogResult<Vec<String>> {
-        let mut tables = self
+        let stream = self
             .table_metadata_manager
             .table_name_manager()
             .tables(catalog, schema)
+            .await;
+        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> {
@@ -193,39 +237,101 @@ impl CatalogManager for KvBackendCatalogManager {
             return Ok(Some(table));
         }
 
-        let key = TableNameKey::new(catalog, schema, table_name);
-        let Some(table_name_value) = self
-            .table_metadata_manager
-            .table_name_manager()
-            .get(key)
-            .await
-            .context(TableMetadataManagerSnafu)?
-        else {
-            return Ok(None);
-        };
-        let table_id = table_name_value.table_id();
+        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(table_name_key)
+                .await
+                .context(TableMetadataManagerSnafu)?
+            else {
+                return TableCacheNotGetSnafu {
+                    key: table_name_key.to_string(),
+                }
+                .fail();
+            };
+            let table_id = table_name_value.table_id();
 
-        let Some(table_info_value) = self
-            .table_metadata_manager
-            .table_info_manager()
-            .get(table_id)
-            .await
-            .context(TableMetadataManagerSnafu)?
-            .map(|v| v.into_inner())
-        else {
-            return Ok(None);
+            let Some(table_info_value) = self
+                .table_metadata_manager
+                .table_info_manager()
+                .get(table_id)
+                .await
+                .context(TableMetadataManagerSnafu)?
+                .map(|v| v.into_inner())
+            else {
+                return TableCacheNotGetSnafu {
+                    key: table_name_key.to_string(),
+                }
+                .fail();
+            };
+            make_table(table_info_value)
         };
-        let table_info = Arc::new(
-            table_info_value
-                .table_info
-                .try_into()
-                .context(catalog_err::InvalidTableInfoInCatalogSnafu)?,
-        );
-        Ok(Some(DistTable::table(table_info)))
+
+        match self
+            .table_cache
+            .try_get_with_by_ref(&format_full_table_name(catalog, schema, table_name), init)
+            .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 {
-        self
+    async fn tables<'a>(
+        &'a self,
+        catalog: &'a str,
+        schema: &'a str,
+    ) -> BoxStream<'a, CatalogResult<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)
+            .await
+            .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::<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 make_table(table_info_value)?;
+                }
+            }
+        });
+
+        Box::pin(sys_tables.chain(user_tables))
     }
 }
 
@@ -238,6 +344,7 @@ 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 +380,12 @@ 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

@@ -20,6 +20,7 @@ use std::fmt::{Debug, Formatter};
 use std::sync::Arc;
 
 use futures::future::BoxFuture;
+use futures_util::stream::BoxStream;
 use table::metadata::TableId;
 use table::requests::CreateTableRequest;
 use table::TableRef;
@@ -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>;

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> {
-        self.catalog_exist_sync(catalog)
-    }
-
-    async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
+    async fn tables<'a>(
+        &'a self,
+        catalog: &'a str,
+        schema: &'a str,
+    ) -> BoxStream<'a, Result<TableRef>> {
         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 catalog_names(&self) -> Result<Vec<String>> {
-        Ok(self.catalogs.read().unwrap().keys().cloned().collect())
-    }
+        let Some(schemas) = catalogs.get(catalog) else {
+            return Box::pin(stream!({
+                yield CatalogNotFoundSnafu {
+                    catalog_name: catalog,
+                }
+                .fail();
+            }));
+        };
 
-    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())
-    }
+        let Some(tables) = schemas.get(schema) else {
+            return Box::pin(stream!({
+                yield SchemaNotFoundSnafu { catalog, schema }.fail();
+            }));
+        };
 
-    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())
-    }
+        let tables = tables.values().cloned().collect::<Vec<_>>();
 
-    fn as_any(&self) -> &dyn Any {
-        self
+        return Box::pin(try_stream!({
+            for table in tables {
+                yield table;
+            }
+        }));
     }
 }
 
@@ -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();
 }

src/client/Cargo.toml

@@ -7,8 +7,12 @@ license.workspace = true
 [features]
 testing = []
 
+[lints]
+workspace = true
+
 [dependencies]
 api.workspace = true
+arc-swap = "1.6"
 arrow-flight.workspace = true
 async-stream.workspace = true
 async-trait.workspace = true
@@ -33,10 +37,12 @@ parking_lot = "0.12"
 prometheus.workspace = true
 prost.workspace = true
 rand.workspace = true
+serde.workspace = true
+serde_json.workspace = true
 session.workspace = true
 snafu.workspace = true
-tokio-stream = { workspace = true, features = ["net"] }
 tokio.workspace = true
+tokio-stream = { workspace = true, features = ["net"] }
 tonic.workspace = true
 
 [dev-dependencies]

src/client/examples/logical.rs

@@ -37,7 +37,7 @@ async fn run() {
         catalog_name: "greptime".to_string(),
         schema_name: "public".to_string(),
         table_name: "test_logical_dist_exec".to_string(),
-        desc: "".to_string(),
+        desc: String::default(),
         column_defs: vec![
             ColumnDef {
                 name: "timestamp".to_string(),

src/client/src/client.rs

@@ -122,7 +122,7 @@ impl Client {
         self.inner.set_peers(urls);
     }
 
-    fn find_channel(&self) -> Result<(String, Channel)> {
+    pub fn find_channel(&self) -> Result<(String, Channel)> {
         let addr = self
             .inner
             .get_peer()

src/client/src/database.rs

@@ -47,6 +47,9 @@ pub struct Database {
     // The dbname follows naming rule as out mysql, postgres and http
     // protocol. The server treat dbname in priority of catalog/schema.
     dbname: String,
+    // The time zone indicates the time zone where the user is located.
+    // Some queries need to be aware of the user's time zone to perform some specific actions.
+    timezone: String,
 
     client: Client,
     ctx: FlightContext,
@@ -58,7 +61,8 @@ impl Database {
         Self {
             catalog: catalog.into(),
             schema: schema.into(),
-            dbname: "".to_string(),
+            dbname: String::default(),
+            timezone: String::default(),
             client,
             ctx: FlightContext::default(),
         }
@@ -73,8 +77,9 @@ impl Database {
     /// environment
     pub fn new_with_dbname(dbname: impl Into<String>, client: Client) -> Self {
         Self {
-            catalog: "".to_string(),
-            schema: "".to_string(),
+            catalog: String::default(),
+            schema: String::default(),
+            timezone: String::default(),
             dbname: dbname.into(),
             client,
             ctx: FlightContext::default(),
@@ -105,6 +110,14 @@ impl Database {
         self.dbname = dbname.into();
     }
 
+    pub fn timezone(&self) -> &String {
+        &self.timezone
+    }
+
+    pub fn set_timezone(&mut self, timezone: impl Into<String>) {
+        self.timezone = timezone.into();
+    }
+
     pub fn set_auth(&mut self, auth: AuthScheme) {
         self.ctx.auth_header = Some(AuthHeader {
             auth_scheme: Some(auth),
@@ -161,6 +174,7 @@ impl Database {
                 schema: self.schema.clone(),
                 authorization: self.ctx.auth_header.clone(),
                 dbname: self.dbname.clone(),
+                timezone: self.timezone.clone(),
                 // TODO(Taylor-lagrange): add client grpc tracing
                 tracing_context: W3cTrace::new(),
             }),
@@ -293,34 +307,40 @@ impl Database {
                         reason: "Expect 'AffectedRows' Flight messages to be the one and the only!"
                     }
                 );
-                Ok(Output::AffectedRows(rows))
+                Ok(Output::new_with_affected_rows(rows))
             }
-            FlightMessage::Recordbatch(_) => IllegalFlightMessagesSnafu {
-                reason: "The first flight message cannot be a RecordBatch message",
+            FlightMessage::Recordbatch(_) | FlightMessage::Metrics(_) => {
+                IllegalFlightMessagesSnafu {
+                    reason: "The first flight message cannot be a RecordBatch or Metrics message",
+                }
+                .fail()
             }
-            .fail(),
             FlightMessage::Schema(schema) => {
                 let stream = Box::pin(stream!({
                     while let Some(flight_message) = flight_message_stream.next().await {
                         let flight_message = flight_message
                             .map_err(BoxedError::new)
                             .context(ExternalSnafu)?;
-                        let FlightMessage::Recordbatch(record_batch) = flight_message else {
-                            yield IllegalFlightMessagesSnafu {reason: "A Schema message must be succeeded exclusively by a set of RecordBatch messages"}
+                        match flight_message {
+                            FlightMessage::Recordbatch(record_batch) => yield Ok(record_batch),
+                            FlightMessage::Metrics(_) => {}
+                            FlightMessage::AffectedRows(_) | FlightMessage::Schema(_) => {
+                                yield IllegalFlightMessagesSnafu {reason: format!("A Schema message must be succeeded exclusively by a set of RecordBatch messages, flight_message: {:?}", flight_message)}
                                         .fail()
                                         .map_err(BoxedError::new)
                                         .context(ExternalSnafu);
-                            break;
-                        };
-                        yield Ok(record_batch);
+                                break;
+                            }
+                        }
                     }
                 }));
                 let record_batch_stream = RecordBatchStreamWrapper {
                     schema,
                     stream,
                     output_ordering: None,
+                    metrics: Default::default(),
                 };
-                Ok(Output::Stream(Box::pin(record_batch_stream)))
+                Ok(Output::new_with_stream(Box::pin(record_batch_stream)))
             }
         }
     }

src/client/src/error.rs

@@ -134,10 +134,17 @@ impl From<Status> for Error {
 
 impl Error {
     pub fn should_retry(&self) -> bool {
-        !matches!(
+        // TODO(weny): figure out each case of these codes.
+        matches!(
             self,
             Self::RegionServer {
-                code: Code::InvalidArgument,
+                code: Code::Cancelled,
+                ..
+            } | Self::RegionServer {
+                code: Code::DeadlineExceeded,
+                ..
+            } | Self::RegionServer {
+                code: Code::Unavailable,
                 ..
             }
         )

src/client/src/lib.rs

@@ -26,7 +26,7 @@ use api::v1::greptime_response::Response;
 use api::v1::{AffectedRows, GreptimeResponse};
 pub use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
 use common_error::status_code::StatusCode;
-pub use common_query::Output;
+pub use common_query::{Output, OutputData, OutputMeta};
 pub use common_recordbatch::{RecordBatches, SendableRecordBatchStream};
 use snafu::OptionExt;
 

src/client/src/region.rs

@@ -12,8 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::sync::Arc;
+
 use api::v1::region::{QueryRequest, RegionRequest, RegionResponse};
 use api::v1::ResponseHeader;
+use arc_swap::ArcSwapOption;
 use arrow_flight::Ticket;
 use async_stream::stream;
 use async_trait::async_trait;
@@ -25,6 +28,7 @@ use common_meta::error::{self as meta_error, Result as MetaResult};
 use common_recordbatch::error::ExternalSnafu;
 use common_recordbatch::{RecordBatchStreamWrapper, SendableRecordBatchStream};
 use common_telemetry::error;
+use common_telemetry::tracing_context::TracingContext;
 use prost::Message;
 use snafu::{location, Location, OptionExt, ResultExt};
 use tokio_stream::StreamExt;
@@ -119,27 +123,44 @@ impl RegionRequester {
             .fail();
         };
 
+        let metrics = Arc::new(ArcSwapOption::from(None));
+        let metrics_ref = metrics.clone();
+
+        let tracing_context = TracingContext::from_current_span();
+
         let stream = Box::pin(stream!({
+            let _span = tracing_context.attach(common_telemetry::tracing::info_span!(
+                "poll_flight_data_stream"
+            ));
             while let Some(flight_message) = flight_message_stream.next().await {
                 let flight_message = flight_message
                     .map_err(BoxedError::new)
                     .context(ExternalSnafu)?;
-                let FlightMessage::Recordbatch(record_batch) = flight_message else {
-                    yield IllegalFlightMessagesSnafu {
+
+                match flight_message {
+                    FlightMessage::Recordbatch(record_batch) => yield Ok(record_batch),
+                    FlightMessage::Metrics(s) => {
+                        let m = serde_json::from_str(&s).ok().map(Arc::new);
+                        metrics_ref.swap(m);
+                        break;
+                    }
+                    _ => {
+                        yield IllegalFlightMessagesSnafu {
                             reason: "A Schema message must be succeeded exclusively by a set of RecordBatch messages"
                         }
                         .fail()
                         .map_err(BoxedError::new)
                         .context(ExternalSnafu);
-                    break;
-                };
-                yield Ok(record_batch);
+                        break;
+                    }
+                }
             }
         }));
         let record_batch_stream = RecordBatchStreamWrapper {
             schema,
             stream,
             output_ordering: None,
+            metrics,
         };
         Ok(Box::pin(record_batch_stream))
     }
@@ -176,7 +197,7 @@ impl RegionRequester {
 
         check_response_header(header)?;
 
-        Ok(affected_rows)
+        Ok(affected_rows as _)
     }
 
     pub async fn handle(&self, request: RegionRequest) -> Result<AffectedRows> {
@@ -230,7 +251,7 @@ mod test {
         let result = check_response_header(Some(ResponseHeader {
             status: Some(PbStatus {
                 status_code: StatusCode::Success as u32,
-                err_msg: "".to_string(),
+                err_msg: String::default(),
             }),
         }));
         assert!(result.is_ok());
@@ -238,7 +259,7 @@ mod test {
         let result = check_response_header(Some(ResponseHeader {
             status: Some(PbStatus {
                 status_code: u32::MAX,
-                err_msg: "".to_string(),
+                err_msg: String::default(),
             }),
         }));
         assert!(matches!(

src/cmd/Cargo.toml

@@ -12,13 +12,16 @@ path = "src/bin/greptime.rs"
 [features]
 tokio-console = ["common-telemetry/tokio-console"]
 
+[lints]
+workspace = true
+
 [dependencies]
 anymap = "1.0.0-beta.2"
 async-trait.workspace = true
 auth.workspace = true
 catalog.workspace = true
 chrono.workspace = true
-clap = { version = "4.4", features = ["derive"] }
+clap.workspace = true
 client.workspace = true
 common-base.workspace = true
 common-catalog.workspace = true
@@ -29,10 +32,12 @@ common-meta.workspace = true
 common-procedure.workspace = true
 common-query.workspace = true
 common-recordbatch.workspace = true
+common-runtime.workspace = true
 common-telemetry = { workspace = true, features = [
     "deadlock_detection",
 ] }
 common-time.workspace = true
+common-wal.workspace = true
 config = "0.13"
 datanode.workspace = true
 datatypes.workspace = true

src/cmd/build.rs

@@ -13,5 +13,5 @@
 // limitations under the License.
 
 fn main() {
-    common_version::setup_git_versions();
+    common_version::setup_build_info();
 }

src/cmd/src/cli/bench.rs

@@ -62,7 +62,9 @@ pub struct BenchTableMetadataCommand {
 
 impl BenchTableMetadataCommand {
     pub async fn build(&self) -> Result<Instance> {
-        let etcd_store = EtcdStore::with_endpoints([&self.etcd_addr]).await.unwrap();
+        let etcd_store = EtcdStore::with_endpoints([&self.etcd_addr], 128)
+            .await
+            .unwrap();
 
         let table_metadata_manager = Arc::new(TableMetadataManager::new(etcd_store));
 
@@ -156,6 +158,7 @@ fn create_region_routes(regions: Vec<RegionNumber>) -> Vec<RegionRoute> {
             }),
             follower_peers: vec![],
             leader_status: None,
+            leader_down_since: None,
         });
     }
 

src/cmd/src/cli/export.rs

@@ -19,8 +19,7 @@ use async_trait::async_trait;
 use clap::{Parser, ValueEnum};
 use client::api::v1::auth_header::AuthScheme;
 use client::api::v1::Basic;
-use client::{Client, Database, DEFAULT_SCHEMA_NAME};
-use common_query::Output;
+use client::{Client, Database, OutputData, DEFAULT_SCHEMA_NAME};
 use common_recordbatch::util::collect;
 use common_telemetry::{debug, error, info, warn};
 use datatypes::scalars::ScalarVector;
@@ -58,8 +57,8 @@ pub struct ExportCommand {
     #[clap(long)]
     output_dir: String,
 
-    /// The name of the catalog to export. Default to "greptime-*"".
-    #[clap(long, default_value = "")]
+    /// The name of the catalog to export.
+    #[clap(long, default_value = "greptime-*")]
     database: String,
 
     /// Parallelism of the export.
@@ -142,7 +141,7 @@ impl Export {
                     .with_context(|_| RequestDatabaseSnafu {
                         sql: "show databases".to_string(),
                     })?;
-            let Output::Stream(stream) = result else {
+            let OutputData::Stream(stream) = result.data else {
                 NotDataFromOutputSnafu.fail()?
             };
             let record_batch = collect(stream)
@@ -183,7 +182,7 @@ impl Export {
             .sql(&sql)
             .await
             .with_context(|_| RequestDatabaseSnafu { sql })?;
-        let Output::Stream(stream) = result else {
+        let OutputData::Stream(stream) = result.data else {
             NotDataFromOutputSnafu.fail()?
         };
         let Some(record_batch) = collect(stream)
@@ -235,7 +234,7 @@ impl Export {
             .sql(&sql)
             .await
             .with_context(|_| RequestDatabaseSnafu { sql })?;
-        let Output::Stream(stream) = result else {
+        let OutputData::Stream(stream) = result.data else {
             NotDataFromOutputSnafu.fail()?
         };
         let record_batch = collect(stream)

src/cmd/src/cli/repl.rs

@@ -16,8 +16,10 @@ use std::path::PathBuf;
 use std::sync::Arc;
 use std::time::Instant;
 
-use catalog::kvbackend::{CachedMetaKvBackend, KvBackendCatalogManager};
-use client::{Client, Database, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
+use catalog::kvbackend::{
+    CachedMetaKvBackend, CachedMetaKvBackendBuilder, KvBackendCatalogManager,
+};
+use client::{Client, Database, OutputData, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
 use common_base::Plugins;
 use common_error::ext::ErrorExt;
 use common_query::Output;
@@ -157,19 +159,20 @@ impl Repl {
         let start = Instant::now();
 
         let output = if let Some(query_engine) = &self.query_engine {
-            let stmt = QueryLanguageParser::parse_sql(&sql)
-                .with_context(|_| ParseSqlSnafu { sql: sql.clone() })?;
-
             let query_ctx = QueryContext::with(self.database.catalog(), self.database.schema());
 
+            let stmt = QueryLanguageParser::parse_sql(&sql, &query_ctx)
+                .with_context(|_| ParseSqlSnafu { sql: sql.clone() })?;
+
             let plan = query_engine
                 .planner()
-                .plan(stmt, query_ctx)
+                .plan(stmt, query_ctx.clone())
                 .await
                 .context(PlanStatementSnafu)?;
 
-            let LogicalPlan::DfPlan(plan) =
-                query_engine.optimize(&plan).context(PlanStatementSnafu)?;
+            let LogicalPlan::DfPlan(plan) = query_engine
+                .optimize(&query_engine.engine_context(query_ctx), &plan)
+                .context(PlanStatementSnafu)?;
 
             let plan = DFLogicalSubstraitConvertor {}
                 .encode(&plan)
@@ -181,15 +184,15 @@ impl Repl {
         }
         .context(RequestDatabaseSnafu { sql: &sql })?;
 
-        let either = match output {
-            Output::Stream(s) => {
+        let either = match output.data {
+            OutputData::Stream(s) => {
                 let x = RecordBatches::try_collect(s)
                     .await
                     .context(CollectRecordBatchesSnafu)?;
                 Either::Left(x)
             }
-            Output::RecordBatches(x) => Either::Left(x),
-            Output::AffectedRows(rows) => Either::Right(rows),
+            OutputData::RecordBatches(x) => Either::Left(x),
+            OutputData::AffectedRows(rows) => Either::Right(rows),
         };
 
         let end = Instant::now();
@@ -247,7 +250,8 @@ async fn create_query_engine(meta_addr: &str) -> Result<DatafusionQueryEngine> {
         .context(StartMetaClientSnafu)?;
     let meta_client = Arc::new(meta_client);
 
-    let cached_meta_backend = Arc::new(CachedMetaKvBackend::new(meta_client.clone()));
+    let cached_meta_backend =
+        Arc::new(CachedMetaKvBackendBuilder::new(meta_client.clone()).build());
 
     let catalog_list =
         KvBackendCatalogManager::new(cached_meta_backend.clone(), cached_meta_backend);
@@ -256,6 +260,7 @@ async fn create_query_engine(meta_addr: &str) -> Result<DatafusionQueryEngine> {
         catalog_list,
         None,
         None,
+        None,
         false,
         plugins.clone(),
     ));

src/cmd/src/cli/upgrade.rs

@@ -70,7 +70,7 @@ impl UpgradeCommand {
                 etcd_addr: &self.etcd_addr,
             })?;
         let tool = MigrateTableMetadata {
-            etcd_store: EtcdStore::with_etcd_client(client),
+            etcd_store: EtcdStore::with_etcd_client(client, 128),
             dryrun: self.dryrun,
             skip_catalog_keys: self.skip_catalog_keys,
             skip_table_global_keys: self.skip_table_global_keys,

src/cmd/src/datanode.rs

@@ -18,10 +18,11 @@ use std::time::Duration;
 use async_trait::async_trait;
 use catalog::kvbackend::MetaKvBackend;
 use clap::Parser;
-use common_config::WalConfig;
 use common_telemetry::{info, logging};
+use common_wal::config::DatanodeWalConfig;
 use datanode::config::DatanodeOptions;
 use datanode::datanode::{Datanode, DatanodeBuilder};
+use datanode::service::DatanodeServiceBuilder;
 use meta_client::MetaClientOptions;
 use servers::Mode;
 use snafu::{OptionExt, ResultExt};
@@ -35,9 +36,17 @@ pub struct Instance {
 }
 
 impl Instance {
-    fn new(datanode: Datanode) -> Self {
+    pub fn new(datanode: Datanode) -> Self {
         Self { datanode }
     }
+
+    pub fn datanode_mut(&mut self) -> &mut Datanode {
+        &mut self.datanode
+    }
+
+    pub fn datanode(&self) -> &Datanode {
+        &self.datanode
+    }
 }
 
 #[async_trait]
@@ -169,7 +178,7 @@ impl StartCommand {
 
         // `wal_dir` only affects raft-engine config.
         if let Some(wal_dir) = &self.wal_dir
-            && let WalConfig::RaftEngine(raft_engine_config) = &mut opts.wal
+            && let DatanodeWalConfig::RaftEngine(raft_engine_config) = &mut opts.wal
         {
             if raft_engine_config
                 .dir
@@ -219,14 +228,20 @@ impl StartCommand {
             client: Arc::new(meta_client.clone()),
         });
 
-        let datanode = DatanodeBuilder::new(opts, plugins)
+        let mut datanode = DatanodeBuilder::new(opts.clone(), plugins)
             .with_meta_client(meta_client)
             .with_kv_backend(meta_backend)
-            .enable_region_server_service()
+            .build()
+            .await
+            .context(StartDatanodeSnafu)?;
+
+        let services = DatanodeServiceBuilder::new(&opts)
+            .with_default_grpc_server(&datanode.region_server())
             .enable_http_service()
             .build()
             .await
             .context(StartDatanodeSnafu)?;
+        datanode.setup_services(services);
 
         Ok(Instance::new(datanode))
     }
@@ -306,7 +321,7 @@ mod tests {
         assert_eq!("127.0.0.1:3001".to_string(), options.rpc_addr);
         assert_eq!(Some(42), options.node_id);
 
-        let WalConfig::RaftEngine(raft_engine_config) = options.wal else {
+        let DatanodeWalConfig::RaftEngine(raft_engine_config) = options.wal else {
             unreachable!()
         };
         assert_eq!("/other/wal", raft_engine_config.dir.unwrap());
@@ -494,7 +509,7 @@ mod tests {
                 };
 
                 // Should be read from env, env > default values.
-                let WalConfig::RaftEngine(raft_engine_config) = opts.wal else {
+                let DatanodeWalConfig::RaftEngine(raft_engine_config) = opts.wal else {
                     unreachable!()
                 };
                 assert_eq!(raft_engine_config.read_batch_size, 100);

src/cmd/src/error.rs

@@ -249,6 +249,12 @@ pub enum Error {
         source: BoxedError,
         location: Location,
     },
+
+    #[snafu(display("Failed to build runtime"))]
+    BuildRuntime {
+        location: Location,
+        source: common_runtime::error::Error,
+    },
 }
 
 pub type Result<T> = std::result::Result<T, Error>;
@@ -298,6 +304,8 @@ impl ErrorExt for Error {
             Error::SerdeJson { .. } | Error::FileIo { .. } => StatusCode::Unexpected,
 
             Error::Other { source, .. } => source.status_code(),
+
+            Error::BuildRuntime { source, .. } => source.status_code(),
         }
     }
 

src/cmd/src/frontend.rs

@@ -16,7 +16,7 @@ use std::sync::Arc;
 use std::time::Duration;
 
 use async_trait::async_trait;
-use catalog::kvbackend::CachedMetaKvBackend;
+use catalog::kvbackend::CachedMetaKvBackendBuilder;
 use clap::Parser;
 use client::client_manager::DatanodeClients;
 use common_meta::heartbeat::handler::parse_mailbox_message::ParseMailboxMessageHandler;
@@ -43,9 +43,17 @@ pub struct Instance {
 }
 
 impl Instance {
-    fn new(frontend: FeInstance) -> Self {
+    pub fn new(frontend: FeInstance) -> Self {
         Self { frontend }
     }
+
+    pub fn mut_inner(&mut self) -> &mut FeInstance {
+        &mut self.frontend
+    }
+
+    pub fn inner(&self) -> &FeInstance {
+        &self.frontend
+    }
 }
 
 #[async_trait]
@@ -224,15 +232,27 @@ impl StartCommand {
         let meta_client_options = opts.meta_client.as_ref().context(MissingConfigSnafu {
             msg: "'meta_client'",
         })?;
+
+        let cache_max_capacity = meta_client_options.metadata_cache_max_capacity;
+        let cache_ttl = meta_client_options.metadata_cache_ttl;
+        let cache_tti = meta_client_options.metadata_cache_tti;
+
         let meta_client = FeInstance::create_meta_client(meta_client_options)
             .await
             .context(StartFrontendSnafu)?;
 
-        let meta_backend = Arc::new(CachedMetaKvBackend::new(meta_client.clone()));
+        let cached_meta_backend = CachedMetaKvBackendBuilder::new(meta_client.clone())
+            .cache_max_capacity(cache_max_capacity)
+            .cache_ttl(cache_ttl)
+            .cache_tti(cache_tti)
+            .build();
+        let cached_meta_backend = Arc::new(cached_meta_backend);
 
         let executor = HandlerGroupExecutor::new(vec![
             Arc::new(ParseMailboxMessageHandler),
-            Arc::new(InvalidateTableCacheHandler::new(meta_backend.clone())),
+            Arc::new(InvalidateTableCacheHandler::new(
+                cached_meta_backend.clone(),
+            )),
         ]);
 
         let heartbeat_task = HeartbeatTask::new(
@@ -242,24 +262,23 @@ impl StartCommand {
         );
 
         let mut instance = FrontendBuilder::new(
-            meta_backend.clone(),
+            cached_meta_backend.clone(),
             Arc::new(DatanodeClients::default()),
             meta_client,
         )
-        .with_cache_invalidator(meta_backend)
+        .with_cache_invalidator(cached_meta_backend)
         .with_plugin(plugins.clone())
         .with_heartbeat_task(heartbeat_task)
         .try_build()
         .await
         .context(StartFrontendSnafu)?;
 
-        let servers = Services::new(plugins)
-            .build(opts.clone(), Arc::new(instance.clone()))
+        let servers = Services::new(opts.clone(), Arc::new(instance.clone()), plugins)
+            .build()
             .await
             .context(StartFrontendSnafu)?;
         instance
             .build_servers(opts, servers)
-            .await
             .context(StartFrontendSnafu)?;
 
         Ok(Instance::new(instance))

src/cmd/src/lib.rs

@@ -32,31 +32,35 @@ lazy_static::lazy_static! {
 }
 
 #[async_trait]
-pub trait App {
+pub trait App: Send {
     fn name(&self) -> &str;
 
+    /// A hook for implementor to make something happened before actual startup. Defaults to no-op.
+    async fn pre_start(&mut self) -> error::Result<()> {
+        Ok(())
+    }
+
     async fn start(&mut self) -> error::Result<()>;
 
     async fn stop(&self) -> error::Result<()>;
 }
 
 pub async fn start_app(mut app: Box<dyn App>) -> error::Result<()> {
-    let name = app.name().to_string();
+    info!("Starting app: {}", app.name());
 
-    tokio::select! {
-        result = app.start() => {
-            if let Err(err) = result {
-                error!(err; "Failed to start app {name}!");
-            }
-        }
-        _ = tokio::signal::ctrl_c() => {
-            if let Err(err) = app.stop().await {
-                error!(err; "Failed to stop app {name}!");
-            }
-            info!("Goodbye!");
-        }
+    app.pre_start().await?;
+
+    app.start().await?;
+
+    if let Err(e) = tokio::signal::ctrl_c().await {
+        error!("Failed to listen for ctrl-c signal: {}", e);
+        // It's unusual to fail to listen for ctrl-c signal, maybe there's something unexpected in
+        // the underlying system. So we stop the app instead of running nonetheless to let people
+        // investigate the issue.
     }
 
+    app.stop().await?;
+    info!("Goodbye!");
     Ok(())
 }
 

src/cmd/src/metasrv.rs

@@ -117,10 +117,12 @@ struct StartCommand {
     /// The working home directory of this metasrv instance.
     #[clap(long)]
     data_home: Option<String>,
-
     /// If it's not empty, the metasrv will store all data with this key prefix.
     #[clap(long, default_value = "")]
     store_key_prefix: String,
+    /// The max operations per txn
+    #[clap(long)]
+    max_txn_ops: Option<usize>,
 }
 
 impl StartCommand {
@@ -181,6 +183,10 @@ impl StartCommand {
             opts.store_key_prefix = self.store_key_prefix.clone()
         }
 
+        if let Some(max_txn_ops) = self.max_txn_ops {
+            opts.max_txn_ops = max_txn_ops;
+        }
+
         // Disable dashboard in metasrv.
         opts.http.disable_dashboard = true;
 

src/cmd/src/options.rs

@@ -14,8 +14,8 @@
 
 use clap::ArgMatches;
 use common_config::KvBackendConfig;
-use common_meta::wal::WalConfig as MetaSrvWalConfig;
 use common_telemetry::logging::{LoggingOptions, TracingOptions};
+use common_wal::config::MetaSrvWalConfig;
 use config::{Config, Environment, File, FileFormat};
 use datanode::config::{DatanodeOptions, ProcedureConfig};
 use frontend::error::{Result as FeResult, TomlFormatSnafu};
@@ -173,8 +173,8 @@ impl Options {
 mod tests {
     use std::io::Write;
 
-    use common_config::WalConfig;
     use common_test_util::temp_dir::create_named_temp_file;
+    use common_wal::config::DatanodeWalConfig;
     use datanode::config::{DatanodeOptions, ObjectStoreConfig};
 
     use super::*;
@@ -281,7 +281,7 @@ mod tests {
                 );
 
                 // Should be the values from config file, not environment variables.
-                let WalConfig::RaftEngine(raft_engine_config) = opts.wal else {
+                let DatanodeWalConfig::RaftEngine(raft_engine_config) = opts.wal else {
                     unreachable!()
                 };
                 assert_eq!(raft_engine_config.dir.unwrap(), "/tmp/greptimedb/wal");

src/cmd/src/standalone.rs

@@ -18,22 +18,22 @@ use std::{fs, path};
 use async_trait::async_trait;
 use clap::Parser;
 use common_catalog::consts::MIN_USER_TABLE_ID;
-use common_config::wal::StandaloneWalConfig;
 use common_config::{metadata_store_dir, KvBackendConfig};
 use common_meta::cache_invalidator::DummyCacheInvalidator;
 use common_meta::datanode_manager::DatanodeManagerRef;
-use common_meta::ddl::table_meta::TableMetadataAllocator;
-use common_meta::ddl::DdlTaskExecutorRef;
+use common_meta::ddl::table_meta::{TableMetadataAllocator, TableMetadataAllocatorRef};
+use common_meta::ddl::ProcedureExecutorRef;
 use common_meta::ddl_manager::DdlManager;
 use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
 use common_meta::kv_backend::KvBackendRef;
 use common_meta::region_keeper::MemoryRegionKeeper;
 use common_meta::sequence::SequenceBuilder;
-use common_meta::wal::{WalOptionsAllocator, WalOptionsAllocatorRef};
+use common_meta::wal_options_allocator::{WalOptionsAllocator, WalOptionsAllocatorRef};
 use common_procedure::ProcedureManagerRef;
 use common_telemetry::info;
 use common_telemetry::logging::LoggingOptions;
 use common_time::timezone::set_default_timezone;
+use common_wal::config::StandaloneWalConfig;
 use datanode::config::{DatanodeOptions, ProcedureConfig, RegionEngineConfig, StorageConfig};
 use datanode::datanode::{Datanode, DatanodeBuilder};
 use file_engine::config::EngineConfig as FileEngineConfig;
@@ -213,6 +213,10 @@ impl App for Instance {
             .await
             .context(StartWalOptionsAllocatorSnafu)?;
 
+        plugins::start_frontend_plugins(self.frontend.plugins().clone())
+            .await
+            .context(StartFrontendSnafu)?;
+
         self.frontend.start().await.context(StartFrontendSnafu)?;
         Ok(())
     }
@@ -368,20 +372,18 @@ impl StartCommand {
     #[allow(unused_variables)]
     #[allow(clippy::diverging_sub_expression)]
     async fn build(self, opts: MixOptions) -> Result<Instance> {
-        let mut fe_opts = opts.frontend.clone();
+        info!("Standalone start command: {:#?}", self);
+        info!("Building standalone instance with {opts:#?}");
+
+        let mut fe_opts = opts.frontend;
         #[allow(clippy::unnecessary_mut_passed)]
         let fe_plugins = plugins::setup_frontend_plugins(&mut fe_opts) // mut ref is MUST, DO NOT change it
             .await
             .context(StartFrontendSnafu)?;
 
-        let dn_opts = opts.datanode.clone();
+        let dn_opts = opts.datanode;
 
-        info!("Standalone start command: {:#?}", self);
-
-        info!("Building standalone instance with {opts:#?}");
-
-        set_default_timezone(opts.frontend.default_timezone.as_deref())
-            .context(InitTimezoneSnafu)?;
+        set_default_timezone(fe_opts.default_timezone.as_deref()).context(InitTimezoneSnafu)?;
 
         // Ensure the data_home directory exists.
         fs::create_dir_all(path::Path::new(&opts.data_home)).context(CreateDirSnafu {
@@ -417,11 +419,11 @@ impl StartCommand {
         let table_metadata_manager =
             Self::create_table_metadata_manager(kv_backend.clone()).await?;
 
-        let table_meta_allocator = TableMetadataAllocator::new(
+        let table_meta_allocator = Arc::new(TableMetadataAllocator::new(
             table_id_sequence,
             wal_options_allocator.clone(),
-            table_metadata_manager.clone(),
-        );
+            table_metadata_manager.table_name_manager().clone(),
+        ));
 
         let ddl_task_executor = Self::create_ddl_task_executor(
             table_metadata_manager,
@@ -437,13 +439,12 @@ impl StartCommand {
             .await
             .context(StartFrontendSnafu)?;
 
-        let servers = Services::new(fe_plugins)
-            .build(opts.clone(), Arc::new(frontend.clone()))
+        let servers = Services::new(fe_opts.clone(), Arc::new(frontend.clone()), fe_plugins)
+            .build()
             .await
             .context(StartFrontendSnafu)?;
         frontend
-            .build_servers(opts, servers)
-            .await
+            .build_servers(fe_opts, servers)
             .context(StartFrontendSnafu)?;
 
         Ok(Instance {
@@ -458,9 +459,9 @@ impl StartCommand {
         table_metadata_manager: TableMetadataManagerRef,
         procedure_manager: ProcedureManagerRef,
         datanode_manager: DatanodeManagerRef,
-        table_meta_allocator: TableMetadataAllocator,
-    ) -> Result<DdlTaskExecutorRef> {
-        let ddl_task_executor: DdlTaskExecutorRef = Arc::new(
+        table_meta_allocator: TableMetadataAllocatorRef,
+    ) -> Result<ProcedureExecutorRef> {
+        let procedure_executor: ProcedureExecutorRef = Arc::new(
             DdlManager::try_new(
                 procedure_manager,
                 datanode_manager,
@@ -472,7 +473,7 @@ impl StartCommand {
             .context(InitDdlManagerSnafu)?,
         );
 
-        Ok(ddl_task_executor)
+        Ok(procedure_executor)
     }
 
     pub async fn create_table_metadata_manager(
@@ -497,8 +498,8 @@ mod tests {
 
     use auth::{Identity, Password, UserProviderRef};
     use common_base::readable_size::ReadableSize;
-    use common_config::WalConfig;
     use common_test_util::temp_dir::create_named_temp_file;
+    use common_wal::config::DatanodeWalConfig;
     use datanode::config::{FileConfig, GcsConfig};
     use servers::Mode;
 
@@ -605,7 +606,7 @@ mod tests {
         assert_eq!(None, fe_opts.mysql.reject_no_database);
         assert!(fe_opts.influxdb.enable);
 
-        let WalConfig::RaftEngine(raft_engine_config) = dn_opts.wal else {
+        let DatanodeWalConfig::RaftEngine(raft_engine_config) = dn_opts.wal else {
             unreachable!()
         };
         assert_eq!("/tmp/greptimedb/test/wal", raft_engine_config.dir.unwrap());

src/common/base/Cargo.toml

@@ -4,6 +4,9 @@ version.workspace = true
 edition.workspace = true
 license.workspace = true
 
+[lints]
+workspace = true
+
 [dependencies]
 anymap = "1.0.0-beta.2"
 bitvec = "1.0"

src/common/base/src/bytes.rs

@@ -216,7 +216,7 @@ mod tests {
         let bytes = StringBytes::from(hello.clone());
         assert_eq!(bytes.len(), hello.len());
 
-        let zero = "".to_string();
+        let zero = String::default();
         let bytes = StringBytes::from(zero);
         assert!(bytes.is_empty());
     }

src/common/base/src/lib.rs

@@ -21,6 +21,8 @@ pub mod readable_size;
 use core::any::Any;
 use std::sync::{Arc, Mutex, MutexGuard};
 
+pub type AffectedRows = usize;
+
 pub use bit_vec::BitVec;
 
 /// [`Plugins`] is a wrapper of Arc contents.

src/common/base/src/readable_size.rs

@@ -33,7 +33,7 @@ pub const GIB: u64 = MIB * BINARY_DATA_MAGNITUDE;
 pub const TIB: u64 = GIB * BINARY_DATA_MAGNITUDE;
 pub const PIB: u64 = TIB * BINARY_DATA_MAGNITUDE;
 
-#[derive(Clone, Copy, PartialEq, Eq, PartialOrd)]
+#[derive(Clone, Copy, PartialEq, Eq, Ord, PartialOrd)]
 pub struct ReadableSize(pub u64);
 
 impl ReadableSize {

src/common/catalog/Cargo.toml

@@ -4,6 +4,9 @@ version.workspace = true
 edition.workspace = true
 license.workspace = true
 
+[lints]
+workspace = true
+
 [dependencies]
 common-error.workspace = true
 common-macro.workspace = true

src/common/catalog/src/consts.rs

@@ -82,6 +82,10 @@ pub const INFORMATION_SCHEMA_GLOBAL_STATUS_TABLE_ID: u32 = 25;
 pub const INFORMATION_SCHEMA_SESSION_STATUS_TABLE_ID: u32 = 26;
 /// id for information_schema.RUNTIME_METRICS
 pub const INFORMATION_SCHEMA_RUNTIME_METRICS_TABLE_ID: u32 = 27;
+/// id for information_schema.PARTITIONS
+pub const INFORMATION_SCHEMA_PARTITIONS_TABLE_ID: u32 = 28;
+/// id for information_schema.REGION_PEERS
+pub const INFORMATION_SCHEMA_REGION_PEERS_TABLE_ID: u32 = 29;
 /// ----- End of information_schema tables -----
 
 pub const MITO_ENGINE: &str = "mito";

src/common/catalog/src/lib.rs

@@ -56,11 +56,22 @@ pub fn build_db_string(catalog: &str, schema: &str) -> String {
 /// - if `[<catalog>-]` is provided, we split database name with `-` and use
 /// `<catalog>` and `<schema>`.
 pub fn parse_catalog_and_schema_from_db_string(db: &str) -> (&str, &str) {
+    match parse_optional_catalog_and_schema_from_db_string(db) {
+        (Some(catalog), schema) => (catalog, schema),
+        (None, schema) => (DEFAULT_CATALOG_NAME, schema),
+    }
+}
+
+/// Attempt to parse catalog and schema from given database name
+///
+/// Similar to [`parse_catalog_and_schema_from_db_string`] but returns an optional
+/// catalog if it's not provided in the database name.
+pub fn parse_optional_catalog_and_schema_from_db_string(db: &str) -> (Option<&str>, &str) {
     let parts = db.splitn(2, '-').collect::<Vec<&str>>();
     if parts.len() == 2 {
-        (parts[0], parts[1])
+        (Some(parts[0]), parts[1])
     } else {
-        (DEFAULT_CATALOG_NAME, db)
+        (None, db)
     }
 }
 
@@ -90,5 +101,20 @@ mod tests {
             ("catalog", "schema1-schema2"),
             parse_catalog_and_schema_from_db_string("catalog-schema1-schema2")
         );
+
+        assert_eq!(
+            (None, "fullschema"),
+            parse_optional_catalog_and_schema_from_db_string("fullschema")
+        );
+
+        assert_eq!(
+            (Some("catalog"), "schema"),
+            parse_optional_catalog_and_schema_from_db_string("catalog-schema")
+        );
+
+        assert_eq!(
+            (Some("catalog"), "schema1-schema2"),
+            parse_optional_catalog_and_schema_from_db_string("catalog-schema1-schema2")
+        );
     }
 }

src/common/config/Cargo.toml

@@ -4,11 +4,12 @@ version.workspace = true
 edition.workspace = true
 license.workspace = true
 
+[lints]
+workspace = true
+
 [dependencies]
 common-base.workspace = true
 humantime-serde.workspace = true
-rskafka.workspace = true
+num_cpus.workspace = true
 serde.workspace = true
-serde_json.workspace = true
-serde_with = "3"
-toml.workspace = true
+sysinfo.workspace = true

src/common/config/src/lib.rs

@@ -12,13 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-pub mod wal;
+pub mod utils;
 
 use common_base::readable_size::ReadableSize;
 use serde::{Deserialize, Serialize};
 
-pub use crate::wal::{KafkaWalOptions, WalConfig, WalOptions, WAL_OPTIONS_KEY};
-
 pub fn metadata_store_dir(store_dir: &str) -> String {
     format!("{store_dir}/metadata")
 }

src/common/config/src/utils.rs

@@ -0,0 +1,55 @@
+// 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 common_base::readable_size::ReadableSize;
+use sysinfo::System;
+
+/// Get the CPU core number of system, aware of cgroups.
+pub fn get_cpus() -> usize {
+    // This function will check cgroups
+    num_cpus::get()
+}
+
+/// Get the total memory of the system.
+/// If `cgroup_limits` is enabled, it will also check it.
+pub fn get_sys_total_memory() -> Option<ReadableSize> {
+    if sysinfo::IS_SUPPORTED_SYSTEM {
+        let mut sys_info = System::new();
+        sys_info.refresh_memory();
+        let mut total_memory = sys_info.total_memory();
+        // Compare with cgroups memory limit, use smaller values
+        // This method is only implemented for Linux. It always returns None for all other systems.
+        if let Some(cgroup_limits) = sys_info.cgroup_limits() {
+            total_memory = total_memory.min(cgroup_limits.total_memory)
+        }
+        Some(ReadableSize(total_memory))
+    } else {
+        None
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_get_cpus() {
+        assert!(get_cpus() > 0);
+    }
+
+    #[test]
+    fn test_get_sys_total_memory() {
+        assert!(get_sys_total_memory().unwrap() > ReadableSize::mb(0));
+    }
+}

src/common/config/src/wal.rs

@@ -1,154 +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.
-
-pub mod kafka;
-pub mod raft_engine;
-
-use serde::{Deserialize, Serialize};
-use serde_with::with_prefix;
-
-pub use crate::wal::kafka::{KafkaConfig, KafkaOptions as KafkaWalOptions, StandaloneKafkaConfig};
-pub use crate::wal::raft_engine::RaftEngineConfig;
-
-/// An encoded wal options will be wrapped into a (WAL_OPTIONS_KEY, encoded wal options) key-value pair
-/// and inserted into the options of a `RegionCreateRequest`.
-pub const WAL_OPTIONS_KEY: &str = "wal_options";
-
-/// Wal config for datanode.
-#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq)]
-#[serde(tag = "provider", rename_all = "snake_case")]
-pub enum WalConfig {
-    RaftEngine(RaftEngineConfig),
-    Kafka(KafkaConfig),
-}
-
-impl From<StandaloneWalConfig> for WalConfig {
-    fn from(value: StandaloneWalConfig) -> Self {
-        match value {
-            StandaloneWalConfig::RaftEngine(config) => WalConfig::RaftEngine(config),
-            StandaloneWalConfig::Kafka(config) => WalConfig::Kafka(config.base),
-        }
-    }
-}
-
-impl Default for WalConfig {
-    fn default() -> Self {
-        WalConfig::RaftEngine(RaftEngineConfig::default())
-    }
-}
-
-/// Wal config for datanode.
-#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq)]
-#[serde(tag = "provider", rename_all = "snake_case")]
-pub enum StandaloneWalConfig {
-    RaftEngine(RaftEngineConfig),
-    Kafka(StandaloneKafkaConfig),
-}
-
-impl Default for StandaloneWalConfig {
-    fn default() -> Self {
-        StandaloneWalConfig::RaftEngine(RaftEngineConfig::default())
-    }
-}
-
-/// Wal options allocated to a region.
-/// A wal options is encoded by metasrv with `serde_json::to_string`, and then decoded
-/// by datanode with `serde_json::from_str`.
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
-#[serde(tag = "wal.provider", rename_all = "snake_case")]
-pub enum WalOptions {
-    #[default]
-    RaftEngine,
-    #[serde(with = "prefix_wal_kafka")]
-    Kafka(KafkaWalOptions),
-}
-
-with_prefix!(prefix_wal_kafka "wal.kafka.");
-
-#[cfg(test)]
-mod tests {
-    use std::time::Duration;
-
-    use common_base::readable_size::ReadableSize;
-    use rskafka::client::partition::Compression as RsKafkaCompression;
-
-    use crate::wal::kafka::KafkaBackoffConfig;
-    use crate::wal::{KafkaConfig, KafkaWalOptions, WalOptions};
-
-    #[test]
-    fn test_serde_kafka_config() {
-        // With all fields.
-        let toml_str = r#"
-            broker_endpoints = ["127.0.0.1:9092"]
-            max_batch_size = "1MB"
-            linger = "200ms"
-            consumer_wait_timeout = "100ms"
-            backoff_init = "500ms"
-            backoff_max = "10s"
-            backoff_base = 2
-            backoff_deadline = "5mins"
-        "#;
-        let decoded: KafkaConfig = toml::from_str(toml_str).unwrap();
-        let expected = KafkaConfig {
-            broker_endpoints: vec!["127.0.0.1:9092".to_string()],
-            compression: RsKafkaCompression::default(),
-            max_batch_size: ReadableSize::mb(1),
-            linger: Duration::from_millis(200),
-            consumer_wait_timeout: Duration::from_millis(100),
-            backoff: KafkaBackoffConfig {
-                init: Duration::from_millis(500),
-                max: Duration::from_secs(10),
-                base: 2,
-                deadline: Some(Duration::from_secs(60 * 5)),
-            },
-        };
-        assert_eq!(decoded, expected);
-
-        // With some fields missing.
-        let toml_str = r#"
-            broker_endpoints = ["127.0.0.1:9092"]
-            linger = "200ms"
-        "#;
-        let decoded: KafkaConfig = toml::from_str(toml_str).unwrap();
-        let expected = KafkaConfig {
-            broker_endpoints: vec!["127.0.0.1:9092".to_string()],
-            linger: Duration::from_millis(200),
-            ..Default::default()
-        };
-        assert_eq!(decoded, expected);
-    }
-
-    #[test]
-    fn test_serde_wal_options() {
-        // Test serde raft-engine wal options.
-        let wal_options = WalOptions::RaftEngine;
-        let encoded = serde_json::to_string(&wal_options).unwrap();
-        let expected = r#"{"wal.provider":"raft_engine"}"#;
-        assert_eq!(&encoded, expected);
-
-        let decoded: WalOptions = serde_json::from_str(&encoded).unwrap();
-        assert_eq!(decoded, wal_options);
-
-        // Test serde kafka wal options.
-        let wal_options = WalOptions::Kafka(KafkaWalOptions {
-            topic: "test_topic".to_string(),
-        });
-        let encoded = serde_json::to_string(&wal_options).unwrap();
-        let expected = r#"{"wal.provider":"kafka","wal.kafka.topic":"test_topic"}"#;
-        assert_eq!(&encoded, expected);
-
-        let decoded: WalOptions = serde_json::from_str(&encoded).unwrap();
-        assert_eq!(decoded, wal_options);
-    }
-}