### Added Deduplication and Merge Functionality

- Introduced `DedupReader` and `MergeReader` in `src/mito2/src/read/sync/dedup.rs` and `src/mito2/src/read/sync/merge.rs` to handle deduplication and merging of sorted batches. ### Enhanced `BulkMemtable` Iteration - Updated `BulkMemtable` in `src/mito2/src/memtable/bulk.rs` to support deduplication and merge modes during iteration. - Added `BulkIterContext` to manage iteration context. ### Testing Enhancements - Added comprehensive tests for `BulkMemtable` and `BulkPart` in `src/mito2/src/memtable/bulk.rs` and `src/mito2/src/memtable/bulk/part.rs`. ### Code Refactoring - Made `BulkPart` and `BulkPartMeta` cloneable in `src/mito2/src/memtable/bulk/part.rs`. - Exposed internal test modules for better test coverage in `src/mito2/src/memtable/time_series.rs` and `src/mito2/src/read/merge.rs`. ### New Modules - Created `sync` module in `src/mito2/src/read.rs` to organize synchronous read operations.
feat: use spawn
2025-12-24 15:09:59 +00:00 · 2025-02-20 08:54:44 +00:00 · 2025-02-13 22:49:02 +08:00 · 2025-02-10 15:14:45 +00:00 · 2025-02-10 14:24:13 +00:00 · 2025-02-10 21:38:10 +08:00
782 changed files with 21704 additions and 38982 deletions
--- a/.cargo/config.toml
+++ b/.cargo/config.toml
@@ -3,12 +3,3 @@ linker = "aarch64-linux-gnu-gcc"

 [alias]
 sqlness = "run --bin sqlness-runner --"
-
-[unstable.git]
-shallow_index = true
-shallow_deps = true
-[unstable.gitoxide]
-fetch = true
-checkout = true
-list_files = true
-internal_use_git2 = false
--- a/.github/actions/build-dev-builder-images/action.yml
+++ b/.github/actions/build-dev-builder-images/action.yml
@@ -41,14 +41,7 @@ runs:
        username: ${{ inputs.dockerhub-image-registry-username }}
        password: ${{ inputs.dockerhub-image-registry-token }}

-    - name: Set up qemu for multi-platform builds
-      uses: docker/setup-qemu-action@v3
-      with:
-        platforms: linux/amd64,linux/arm64
-        # The latest version will lead to segmentation fault.
-        image: tonistiigi/binfmt:qemu-v7.0.0-28
-
-    - name: Build and push dev-builder-ubuntu image # Build image for amd64 and arm64 platform.
+    - name: Build and push dev-builder-ubuntu image
      shell: bash
      if: ${{ inputs.build-dev-builder-ubuntu == 'true' }}
      run: |
@@ -59,7 +52,7 @@ runs:
          IMAGE_NAMESPACE=${{ inputs.dockerhub-image-namespace }} \
          DEV_BUILDER_IMAGE_TAG=${{ inputs.version }}

-    - name: Build and push dev-builder-centos image # Only build image for amd64 platform.
+    - name: Build and push dev-builder-centos image
      shell: bash
      if: ${{ inputs.build-dev-builder-centos == 'true' }}
      run: |
@@ -76,7 +69,8 @@ runs:
      run: |
        make dev-builder \
          BASE_IMAGE=android \
-          BUILDX_MULTI_PLATFORM_BUILD=amd64 \
          IMAGE_REGISTRY=${{ inputs.dockerhub-image-registry }} \
          IMAGE_NAMESPACE=${{ inputs.dockerhub-image-namespace }} \
-          DEV_BUILDER_IMAGE_TAG=${{ inputs.version }}
+          DEV_BUILDER_IMAGE_TAG=${{ inputs.version }} && \
+
+        docker push ${{ inputs.dockerhub-image-registry }}/${{ inputs.dockerhub-image-namespace }}/dev-builder-android:${{ inputs.version }}
--- a/.github/actions/build-greptime-images/action.yml
+++ b/.github/actions/build-greptime-images/action.yml
@@ -34,8 +34,8 @@ inputs:
    required: true
  push-latest-tag:
    description: Whether to push the latest tag
-    required: true
-    default: 'false'
+    required: false
+    default: 'true'
 runs:
  using: composite
  steps:
@@ -47,11 +47,7 @@ runs:
        password: ${{ inputs.image-registry-password }}

    - name: Set up qemu for multi-platform builds
-      uses: docker/setup-qemu-action@v3
-      with:
-        platforms: linux/amd64,linux/arm64
-        # The latest version will lead to segmentation fault.
-        image: tonistiigi/binfmt:qemu-v7.0.0-28
+      uses: docker/setup-qemu-action@v2

    - name: Set up buildx
      uses: docker/setup-buildx-action@v2
--- a/.github/actions/build-images/action.yml
+++ b/.github/actions/build-images/action.yml
@@ -22,8 +22,8 @@ inputs:
    required: true
  push-latest-tag:
    description: Whether to push the latest tag
-    required: true
-    default: 'false'
+    required: false
+    default: 'true'
  dev-mode:
    description: Enable dev mode, only build standard greptime
    required: false
--- a/.github/actions/build-linux-artifacts/action.yml
+++ b/.github/actions/build-linux-artifacts/action.yml
@@ -52,7 +52,7 @@ runs:
      uses: ./.github/actions/build-greptime-binary
      with:
        base-image: ubuntu
-        features: servers/dashboard,pg_kvbackend,mysql_kvbackend
+        features: servers/dashboard,pg_kvbackend
        cargo-profile: ${{ inputs.cargo-profile }}
        artifacts-dir: greptime-linux-${{ inputs.arch }}-${{ inputs.version }}
        version: ${{ inputs.version }}
@@ -70,7 +70,7 @@ runs:
      if: ${{ inputs.arch == 'amd64' && inputs.dev-mode == 'false' }} # Builds greptime for centos if the host machine is amd64.
      with:
        base-image: centos
-        features: servers/dashboard,pg_kvbackend,mysql_kvbackend
+        features: servers/dashboard,pg_kvbackend
        cargo-profile: ${{ inputs.cargo-profile }}
        artifacts-dir: greptime-linux-${{ inputs.arch }}-centos-${{ inputs.version }}
        version: ${{ inputs.version }}
--- a/.github/actions/release-cn-artifacts/action.yaml
+++ b/.github/actions/release-cn-artifacts/action.yaml
@@ -51,8 +51,8 @@ inputs:
    required: true
  upload-to-s3:
    description: Upload to S3
-    required: true
-    default: 'false'
+    required: false
+    default: 'true'
  artifacts-dir:
    description: Directory to store artifacts
    required: false
@@ -77,21 +77,13 @@ runs:
      with:
        path: ${{ inputs.artifacts-dir }}

-    - name: Install s5cmd
-      shell: bash
-      run: |
-        wget https://github.com/peak/s5cmd/releases/download/v2.3.0/s5cmd_2.3.0_Linux-64bit.tar.gz
-        tar -xzf s5cmd_2.3.0_Linux-64bit.tar.gz
-        sudo mv s5cmd /usr/local/bin/
-        sudo chmod +x /usr/local/bin/s5cmd
-
    - name: Release artifacts to cn region
      uses: nick-invision/retry@v2
      if: ${{ inputs.upload-to-s3 == 'true' }}
      env:
        AWS_ACCESS_KEY_ID: ${{ inputs.aws-cn-access-key-id }}
        AWS_SECRET_ACCESS_KEY: ${{ inputs.aws-cn-secret-access-key }}
-        AWS_REGION: ${{ inputs.aws-cn-region }}
+        AWS_DEFAULT_REGION: ${{ inputs.aws-cn-region }}
        UPDATE_VERSION_INFO: ${{ inputs.update-version-info }}
      with:
        max_attempts: ${{ inputs.upload-max-retry-times }}
--- a/.github/actions/start-runner/action.yml
+++ b/.github/actions/start-runner/action.yml
@@ -56,7 +56,7 @@ runs:

    - name: Start EC2 runner
      if: startsWith(inputs.runner, 'ec2')
-      uses: machulav/ec2-github-runner@v2.3.8
+      uses: machulav/ec2-github-runner@v2
      id: start-linux-arm64-ec2-runner
      with:
        mode: start
--- a/.github/actions/stop-runner/action.yml
+++ b/.github/actions/stop-runner/action.yml
@@ -33,7 +33,7 @@ runs:

    - name: Stop EC2 runner
      if: ${{ inputs.label && inputs.ec2-instance-id }}
-      uses: machulav/ec2-github-runner@v2.3.8
+      uses: machulav/ec2-github-runner@v2
      with:
        mode: stop
        label: ${{ inputs.label }}
--- a/.github/scripts/upload-artifacts-to-s3.sh
+++ b/.github/scripts/upload-artifacts-to-s3.sh
@@ -33,7 +33,7 @@ function upload_artifacts() {
  #    ├── greptime-darwin-amd64-v0.2.0.sha256sum
  #    └── greptime-darwin-amd64-v0.2.0.tar.gz
  find "$ARTIFACTS_DIR" -type f \( -name "*.tar.gz" -o -name "*.sha256sum" \) | while IFS= read -r file; do
-    s5cmd cp \
+    aws s3 cp \
      "$file" "s3://$AWS_S3_BUCKET/$RELEASE_DIRS/$VERSION/$(basename "$file")"
  done
 }
@@ -45,7 +45,7 @@ function update_version_info() {
    if [[ "$VERSION" =~ ^v[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
      echo "Updating latest-version.txt"
      echo "$VERSION" > latest-version.txt
-      s5cmd cp \
+      aws s3 cp \
        latest-version.txt "s3://$AWS_S3_BUCKET/$RELEASE_DIRS/latest-version.txt"
    fi

@@ -53,7 +53,7 @@ function update_version_info() {
    if [[ "$VERSION" == *"nightly"* ]]; then
      echo "Updating latest-nightly-version.txt"
      echo "$VERSION" > latest-nightly-version.txt
-      s5cmd cp \
+      aws s3 cp \
        latest-nightly-version.txt "s3://$AWS_S3_BUCKET/$RELEASE_DIRS/latest-nightly-version.txt"
    fi
  fi
--- a/.github/workflows/apidoc.yml
+++ b/.github/workflows/apidoc.yml
@@ -14,11 +14,9 @@ name: Build API docs

 jobs:
  apidoc:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
    - uses: actions/checkout@v4
-      with:
-        persist-credentials: false
    - uses: arduino/setup-protoc@v3
      with:
        repo-token: ${{ secrets.GITHUB_TOKEN }}
--- a/.github/workflows/dependency-check.yml
+++ b/.github/workflows/dependency-check.yml
@@ -12,8 +12,6 @@ jobs:
    steps:
    - name: Checkout code
      uses: actions/checkout@v4
-      with:
-        persist-credentials: false

    - name: Set up Rust
      uses: actions-rust-lang/setup-rust-toolchain@v1
--- a/.github/workflows/dev-build.yml
+++ b/.github/workflows/dev-build.yml
@@ -16,11 +16,11 @@ on:
        description: The runner uses to build linux-amd64 artifacts
        default: ec2-c6i.4xlarge-amd64
        options:
-          - ubuntu-22.04
-          - ubuntu-22.04-8-cores
-          - ubuntu-22.04-16-cores
-          - ubuntu-22.04-32-cores
-          - ubuntu-22.04-64-cores
+          - ubuntu-20.04
+          - ubuntu-20.04-8-cores
+          - ubuntu-20.04-16-cores
+          - ubuntu-20.04-32-cores
+          - ubuntu-20.04-64-cores
          - ec2-c6i.xlarge-amd64 # 4C8G
          - ec2-c6i.2xlarge-amd64 # 8C16G
          - ec2-c6i.4xlarge-amd64 # 16C32G
@@ -76,14 +76,20 @@ env:

  NIGHTLY_RELEASE_PREFIX: nightly

+  # Use the different image name to avoid conflict with the release images.
+  IMAGE_NAME: greptimedb-dev
+
  # The source code will check out in the following path: '${WORKING_DIR}/dev/greptime'.
  CHECKOUT_GREPTIMEDB_PATH: dev/greptimedb

+permissions:
+  issues: write
+
 jobs:
  allocate-runners:
    name: Allocate runners
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    outputs:
      linux-amd64-runner: ${{ steps.start-linux-amd64-runner.outputs.label }}
      linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
@@ -101,7 +107,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Create version
        id: create-version
@@ -156,7 +161,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Checkout greptimedb
        uses: actions/checkout@v4
@@ -164,7 +168,6 @@ jobs:
          repository: ${{ inputs.repository }}
          ref: ${{ inputs.commit }}
          path: ${{ env.CHECKOUT_GREPTIMEDB_PATH }}
-          persist-credentials: true

      - uses: ./.github/actions/build-linux-artifacts
        with:
@@ -189,7 +192,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Checkout greptimedb
        uses: actions/checkout@v4
@@ -197,7 +199,6 @@ jobs:
          repository: ${{ inputs.repository }}
          ref: ${{ inputs.commit }}
          path: ${{ env.CHECKOUT_GREPTIMEDB_PATH }}
-          persist-credentials: true

      - uses: ./.github/actions/build-linux-artifacts
        with:
@@ -218,33 +219,25 @@ jobs:
      build-linux-amd64-artifacts,
      build-linux-arm64-artifacts,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    outputs:
      build-result: ${{ steps.set-build-result.outputs.build-result }}
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Build and push images to dockerhub
        uses: ./.github/actions/build-images
        with:
          image-registry: docker.io
          image-namespace: ${{ vars.IMAGE_NAMESPACE }}
-          image-name: ${{ vars.DEV_BUILD_IMAGE_NAME }}
+          image-name: ${{ env.IMAGE_NAME }}
          image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
          image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
          version: ${{ needs.allocate-runners.outputs.version }}
          push-latest-tag: false # Don't push the latest tag to registry.
          dev-mode: true # Only build the standard images.
-          
-      - name: Echo Docker image tag to step summary
-        run: |
-          echo "## Docker Image Tag" >> $GITHUB_STEP_SUMMARY
-          echo "Image Tag: \`${{ needs.allocate-runners.outputs.version }}\`" >> $GITHUB_STEP_SUMMARY
-          echo "Full Image Name: \`docker.io/${{ vars.IMAGE_NAMESPACE }}/${{ vars.DEV_BUILD_IMAGE_NAME }}:${{ needs.allocate-runners.outputs.version }}\`" >> $GITHUB_STEP_SUMMARY
-          echo "Pull Command: \`docker pull docker.io/${{ vars.IMAGE_NAMESPACE }}/${{ vars.DEV_BUILD_IMAGE_NAME }}:${{ needs.allocate-runners.outputs.version }}\`" >> $GITHUB_STEP_SUMMARY

      - name: Set build result
        id: set-build-result
@@ -258,20 +251,19 @@ jobs:
      allocate-runners,
      release-images-to-dockerhub,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    continue-on-error: true
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Release artifacts to CN region
        uses: ./.github/actions/release-cn-artifacts
        with:
          src-image-registry: docker.io
          src-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
-          src-image-name: ${{ vars.DEV_BUILD_IMAGE_NAME }}
+          src-image-name: ${{ env.IMAGE_NAME }}
          dst-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
          dst-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
          dst-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
@@ -281,7 +273,6 @@ jobs:
          aws-cn-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
          aws-cn-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
          aws-cn-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
-          upload-to-s3: false
          dev-mode: true                     # Only build the standard images(exclude centos images).
          push-latest-tag: false             # Don't push the latest tag to registry.
          update-version-info: false         # Don't update the version info in S3.
@@ -290,7 +281,7 @@ jobs:
    name: Stop linux-amd64 runner
    # Only run this job when the runner is allocated.
    if: ${{ always() }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      allocate-runners,
      build-linux-amd64-artifacts,
@@ -300,7 +291,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Stop EC2 runner
        uses: ./.github/actions/stop-runner
@@ -316,7 +306,7 @@ jobs:
    name: Stop linux-arm64 runner
    # Only run this job when the runner is allocated.
    if: ${{ always() }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      allocate-runners,
      build-linux-arm64-artifacts,
@@ -326,7 +316,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Stop EC2 runner
        uses: ./.github/actions/stop-runner
@@ -344,17 +333,11 @@ jobs:
    needs: [
      release-images-to-dockerhub
    ]
-    runs-on: ubuntu-latest
-    permissions:
-      issues: write
-
+    runs-on: ubuntu-20.04
    env:
      SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
    steps:
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Report CI status
        id: report-ci-status
--- a/.github/workflows/develop.yml
+++ b/.github/workflows/develop.yml
@@ -23,11 +23,9 @@ concurrency:
 jobs:
  check-typos-and-docs:
    name: Check typos and docs
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: crate-ci/typos@master
      - name: Check the config docs
        run: |
@@ -36,12 +34,10 @@ jobs:
          || (echo "'config/config.md' is not up-to-date, please run 'make config-docs'." && exit 1)

  license-header-check:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    name: Check License Header
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: korandoru/hawkeye@v5

  check:
@@ -49,12 +45,10 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        os: [ ubuntu-latest ]
+        os: [ ubuntu-20.04 ]
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -72,12 +66,10 @@ jobs:

  toml:
    name: Toml Check
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: actions-rust-lang/setup-rust-toolchain@v1
      - name: Install taplo
        run: cargo +stable install taplo-cli --version ^0.9 --locked --force
@@ -89,12 +81,10 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        os: [ ubuntu-latest ]
+        os: [ ubuntu-20.04 ]
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -111,7 +101,7 @@ jobs:
      - name: Build greptime binaries
        shell: bash
        # `cargo gc` will invoke `cargo build` with specified args
-        run: cargo gc -- --bin greptime --bin sqlness-runner --features "pg_kvbackend,mysql_kvbackend"
+        run: cargo gc -- --bin greptime --bin sqlness-runner --features pg_kvbackend
      - name: Pack greptime binaries
        shell: bash
        run: |
@@ -149,8 +139,6 @@ jobs:
          echo "Disk space after:"
          df -h
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -204,8 +192,6 @@ jobs:
          echo "Disk space after:"
          df -h
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -248,12 +234,10 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        os: [ ubuntu-latest ]
+        os: [ ubuntu-20.04 ]
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -270,7 +254,7 @@ jobs:
      - name: Build greptime bianry
        shell: bash
        # `cargo gc` will invoke `cargo build` with specified args
-        run: cargo gc --profile ci -- --bin greptime --features "pg_kvbackend,mysql_kvbackend"
+        run: cargo gc --profile ci -- --bin greptime --features pg_kvbackend
      - name: Pack greptime binary
        shell: bash
        run: |
@@ -311,8 +295,6 @@ jobs:
          echo "Disk space after:"
          df -h
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - name: Setup Kind
        uses: ./.github/actions/setup-kind
      - if: matrix.mode.minio
@@ -455,8 +437,6 @@ jobs:
          echo "Disk space after:"
          df -h
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - name: Setup Kind
        uses: ./.github/actions/setup-kind
      - name: Setup Chaos Mesh
@@ -568,7 +548,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        os: [ ubuntu-latest ]
+        os: [ ubuntu-20.04 ]
        mode:
          - name: "Basic"
            opts: ""
@@ -582,8 +562,6 @@ jobs:
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - if: matrix.mode.kafka
        name: Setup kafka server
        working-directory: tests-integration/fixtures
@@ -607,12 +585,10 @@ jobs:

  fmt:
    name: Rustfmt
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -624,12 +600,10 @@ jobs:

  clippy:
    name: Clippy
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -652,8 +626,6 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - name: Merge Conflict Finder
        uses: olivernybroe/action-conflict-finder@v4.0

@@ -664,8 +636,6 @@ jobs:
    needs:  [conflict-check, clippy, fmt]
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -687,7 +657,7 @@ jobs:
        working-directory: tests-integration/fixtures
        run: docker compose up -d --wait
      - name: Run nextest cases
-        run: cargo nextest run --workspace -F dashboard -F pg_kvbackend -F mysql_kvbackend
+        run: cargo nextest run --workspace -F dashboard -F pg_kvbackend
        env:
          CARGO_BUILD_RUSTFLAGS: "-C link-arg=-fuse-ld=mold"
          RUST_BACKTRACE: 1
@@ -704,19 +674,16 @@ jobs:
          GT_MINIO_ENDPOINT_URL: http://127.0.0.1:9000
          GT_ETCD_ENDPOINTS: http://127.0.0.1:2379
          GT_POSTGRES_ENDPOINTS: postgres://greptimedb:admin@127.0.0.1:5432/postgres
-          GT_MYSQL_ENDPOINTS: mysql://greptimedb:admin@127.0.0.1:3306/mysql
          GT_KAFKA_ENDPOINTS: 127.0.0.1:9092
          GT_KAFKA_SASL_ENDPOINTS: 127.0.0.1:9093
          UNITTEST_LOG_DIR: "__unittest_logs"

  coverage:
    if: github.event_name == 'merge_group'
-    runs-on: ubuntu-22.04-8-cores
+    runs-on: ubuntu-20.04-8-cores
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: arduino/setup-protoc@v3
        with:
          repo-token: ${{ secrets.GITHUB_TOKEN }}
@@ -740,7 +707,7 @@ jobs:
        working-directory: tests-integration/fixtures
        run: docker compose up -d --wait
      - name: Run nextest cases
-        run: cargo llvm-cov nextest --workspace --lcov --output-path lcov.info -F dashboard -F pg_kvbackend -F mysql_kvbackend
+        run: cargo llvm-cov nextest --workspace --lcov --output-path lcov.info -F dashboard -F pg_kvbackend
        env:
          CARGO_BUILD_RUSTFLAGS: "-C link-arg=-fuse-ld=mold"
          RUST_BACKTRACE: 1
@@ -756,7 +723,6 @@ jobs:
          GT_MINIO_ENDPOINT_URL: http://127.0.0.1:9000
          GT_ETCD_ENDPOINTS: http://127.0.0.1:2379
          GT_POSTGRES_ENDPOINTS: postgres://greptimedb:admin@127.0.0.1:5432/postgres
-          GT_MYSQL_ENDPOINTS: mysql://greptimedb:admin@127.0.0.1:3306/mysql
          GT_KAFKA_ENDPOINTS: 127.0.0.1:9092
          GT_KAFKA_SASL_ENDPOINTS: 127.0.0.1:9093
          UNITTEST_LOG_DIR: "__unittest_logs"
@@ -772,7 +738,7 @@ jobs:
  # compat:
  #   name: Compatibility Test
  #   needs: build
-  #   runs-on: ubuntu-22.04
+  #   runs-on: ubuntu-20.04
  #   timeout-minutes: 60
  #   steps:
  #     - uses: actions/checkout@v4
--- a/.github/workflows/docbot.yml
+++ b/.github/workflows/docbot.yml
@@ -3,21 +3,16 @@ on:
  pull_request_target:
    types: [opened, edited]

-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
-  cancel-in-progress: true
+permissions:
+  pull-requests: write
+  contents: read

 jobs:
  docbot:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-      contents: read
+    runs-on: ubuntu-20.04
    timeout-minutes: 10
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Maybe Follow Up Docs Issue
        working-directory: cyborg
--- a/.github/workflows/docs.yml
+++ b/.github/workflows/docs.yml
@@ -31,47 +31,43 @@ name: CI
 jobs:
  typos:
    name: Spell Check with Typos
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: crate-ci/typos@master

  license-header-check:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    name: Check License Header
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: korandoru/hawkeye@v5

  check:
    name: Check
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - run: 'echo "No action required"'

  fmt:
    name: Rustfmt
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - run: 'echo "No action required"'

  clippy:
    name: Clippy
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - run: 'echo "No action required"'

  coverage:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - run: 'echo "No action required"'

  test:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - run: 'echo "No action required"'

@@ -80,7 +76,7 @@ jobs:
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
-        os: [ ubuntu-latest ]
+        os: [ ubuntu-20.04 ]
        mode:
          - name: "Basic"
          - name: "Remote WAL"
--- a/.github/workflows/grafana.yml
+++ b/.github/workflows/grafana.yml
@@ -1,52 +0,0 @@
-name: Check Grafana Panels
-
-on:
-  pull_request:
-    branches:
-      - main
-    paths:
-      - 'grafana/**'  # Trigger only when files under the grafana/ directory change
-
-jobs:
-  check-panels:
-    runs-on: ubuntu-latest
-
-    steps:
-      # Check out the repository
-      - name: Checkout repository
-        uses: actions/checkout@v4
-
-      # Install jq (required for the script)
-      - name: Install jq
-        run: sudo apt-get install -y jq
-
-      # Make the check.sh script executable
-      - name: Make check.sh executable
-        run: chmod +x grafana/check.sh
-
-      # Run the check.sh script
-      - name: Run check.sh
-        run: ./grafana/check.sh
-
-      # Only run summary.sh for pull_request events (not for merge queues or final pushes)
-      - name: Check if this is a pull request
-        id: check-pr
-        run: |
-          if [[ "${{ github.event_name }}" == "pull_request" ]]; then
-            echo "is_pull_request=true" >> $GITHUB_OUTPUT
-          else
-            echo "is_pull_request=false" >> $GITHUB_OUTPUT
-          fi
-
-      # Make the summary.sh script executable
-      - name: Make summary.sh executable
-        if: steps.check-pr.outputs.is_pull_request == 'true'
-        run: chmod +x grafana/summary.sh
-
-      # Run the summary.sh script and add its output to the GitHub Job Summary
-      - name: Run summary.sh and add to Job Summary
-        if: steps.check-pr.outputs.is_pull_request == 'true'
-        run: |
-          SUMMARY=$(./grafana/summary.sh)
-          echo "### Summary of Grafana Panels" >> $GITHUB_STEP_SUMMARY
-          echo "$SUMMARY" >> $GITHUB_STEP_SUMMARY
--- a/.github/workflows/nightly-build.yml
+++ b/.github/workflows/nightly-build.yml
@@ -14,11 +14,11 @@ on:
        description: The runner uses to build linux-amd64 artifacts
        default: ec2-c6i.4xlarge-amd64
        options:
-          - ubuntu-22.04
-          - ubuntu-22.04-8-cores
-          - ubuntu-22.04-16-cores
-          - ubuntu-22.04-32-cores
-          - ubuntu-22.04-64-cores
+          - ubuntu-20.04
+          - ubuntu-20.04-8-cores
+          - ubuntu-20.04-16-cores
+          - ubuntu-20.04-32-cores
+          - ubuntu-20.04-64-cores
          - ec2-c6i.xlarge-amd64 # 4C8G
          - ec2-c6i.2xlarge-amd64 # 8C16G
          - ec2-c6i.4xlarge-amd64 # 16C32G
@@ -66,11 +66,18 @@ env:

  NIGHTLY_RELEASE_PREFIX: nightly

+  # Use the different image name to avoid conflict with the release images.
+  # The DockerHub image will be greptime/greptimedb-nightly.
+  IMAGE_NAME: greptimedb-nightly
+
+permissions:
+  issues: write
+
 jobs:
  allocate-runners:
    name: Allocate runners
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    outputs:
      linux-amd64-runner: ${{ steps.start-linux-amd64-runner.outputs.label }}
      linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
@@ -88,7 +95,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Create version
        id: create-version
@@ -141,7 +147,6 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - uses: ./.github/actions/build-linux-artifacts
        with:
@@ -163,7 +168,6 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - uses: ./.github/actions/build-linux-artifacts
        with:
@@ -182,25 +186,24 @@ jobs:
      build-linux-amd64-artifacts,
      build-linux-arm64-artifacts,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    outputs:
      nightly-build-result: ${{ steps.set-nightly-build-result.outputs.nightly-build-result }}
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Build and push images to dockerhub
        uses: ./.github/actions/build-images
        with:
          image-registry: docker.io
          image-namespace: ${{ vars.IMAGE_NAMESPACE }}
-          image-name: ${{ vars.NIGHTLY_BUILD_IMAGE_NAME }}
+          image-name: ${{ env.IMAGE_NAME }}
          image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
          image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
          version: ${{ needs.allocate-runners.outputs.version }}
-          push-latest-tag: false
+          push-latest-tag: true

      - name: Set nightly build result
        id: set-nightly-build-result
@@ -214,7 +217,7 @@ jobs:
      allocate-runners,
      release-images-to-dockerhub,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    # When we push to ACR, it's easy to fail due to some unknown network issues.
    # However, we don't want to fail the whole workflow because of this.
    # The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
@@ -223,14 +226,13 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Release artifacts to CN region
        uses: ./.github/actions/release-cn-artifacts
        with:
          src-image-registry: docker.io
          src-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
-          src-image-name: ${{ vars.NIGHTLY_BUILD_IMAGE_NAME }}
+          src-image-name: ${{ env.IMAGE_NAME }}
          dst-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
          dst-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
          dst-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
@@ -240,16 +242,15 @@ jobs:
          aws-cn-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
          aws-cn-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
          aws-cn-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
-          upload-to-s3: false
          dev-mode: false
          update-version-info: false  # Don't update version info in S3.
-          push-latest-tag: false
+          push-latest-tag: true

  stop-linux-amd64-runner: # It's always run as the last job in the workflow to make sure that the runner is released.
    name: Stop linux-amd64 runner
    # Only run this job when the runner is allocated.
    if: ${{ always() }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      allocate-runners,
      build-linux-amd64-artifacts,
@@ -259,7 +260,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Stop EC2 runner
        uses: ./.github/actions/stop-runner
@@ -275,7 +275,7 @@ jobs:
    name: Stop linux-arm64 runner
    # Only run this job when the runner is allocated.
    if: ${{ always() }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      allocate-runners,
      build-linux-arm64-artifacts,
@@ -285,7 +285,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Stop EC2 runner
        uses: ./.github/actions/stop-runner
@@ -303,15 +302,11 @@ jobs:
    needs: [
      release-images-to-dockerhub
    ]
-    runs-on: ubuntu-latest
-    permissions:
-      issues: write
+    runs-on: ubuntu-20.04
    env:
      SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Report CI status
        id: report-ci-status
--- a/.github/workflows/nightly-ci.yml
+++ b/.github/workflows/nightly-ci.yml
@@ -9,17 +9,19 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
  cancel-in-progress: true

+permissions:
+  issues: write
+
 jobs:
  sqlness-test:
    name: Run sqlness test
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Check install.sh
        run: ./.github/scripts/check-install-script.sh
@@ -44,14 +46,9 @@ jobs:
    name: Sqlness tests on Windows
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
    runs-on: windows-2022-8-cores
-    permissions:
-      issues: write
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - uses: arduino/setup-protoc@v3
        with:
@@ -79,9 +76,6 @@ jobs:
    steps:
      - run: git config --global core.autocrlf false
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - uses: arduino/setup-protoc@v3
        with:
@@ -117,13 +111,9 @@ jobs:
  cleanbuild-linux-nix:
    name: Run clean build on Linux
    runs-on: ubuntu-latest
-    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: cachix/install-nix-action@v27
        with:
          nix_path: nixpkgs=channel:nixos-24.11
@@ -133,7 +123,7 @@ jobs:
    name: Check status
    needs: [sqlness-test, sqlness-windows, test-on-windows]
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    outputs:
      check-result: ${{ steps.set-check-result.outputs.check-result }}
    steps:
@@ -146,14 +136,11 @@ jobs:
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && always() }} # Not requiring successful dependent jobs, always run.
    name: Send notification to Greptime team
    needs: [check-status]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    env:
      SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
    steps:
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Report CI status
        id: report-ci-status
--- a/.github/workflows/release-dev-builder-images.yaml
+++ b/.github/workflows/release-dev-builder-images.yaml
@@ -29,7 +29,7 @@ jobs:
  release-dev-builder-images:
    name: Release dev builder images
    if: ${{ inputs.release_dev_builder_ubuntu_image || inputs.release_dev_builder_centos_image || inputs.release_dev_builder_android_image }} # Only manually trigger this job.
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04-16-cores
    outputs:
      version: ${{ steps.set-version.outputs.version }}
    steps:
@@ -37,7 +37,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Configure build image version
        id: set-version
@@ -63,7 +62,7 @@ jobs:

  release-dev-builder-images-ecr:
    name: Release dev builder images to AWS ECR
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      release-dev-builder-images
    ]
@@ -86,69 +85,51 @@ jobs:
      - name: Push dev-builder-ubuntu image
        shell: bash
        if: ${{ inputs.release_dev_builder_ubuntu_image }}
-        env:
-          IMAGE_VERSION: ${{ needs.release-dev-builder-images.outputs.version }}
-          IMAGE_NAMESPACE: ${{ vars.IMAGE_NAMESPACE }}
-          ECR_IMAGE_REGISTRY: ${{ vars.ECR_IMAGE_REGISTRY }}
-          ECR_IMAGE_NAMESPACE: ${{ vars.ECR_IMAGE_NAMESPACE }}
        run: |
          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-ubuntu:$IMAGE_VERSION \
-            docker://$ECR_IMAGE_REGISTRY/$ECR_IMAGE_NAMESPACE/dev-builder-ubuntu:$IMAGE_VERSION
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ needs.release-dev-builder-images.outputs.version }} \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ needs.release-dev-builder-images.outputs.version }}

          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-ubuntu:latest \
-            docker://$ECR_IMAGE_REGISTRY/$ECR_IMAGE_NAMESPACE/dev-builder-ubuntu:latest
-
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:latest \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-ubuntu:latest
      - name: Push dev-builder-centos image
        shell: bash
        if: ${{ inputs.release_dev_builder_centos_image }}
-        env:
-          IMAGE_VERSION: ${{ needs.release-dev-builder-images.outputs.version }}
-          IMAGE_NAMESPACE: ${{ vars.IMAGE_NAMESPACE }}
-          ECR_IMAGE_REGISTRY: ${{ vars.ECR_IMAGE_REGISTRY }}
-          ECR_IMAGE_NAMESPACE: ${{ vars.ECR_IMAGE_NAMESPACE }}
        run: |
          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-centos:$IMAGE_VERSION \
-            docker://$ECR_IMAGE_REGISTRY/$ECR_IMAGE_NAMESPACE/dev-builder-centos:$IMAGE_VERSION
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ needs.release-dev-builder-images.outputs.version }} \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-centos:${{ needs.release-dev-builder-images.outputs.version }}

          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-centos:latest \
-            docker://$ECR_IMAGE_REGISTRY/$ECR_IMAGE_NAMESPACE/dev-builder-centos:latest
-
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:latest \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-centos:latest
      - name: Push dev-builder-android image
        shell: bash
        if: ${{ inputs.release_dev_builder_android_image }}
-        env:
-          IMAGE_VERSION: ${{ needs.release-dev-builder-images.outputs.version }}
-          IMAGE_NAMESPACE: ${{ vars.IMAGE_NAMESPACE }}
-          ECR_IMAGE_REGISTRY: ${{ vars.ECR_IMAGE_REGISTRY }}
-          ECR_IMAGE_NAMESPACE: ${{ vars.ECR_IMAGE_NAMESPACE }}
        run: |
          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-android:$IMAGE_VERSION \
-            docker://$ECR_IMAGE_REGISTRY/$ECR_IMAGE_NAMESPACE/dev-builder-android:$IMAGE_VERSION
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ needs.release-dev-builder-images.outputs.version }} \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-android:${{ needs.release-dev-builder-images.outputs.version }}

          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-android:latest \
-            docker://$ECR_IMAGE_REGISTRY/$ECR_IMAGE_NAMESPACE/dev-builder-android:latest
-
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:latest \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-android:latest
  release-dev-builder-images-cn: # Note: Be careful issue: https://github.com/containers/skopeo/issues/1874 and we decide to use the latest stable skopeo container.
    name: Release dev builder images to CN region
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      release-dev-builder-images
    ]
@@ -163,41 +144,29 @@ jobs:
      - name: Push dev-builder-ubuntu image
        shell: bash
        if: ${{ inputs.release_dev_builder_ubuntu_image }}
-        env:
-          IMAGE_VERSION: ${{ needs.release-dev-builder-images.outputs.version }}
-          IMAGE_NAMESPACE: ${{ vars.IMAGE_NAMESPACE }}
-          ACR_IMAGE_REGISTRY: ${{ vars.ACR_IMAGE_REGISTRY }}
        run: |
          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-ubuntu:$IMAGE_VERSION \
-            docker://$ACR_IMAGE_REGISTRY/$IMAGE_NAMESPACE/dev-builder-ubuntu:$IMAGE_VERSION
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ needs.release-dev-builder-images.outputs.version }} \
+            docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ needs.release-dev-builder-images.outputs.version }}

      - name: Push dev-builder-centos image
        shell: bash
        if: ${{ inputs.release_dev_builder_centos_image }}
-        env:
-          IMAGE_VERSION: ${{ needs.release-dev-builder-images.outputs.version }}
-          IMAGE_NAMESPACE: ${{ vars.IMAGE_NAMESPACE }}
-          ACR_IMAGE_REGISTRY: ${{ vars.ACR_IMAGE_REGISTRY }}
        run: |
          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-centos:$IMAGE_VERSION \
-            docker://$ACR_IMAGE_REGISTRY/$IMAGE_NAMESPACE/dev-builder-centos:$IMAGE_VERSION
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ needs.release-dev-builder-images.outputs.version }} \
+            docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ needs.release-dev-builder-images.outputs.version }}

      - name: Push dev-builder-android image
        shell: bash
        if: ${{ inputs.release_dev_builder_android_image }}
-        env:
-          IMAGE_VERSION: ${{ needs.release-dev-builder-images.outputs.version }}
-          IMAGE_NAMESPACE: ${{ vars.IMAGE_NAMESPACE }}
-          ACR_IMAGE_REGISTRY: ${{ vars.ACR_IMAGE_REGISTRY }}
        run: |
          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
            quay.io/skopeo/stable:latest \
-            copy -a docker://docker.io/$IMAGE_NAMESPACE/dev-builder-android:$IMAGE_VERSION \
-            docker://$ACR_IMAGE_REGISTRY/$IMAGE_NAMESPACE/dev-builder-android:$IMAGE_VERSION
+            copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ needs.release-dev-builder-images.outputs.version }} \
+            docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ needs.release-dev-builder-images.outputs.version }}
--- a/.github/workflows/release.yml
+++ b/.github/workflows/release.yml
@@ -18,11 +18,11 @@ on:
        description: The runner uses to build linux-amd64 artifacts
        default: ec2-c6i.4xlarge-amd64
        options:
-          - ubuntu-22.04
-          - ubuntu-22.04-8-cores
-          - ubuntu-22.04-16-cores
-          - ubuntu-22.04-32-cores
-          - ubuntu-22.04-64-cores
+          - ubuntu-20.04
+          - ubuntu-20.04-8-cores
+          - ubuntu-20.04-16-cores
+          - ubuntu-20.04-32-cores
+          - ubuntu-20.04-64-cores
          - ec2-c6i.xlarge-amd64 # 4C8G
          - ec2-c6i.2xlarge-amd64 # 8C16G
          - ec2-c6i.4xlarge-amd64 # 16C32G
@@ -91,13 +91,18 @@ env:
  # The scheduled version is '${{ 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: v0.13.0
+  NEXT_RELEASE_VERSION: v0.12.0
+
+# Permission reference: https://docs.github.com/en/actions/using-jobs/assigning-permissions-to-jobs
+permissions:
+  issues: write # Allows the action to create issues for cyborg.
+  contents: write # Allows the action to create a release.

 jobs:
  allocate-runners:
    name: Allocate runners
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    outputs:
      linux-amd64-runner: ${{ steps.start-linux-amd64-runner.outputs.label }}
      linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
@@ -117,7 +122,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Check Rust toolchain version
        shell: bash
@@ -177,7 +181,6 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - uses: ./.github/actions/build-linux-artifacts
        with:
@@ -199,7 +202,6 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - uses: ./.github/actions/build-linux-artifacts
        with:
@@ -235,7 +237,6 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - uses: ./.github/actions/build-macos-artifacts
        with:
@@ -275,7 +276,6 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - uses: ./.github/actions/build-windows-artifacts
        with:
@@ -299,25 +299,22 @@ jobs:
      build-linux-amd64-artifacts,
      build-linux-arm64-artifacts,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-2004-16-cores
    outputs:
      build-image-result: ${{ steps.set-build-image-result.outputs.build-image-result }}
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Build and push images to dockerhub
        uses: ./.github/actions/build-images
        with:
          image-registry: docker.io
          image-namespace: ${{ vars.IMAGE_NAMESPACE }}
-          image-name: ${{ vars.GREPTIMEDB_IMAGE_NAME }}
          image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
          image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
          version: ${{ needs.allocate-runners.outputs.version }}
-          push-latest-tag: true

      - name: Set build image result
        id: set-build-image-result
@@ -335,7 +332,7 @@ jobs:
      build-windows-artifacts,
      release-images-to-dockerhub,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    # When we push to ACR, it's easy to fail due to some unknown network issues.
    # However, we don't want to fail the whole workflow because of this.
    # The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
@@ -344,14 +341,13 @@ jobs:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Release artifacts to CN region
        uses: ./.github/actions/release-cn-artifacts
        with:
          src-image-registry: docker.io
          src-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
-          src-image-name: ${{ vars.GREPTIMEDB_IMAGE_NAME }}
+          src-image-name: greptimedb
          dst-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
          dst-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
          dst-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
@@ -362,7 +358,6 @@ jobs:
          aws-cn-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
          aws-cn-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
          dev-mode: false
-          upload-to-s3: true
          update-version-info: true
          push-latest-tag: true

@@ -377,12 +372,11 @@ jobs:
      build-windows-artifacts,
      release-images-to-dockerhub,
    ]
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Publish GitHub release
        uses: ./.github/actions/publish-github-release
@@ -396,7 +390,7 @@ jobs:
    name: Stop linux-amd64 runner
    # Only run this job when the runner is allocated.
    if: ${{ always() }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      allocate-runners,
      build-linux-amd64-artifacts,
@@ -406,7 +400,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Stop EC2 runner
        uses: ./.github/actions/stop-runner
@@ -422,7 +415,7 @@ jobs:
    name: Stop linux-arm64 runner
    # Only run this job when the runner is allocated.
    if: ${{ always() }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    needs: [
      allocate-runners,
      build-linux-arm64-artifacts,
@@ -432,7 +425,6 @@ jobs:
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
-          persist-credentials: false

      - name: Stop EC2 runner
        uses: ./.github/actions/stop-runner
@@ -448,16 +440,9 @@ jobs:
    name: Bump doc version
    if: ${{ github.event_name == 'push' || github.event_name == 'schedule' }}
    needs: [allocate-runners]
-    runs-on: ubuntu-latest
-    # Permission reference: https://docs.github.com/en/actions/using-jobs/assigning-permissions-to-jobs
-    permissions:
-      issues: write # Allows the action to create issues for cyborg.
-      contents: write # Allows the action to create a release.
+    runs-on: ubuntu-20.04
    steps:
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Bump doc version
        working-directory: cyborg
@@ -475,18 +460,11 @@ jobs:
      build-macos-artifacts,
      build-windows-artifacts,
    ]
-    runs-on: ubuntu-latest
-    # Permission reference: https://docs.github.com/en/actions/using-jobs/assigning-permissions-to-jobs
-    permissions:
-      issues: write # Allows the action to create issues for cyborg.
-      contents: write # Allows the action to create a release.
+    runs-on: ubuntu-20.04
    env:
      SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
    steps:
      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Report CI status
        id: report-ci-status
--- a/.github/workflows/schedule.yml
+++ b/.github/workflows/schedule.yml
@@ -4,20 +4,18 @@ on:
    - cron: '4 2 * * *'
  workflow_dispatch:

+permissions:
+  contents: read
+  issues: write
+  pull-requests: write

 jobs:
  maintenance:
    name: Periodic Maintenance
    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      issues: write
-      pull-requests: write
    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Do Maintenance
        working-directory: cyborg
--- a/.github/workflows/semantic-pull-request.yml
+++ b/.github/workflows/semantic-pull-request.yml
@@ -1,24 +1,18 @@
 name: "Semantic Pull Request"

 on:
-  pull_request:
+  pull_request_target:
    types:
      - opened
      - reopened
      - edited

-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
-  cancel-in-progress: true
-
 jobs:
  check:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
    timeout-minutes: 10
    steps:
      - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
      - uses: ./.github/actions/setup-cyborg
      - name: Check Pull Request
        working-directory: cyborg
--- a/AUTHOR.md
+++ b/AUTHOR.md
@@ -3,28 +3,30 @@
 ## Individual Committers (in alphabetical order)

 * [CookiePieWw](https://github.com/CookiePieWw)
+* [KKould](https://github.com/KKould)
+* [NiwakaDev](https://github.com/NiwakaDev)
 * [etolbakov](https://github.com/etolbakov)
 * [irenjj](https://github.com/irenjj)
-* [KKould](https://github.com/KKould)
-* [Lanqing Yang](https://github.com/lyang24)
-* [NiwakaDev](https://github.com/NiwakaDev)
 * [tisonkun](https://github.com/tisonkun)
-
+* [Lanqing Yang](https://github.com/lyang24)

 ## Team Members (in alphabetical order)

+* [Breeze-P](https://github.com/Breeze-P)
+* [GrepTime](https://github.com/GrepTime)
+* [MichaelScofield](https://github.com/MichaelScofield)
+* [Wenjie0329](https://github.com/Wenjie0329)
+* [WenyXu](https://github.com/WenyXu)
+* [ZonaHex](https://github.com/ZonaHex)
 * [apdong2022](https://github.com/apdong2022)
 * [beryl678](https://github.com/beryl678)
-* [Breeze-P](https://github.com/Breeze-P)
 * [daviderli614](https://github.com/daviderli614)
 * [discord9](https://github.com/discord9)
 * [evenyag](https://github.com/evenyag)
 * [fengjiachun](https://github.com/fengjiachun)
 * [fengys1996](https://github.com/fengys1996)
-* [GrepTime](https://github.com/GrepTime)
 * [holalengyu](https://github.com/holalengyu)
 * [killme2008](https://github.com/killme2008)
-* [MichaelScofield](https://github.com/MichaelScofield)
 * [nicecui](https://github.com/nicecui)
 * [paomian](https://github.com/paomian)
 * [shuiyisong](https://github.com/shuiyisong)
@@ -32,14 +34,11 @@
 * [sunng87](https://github.com/sunng87)
 * [v0y4g3r](https://github.com/v0y4g3r)
 * [waynexia](https://github.com/waynexia)
-* [Wenjie0329](https://github.com/Wenjie0329)
-* [WenyXu](https://github.com/WenyXu)
 * [xtang](https://github.com/xtang)
 * [zhaoyingnan01](https://github.com/zhaoyingnan01)
 * [zhongzc](https://github.com/zhongzc)
-* [ZonaHex](https://github.com/ZonaHex)
 * [zyy17](https://github.com/zyy17)

 ## All Contributors

-To see the full list of contributors, please visit our [Contributors page](https://github.com/GreptimeTeam/greptimedb/graphs/contributors)
+[![All Contributors](https://contrib.rocks/image?repo=GreptimeTeam/greptimedb)](https://github.com/GreptimeTeam/greptimedb/graphs/contributors)
--- a/Cargo.lock
+++ b/Cargo.lock
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -67,7 +67,7 @@ members = [
 resolver = "2"

 [workspace.package]
-version = "0.13.0"
+version = "0.12.0"
 edition = "2021"
 license = "Apache-2.0"

@@ -81,7 +81,6 @@ rust.unknown_lints = "deny"
 rust.unexpected_cfgs = { level = "warn", check-cfg = ['cfg(tokio_unstable)'] }

 [workspace.dependencies]
-# DO_NOT_REMOVE_THIS: BEGIN_OF_EXTERNAL_DEPENDENCIES
 # We turn off default-features for some dependencies here so the workspaces which inherit them can
 # selectively turn them on if needed, since we can override default-features = true (from false)
 # for the inherited dependency but cannot do the reverse (override from true to false).
@@ -107,7 +106,6 @@ bitflags = "2.4.1"
 bytemuck = "1.12"
 bytes = { version = "1.7", features = ["serde"] }
 chrono = { version = "0.4", features = ["serde"] }
-chrono-tz = "0.10.1"
 clap = { version = "4.4", features = ["derive"] }
 config = "0.13.0"
 crossbeam-utils = "0.8"
@@ -126,11 +124,11 @@ deadpool-postgres = "0.12"
 derive_builder = "0.12"
 dotenv = "0.15"
 etcd-client = "0.14"
-flate2 = { version = "1.1.0", default-features = false, features = ["zlib-rs"] }
 fst = "0.4.7"
 futures = "0.3"
 futures-util = "0.3"
-greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "c5419bbd20cb42e568ec325a4d71a3c94cc327e1" }
+# branch: poc-write-path
+greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "1915576b113a494f5352fd61f211d899b7f87aab" }
 hex = "0.4"
 http = "1"
 humantime = "2.1"
@@ -141,8 +139,8 @@ itertools = "0.10"
 jsonb = { git = "https://github.com/databendlabs/jsonb.git", rev = "8c8d2fc294a39f3ff08909d60f718639cfba3875", default-features = false }
 lazy_static = "1.4"
 local-ip-address = "0.6"
-loki-proto = { git = "https://github.com/GreptimeTeam/loki-proto.git", rev = "1434ecf23a2654025d86188fb5205e7a74b225d3" }
-meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "5618e779cf2bb4755b499c630fba4c35e91898cb" }
+loki-api = { git = "https://github.com/shuiyisong/tracing-loki", branch = "chore/prost_version" }
+meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "a10facb353b41460eeb98578868ebf19c2084fac" }
 mockall = "0.11.4"
 moka = "0.12"
 nalgebra = "0.33"
@@ -161,7 +159,7 @@ parquet = { version = "53.0.0", default-features = false, features = ["arrow", "
 paste = "1.0"
 pin-project = "1.0"
 prometheus = { version = "0.13.3", features = ["process"] }
-promql-parser = { version = "0.5", features = ["ser"] }
+promql-parser = { version = "0.4.3", features = ["ser"] }
 prost = "0.13"
 raft-engine = { version = "0.4.1", default-features = false }
 rand = "0.8"
@@ -189,10 +187,6 @@ shadow-rs = "0.38"
 similar-asserts = "1.6.0"
 smallvec = { version = "1", features = ["serde"] }
 snafu = "0.8"
-sqlx = { version = "0.8", features = [
-    "runtime-tokio-rustls",
-    "mysql",
-] }
 sysinfo = "0.30"
 # on branch v0.52.x
 sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "71dd86058d2af97b9925093d40c4e03360403170", features = [
@@ -214,7 +208,6 @@ tracing-subscriber = { version = "0.3", features = ["env-filter", "json", "fmt"]
 typetag = "0.2"
 uuid = { version = "1.7", features = ["serde", "v4", "fast-rng"] }
 zstd = "0.13"
-# DO_NOT_REMOVE_THIS: END_OF_EXTERNAL_DEPENDENCIES

 ## workspaces members
 api = { path = "src/api" }
@@ -286,10 +279,12 @@ tokio-rustls = { git = "https://github.com/GreptimeTeam/tokio-rustls", rev = "46
 # This is commented, since we are not using aws-lc-sys, if we need to use it, we need to uncomment this line or use a release after this commit, or it wouldn't compile with gcc < 8.1
 # see https://github.com/aws/aws-lc-rs/pull/526
 # aws-lc-sys = { git ="https://github.com/aws/aws-lc-rs", rev = "556558441e3494af4b156ae95ebc07ebc2fd38aa" }
+# Apply a fix for pprof for unaligned pointer access
+pprof = { git = "https://github.com/GreptimeTeam/pprof-rs", rev = "1bd1e21" }

 [workspace.dependencies.meter-macros]
 git = "https://github.com/GreptimeTeam/greptime-meter.git"
-rev = "5618e779cf2bb4755b499c630fba4c35e91898cb"
+rev = "a10facb353b41460eeb98578868ebf19c2084fac"

 [profile.release]
 debug = 1
--- a/Cross.toml
+++ b/Cross.toml
@@ -1,6 +1,3 @@
-[target.aarch64-unknown-linux-gnu]
-image = "ghcr.io/cross-rs/aarch64-unknown-linux-gnu:0.2.5"
-
 [build]
 pre-build = [
    "dpkg --add-architecture $CROSS_DEB_ARCH",
@@ -8,8 +5,3 @@ pre-build = [
    "curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v3.15.8/protoc-3.15.8-linux-x86_64.zip && unzip protoc-3.15.8-linux-x86_64.zip -d /usr/",
    "chmod a+x /usr/bin/protoc && chmod -R a+rx /usr/include/google",
 ]
-
-[build.env]
-passthrough = [
-    "JEMALLOC_SYS_WITH_LG_PAGE",
-]
--- a/4
+++ b/4
@@ -8,7 +8,7 @@ CARGO_BUILD_OPTS := --locked
 IMAGE_REGISTRY ?= docker.io
 IMAGE_NAMESPACE ?= greptime
 IMAGE_TAG ?= latest
-DEV_BUILDER_IMAGE_TAG ?= 2024-12-25-a71b93dd-20250305072908
+DEV_BUILDER_IMAGE_TAG ?= 2024-12-25-9d0fa5d5-20250124085746
 BUILDX_MULTI_PLATFORM_BUILD ?= false
 BUILDX_BUILDER_NAME ?= gtbuilder
 BASE_IMAGE ?= ubuntu
@@ -60,8 +60,6 @@ ifeq ($(BUILDX_MULTI_PLATFORM_BUILD), all)
 	BUILDX_MULTI_PLATFORM_BUILD_OPTS := --platform linux/amd64,linux/arm64 --push
 else ifeq ($(BUILDX_MULTI_PLATFORM_BUILD), amd64)
 	BUILDX_MULTI_PLATFORM_BUILD_OPTS := --platform linux/amd64 --push
-else ifeq ($(BUILDX_MULTI_PLATFORM_BUILD), arm64)
-	BUILDX_MULTI_PLATFORM_BUILD_OPTS := --platform linux/arm64 --push
 else
 	BUILDX_MULTI_PLATFORM_BUILD_OPTS := -o type=docker
 endif
--- a/README.md
+++ b/README.md
@@ -116,7 +116,7 @@ docker run -p 127.0.0.1:4000-4003:4000-4003 \
  --name greptime --rm \
  greptime/greptimedb:latest standalone start \
  --http-addr 0.0.0.0:4000 \
-  --rpc-bind-addr 0.0.0.0:4001 \
+  --rpc-addr 0.0.0.0:4001 \
  --mysql-addr 0.0.0.0:4002 \
  --postgres-addr 0.0.0.0:4003
 ```
--- a/config/config.md
+++ b/config/config.md
@@ -29,7 +29,7 @@
 | `http.enable_cors` | Bool | `true` | HTTP CORS support, it's turned on by default<br/>This allows browser to access http APIs without CORS restrictions |
 | `http.cors_allowed_origins` | Array | Unset | Customize allowed origins for HTTP CORS. |
 | `grpc` | -- | -- | The gRPC server options. |
-| `grpc.bind_addr` | String | `127.0.0.1:4001` | The address to bind the gRPC server. |
+| `grpc.addr` | String | `127.0.0.1:4001` | The address to bind the gRPC server. |
 | `grpc.runtime_size` | Integer | `8` | The number of server worker threads. |
 | `grpc.tls` | -- | -- | gRPC server TLS options, see `mysql.tls` section. |
 | `grpc.tls.mode` | String | `disable` | TLS mode. |
@@ -40,7 +40,6 @@
 | `mysql.enable` | Bool | `true` | Whether to enable. |
 | `mysql.addr` | String | `127.0.0.1:4002` | The addr to bind the MySQL server. |
 | `mysql.runtime_size` | Integer | `2` | The number of server worker threads. |
-| `mysql.keep_alive` | String | `0s` | Server-side keep-alive time.<br/>Set to 0 (default) to disable. |
 | `mysql.tls` | -- | -- | -- |
 | `mysql.tls.mode` | String | `disable` | TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html<br/>- `disable` (default value)<br/>- `prefer`<br/>- `require`<br/>- `verify-ca`<br/>- `verify-full` |
 | `mysql.tls.cert_path` | String | Unset | Certificate file path. |
@@ -50,7 +49,6 @@
 | `postgres.enable` | Bool | `true` | Whether to enable |
 | `postgres.addr` | String | `127.0.0.1:4003` | The addr to bind the PostgresSQL server. |
 | `postgres.runtime_size` | Integer | `2` | The number of server worker threads. |
-| `postgres.keep_alive` | String | `0s` | Server-side keep-alive time.<br/>Set to 0 (default) to disable. |
 | `postgres.tls` | -- | -- | PostgresSQL server TLS options, see `mysql.tls` section. |
 | `postgres.tls.mode` | String | `disable` | TLS mode. |
 | `postgres.tls.cert_path` | String | Unset | Certificate file path. |
@@ -60,8 +58,6 @@
 | `opentsdb.enable` | Bool | `true` | Whether to enable OpenTSDB put in HTTP API. |
 | `influxdb` | -- | -- | InfluxDB protocol options. |
 | `influxdb.enable` | Bool | `true` | Whether to enable InfluxDB protocol in HTTP API. |
-| `jaeger` | -- | -- | Jaeger protocol options. |
-| `jaeger.enable` | Bool | `true` | Whether to enable Jaeger protocol in HTTP API. |
 | `prom_store` | -- | -- | Prometheus remote storage options |
 | `prom_store.enable` | Bool | `true` | Whether to enable Prometheus remote write and read in HTTP API. |
 | `prom_store.with_metric_engine` | Bool | `true` | Whether to store the data from Prometheus remote write in metric engine. |
@@ -69,8 +65,8 @@
 | `wal.provider` | String | `raft_engine` | The provider of the WAL.<br/>- `raft_engine`: the wal is stored in the local file system by raft-engine.<br/>- `kafka`: it's remote wal that data is stored in Kafka. |
 | `wal.dir` | String | Unset | The directory to store the WAL files.<br/>**It's only used when the provider is `raft_engine`**. |
 | `wal.file_size` | String | `128MB` | The size of the WAL segment file.<br/>**It's only used when the provider is `raft_engine`**. |
-| `wal.purge_threshold` | String | `1GB` | The threshold of the WAL size to trigger a purge.<br/>**It's only used when the provider is `raft_engine`**. |
-| `wal.purge_interval` | String | `1m` | The interval to trigger a purge.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.purge_threshold` | String | `1GB` | The threshold of the WAL size to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
+| `wal.purge_interval` | String | `1m` | The interval to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
 | `wal.read_batch_size` | Integer | `128` | The read batch size.<br/>**It's only used when the provider is `raft_engine`**. |
 | `wal.sync_write` | Bool | `false` | Whether to use sync write.<br/>**It's only used when the provider is `raft_engine`**. |
 | `wal.enable_log_recycle` | Bool | `true` | Whether to reuse logically truncated log files.<br/>**It's only used when the provider is `raft_engine`**. |
@@ -92,9 +88,8 @@
 | `wal.backoff_deadline` | String | `5mins` | The deadline of retries.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.overwrite_entry_start_id` | Bool | `false` | Ignore missing entries during read WAL.<br/>**It's only used when the provider is `kafka`**.<br/><br/>This option ensures that when Kafka messages are deleted, the system<br/>can still successfully replay memtable data without throwing an<br/>out-of-range error.<br/>However, enabling this option might lead to unexpected data loss,<br/>as the system will skip over missing entries instead of treating<br/>them as critical errors. |
 | `metadata_store` | -- | -- | Metadata storage options. |
-| `metadata_store.file_size` | String | `64MB` | The size of the metadata store log file. |
-| `metadata_store.purge_threshold` | String | `256MB` | The threshold of the metadata store size to trigger a purge. |
-| `metadata_store.purge_interval` | String | `1m` | The interval of the metadata store to trigger a purge. |
+| `metadata_store.file_size` | String | `256MB` | Kv file size in bytes. |
+| `metadata_store.purge_threshold` | String | `4GB` | Kv purge threshold. |
 | `procedure` | -- | -- | Procedure storage options. |
 | `procedure.max_retry_times` | Integer | `3` | Procedure max retry time. |
 | `procedure.retry_delay` | String | `500ms` | Initial retry delay of procedures, increases exponentially |
@@ -152,7 +147,6 @@
 | `region_engine.mito.index` | -- | -- | The options for index in Mito engine. |
 | `region_engine.mito.index.aux_path` | String | `""` | Auxiliary directory path for the index in filesystem, used to store intermediate files for<br/>creating the index and staging files for searching the index, defaults to `{data_home}/index_intermediate`.<br/>The default name for this directory is `index_intermediate` for backward compatibility.<br/><br/>This path contains two subdirectories:<br/>- `__intm`: for storing intermediate files used during creating index.<br/>- `staging`: for storing staging files used during searching index. |
 | `region_engine.mito.index.staging_size` | String | `2GB` | The max capacity of the staging directory. |
-| `region_engine.mito.index.staging_ttl` | String | `7d` | The TTL of the staging directory.<br/>Defaults to 7 days.<br/>Setting it to "0s" to disable TTL. |
 | `region_engine.mito.index.metadata_cache_size` | String | `64MiB` | Cache size for inverted index metadata. |
 | `region_engine.mito.index.content_cache_size` | String | `128MiB` | Cache size for inverted index content. |
 | `region_engine.mito.index.content_cache_page_size` | String | `64KiB` | Page size for inverted index content cache. |
@@ -227,8 +221,8 @@
 | `http.enable_cors` | Bool | `true` | HTTP CORS support, it's turned on by default<br/>This allows browser to access http APIs without CORS restrictions |
 | `http.cors_allowed_origins` | Array | Unset | Customize allowed origins for HTTP CORS. |
 | `grpc` | -- | -- | The gRPC server options. |
-| `grpc.bind_addr` | String | `127.0.0.1:4001` | The address to bind the gRPC server. |
-| `grpc.server_addr` | String | `127.0.0.1:4001` | The address advertised to the metasrv, and used for connections from outside the host.<br/>If left empty or unset, the server will automatically use the IP address of the first network interface<br/>on the host, with the same port number as the one specified in `grpc.bind_addr`. |
+| `grpc.addr` | String | `127.0.0.1:4001` | The address to bind the gRPC server. |
+| `grpc.hostname` | String | `127.0.0.1:4001` | The hostname advertised to the metasrv,<br/>and used for connections from outside the host |
 | `grpc.runtime_size` | Integer | `8` | The number of server worker threads. |
 | `grpc.tls` | -- | -- | gRPC server TLS options, see `mysql.tls` section. |
 | `grpc.tls.mode` | String | `disable` | TLS mode. |
@@ -239,7 +233,6 @@
 | `mysql.enable` | Bool | `true` | Whether to enable. |
 | `mysql.addr` | String | `127.0.0.1:4002` | The addr to bind the MySQL server. |
 | `mysql.runtime_size` | Integer | `2` | The number of server worker threads. |
-| `mysql.keep_alive` | String | `0s` | Server-side keep-alive time.<br/>Set to 0 (default) to disable. |
 | `mysql.tls` | -- | -- | -- |
 | `mysql.tls.mode` | String | `disable` | TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html<br/>- `disable` (default value)<br/>- `prefer`<br/>- `require`<br/>- `verify-ca`<br/>- `verify-full` |
 | `mysql.tls.cert_path` | String | Unset | Certificate file path. |
@@ -249,7 +242,6 @@
 | `postgres.enable` | Bool | `true` | Whether to enable |
 | `postgres.addr` | String | `127.0.0.1:4003` | The addr to bind the PostgresSQL server. |
 | `postgres.runtime_size` | Integer | `2` | The number of server worker threads. |
-| `postgres.keep_alive` | String | `0s` | Server-side keep-alive time.<br/>Set to 0 (default) to disable. |
 | `postgres.tls` | -- | -- | PostgresSQL server TLS options, see `mysql.tls` section. |
 | `postgres.tls.mode` | String | `disable` | TLS mode. |
 | `postgres.tls.cert_path` | String | Unset | Certificate file path. |
@@ -259,8 +251,6 @@
 | `opentsdb.enable` | Bool | `true` | Whether to enable OpenTSDB put in HTTP API. |
 | `influxdb` | -- | -- | InfluxDB protocol options. |
 | `influxdb.enable` | Bool | `true` | Whether to enable InfluxDB protocol in HTTP API. |
-| `jaeger` | -- | -- | Jaeger protocol options. |
-| `jaeger.enable` | Bool | `true` | Whether to enable Jaeger protocol in HTTP API. |
 | `prom_store` | -- | -- | Prometheus remote storage options |
 | `prom_store.enable` | Bool | `true` | Whether to enable Prometheus remote write and read in HTTP API. |
 | `prom_store.with_metric_engine` | Bool | `true` | Whether to store the data from Prometheus remote write in metric engine. |
@@ -310,7 +300,7 @@
 | --- | -----| ------- | ----------- |
 | `data_home` | String | `/tmp/metasrv/` | The working home directory. |
 | `bind_addr` | String | `127.0.0.1:3002` | The bind address of metasrv. |
-| `server_addr` | String | `127.0.0.1:3002` | The communication server address for the frontend and datanode to connect to metasrv.<br/>If left empty or unset, the server will automatically use the IP address of the first network interface<br/>on the host, with the same port number as the one specified in `bind_addr`. |
+| `server_addr` | String | `127.0.0.1:3002` | The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost. |
 | `store_addrs` | Array | -- | Store server address default to etcd store.<br/>For postgres store, the format is:<br/>"password=password dbname=postgres user=postgres host=localhost port=5432"<br/>For etcd store, the format is:<br/>"127.0.0.1:2379" |
 | `store_key_prefix` | String | `""` | If it's not empty, the metasrv will store all data with this key prefix. |
 | `backend` | String | `etcd_store` | The datastore for meta server.<br/>Available values:<br/>- `etcd_store` (default value)<br/>- `memory_store`<br/>- `postgres_store` |
@@ -319,7 +309,6 @@
 | `selector` | String | `round_robin` | Datanode selector type.<br/>- `round_robin` (default value)<br/>- `lease_based`<br/>- `load_based`<br/>For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector". |
 | `use_memory_store` | Bool | `false` | Store data in memory. |
 | `enable_region_failover` | Bool | `false` | Whether to enable region failover.<br/>This feature is only available on GreptimeDB running on cluster mode and<br/>- Using Remote WAL<br/>- Using shared storage (e.g., s3). |
-| `node_max_idle_time` | String | `24hours` | Max allowed idle time before removing node info from metasrv memory. |
 | `enable_telemetry` | Bool | `true` | Whether to enable greptimedb telemetry. Enabled by default. |
 | `runtime` | -- | -- | The runtime options. |
 | `runtime.global_rt_size` | Integer | `8` | The number of threads to execute the runtime for global read operations. |
@@ -387,14 +376,19 @@
 | `init_regions_in_background` | Bool | `false` | Initialize all regions in the background during the startup.<br/>By default, it provides services after all regions have been initialized. |
 | `init_regions_parallelism` | Integer | `16` | Parallelism of initializing regions. |
 | `max_concurrent_queries` | Integer | `0` | The maximum current queries allowed to be executed. Zero means unlimited. |
+| `rpc_addr` | String | Unset | Deprecated, use `grpc.addr` instead. |
+| `rpc_hostname` | String | Unset | Deprecated, use `grpc.hostname` instead. |
+| `rpc_runtime_size` | Integer | Unset | Deprecated, use `grpc.runtime_size` instead. |
+| `rpc_max_recv_message_size` | String | Unset | Deprecated, use `grpc.rpc_max_recv_message_size` instead. |
+| `rpc_max_send_message_size` | String | Unset | Deprecated, use `grpc.rpc_max_send_message_size` instead. |
 | `enable_telemetry` | Bool | `true` | Enable telemetry to collect anonymous usage data. Enabled by default. |
 | `http` | -- | -- | The HTTP server options. |
 | `http.addr` | String | `127.0.0.1:4000` | The address to bind the HTTP server. |
 | `http.timeout` | String | `30s` | HTTP request timeout. Set to 0 to disable timeout. |
 | `http.body_limit` | String | `64MB` | HTTP request body limit.<br/>The following units are supported: `B`, `KB`, `KiB`, `MB`, `MiB`, `GB`, `GiB`, `TB`, `TiB`, `PB`, `PiB`.<br/>Set to 0 to disable limit. |
 | `grpc` | -- | -- | The gRPC server options. |
-| `grpc.bind_addr` | String | `127.0.0.1:3001` | The address to bind the gRPC server. |
-| `grpc.server_addr` | String | `127.0.0.1:3001` | The address advertised to the metasrv, and used for connections from outside the host.<br/>If left empty or unset, the server will automatically use the IP address of the first network interface<br/>on the host, with the same port number as the one specified in `grpc.bind_addr`. |
+| `grpc.addr` | String | `127.0.0.1:3001` | The address to bind the gRPC server. |
+| `grpc.hostname` | String | `127.0.0.1:3001` | The hostname advertised to the metasrv,<br/>and used for connections from outside the host |
 | `grpc.runtime_size` | Integer | `8` | The number of server worker threads. |
 | `grpc.max_recv_message_size` | String | `512MB` | The maximum receive message size for gRPC server. |
 | `grpc.max_send_message_size` | String | `512MB` | The maximum send message size for gRPC server. |
@@ -493,7 +487,6 @@
 | `region_engine.mito.index` | -- | -- | The options for index in Mito engine. |
 | `region_engine.mito.index.aux_path` | String | `""` | Auxiliary directory path for the index in filesystem, used to store intermediate files for<br/>creating the index and staging files for searching the index, defaults to `{data_home}/index_intermediate`.<br/>The default name for this directory is `index_intermediate` for backward compatibility.<br/><br/>This path contains two subdirectories:<br/>- `__intm`: for storing intermediate files used during creating index.<br/>- `staging`: for storing staging files used during searching index. |
 | `region_engine.mito.index.staging_size` | String | `2GB` | The max capacity of the staging directory. |
-| `region_engine.mito.index.staging_ttl` | String | `7d` | The TTL of the staging directory.<br/>Defaults to 7 days.<br/>Setting it to "0s" to disable TTL. |
 | `region_engine.mito.index.metadata_cache_size` | String | `64MiB` | Cache size for inverted index metadata. |
 | `region_engine.mito.index.content_cache_size` | String | `128MiB` | Cache size for inverted index content. |
 | `region_engine.mito.index.content_cache_page_size` | String | `64KiB` | Page size for inverted index content cache. |
@@ -556,8 +549,8 @@
 | `flow` | -- | -- | flow engine options. |
 | `flow.num_workers` | Integer | `0` | The number of flow worker in flownode.<br/>Not setting(or set to 0) this value will use the number of CPU cores divided by 2. |
 | `grpc` | -- | -- | The gRPC server options. |
-| `grpc.bind_addr` | String | `127.0.0.1:6800` | The address to bind the gRPC server. |
-| `grpc.server_addr` | String | `127.0.0.1:6800` | The address advertised to the metasrv,<br/>and used for connections from outside the host |
+| `grpc.addr` | String | `127.0.0.1:6800` | The address to bind the gRPC server. |
+| `grpc.hostname` | String | `127.0.0.1` | The hostname advertised to the metasrv,<br/>and used for connections from outside the host |
 | `grpc.runtime_size` | Integer | `2` | The number of server worker threads. |
 | `grpc.max_recv_message_size` | String | `512MB` | The maximum receive message size for gRPC server. |
 | `grpc.max_send_message_size` | String | `512MB` | The maximum send message size for gRPC server. |
--- a/config/datanode.example.toml
+++ b/config/datanode.example.toml
@@ -19,6 +19,26 @@ init_regions_parallelism = 16
 ## The maximum current queries allowed to be executed. Zero means unlimited.
 max_concurrent_queries = 0

+## Deprecated, use `grpc.addr` instead.
+## @toml2docs:none-default
+rpc_addr = "127.0.0.1:3001"
+
+## Deprecated, use `grpc.hostname` instead.
+## @toml2docs:none-default
+rpc_hostname = "127.0.0.1"
+
+## Deprecated, use `grpc.runtime_size` instead.
+## @toml2docs:none-default
+rpc_runtime_size = 8
+
+## Deprecated, use `grpc.rpc_max_recv_message_size` instead.
+## @toml2docs:none-default
+rpc_max_recv_message_size = "512MB"
+
+## Deprecated, use `grpc.rpc_max_send_message_size` instead.
+## @toml2docs:none-default
+rpc_max_send_message_size = "512MB"
+
 ## Enable telemetry to collect anonymous usage data. Enabled by default.
 #+ enable_telemetry = true

@@ -36,11 +56,10 @@ body_limit = "64MB"
 ## The gRPC server options.
 [grpc]
 ## The address to bind the gRPC server.
-bind_addr = "127.0.0.1:3001"
-## The address advertised to the metasrv, and used for connections from outside the host.
-## If left empty or unset, the server will automatically use the IP address of the first network interface
-## on the host, with the same port number as the one specified in `grpc.bind_addr`.
-server_addr = "127.0.0.1:3001"
+addr = "127.0.0.1:3001"
+## The hostname advertised to the metasrv,
+## and used for connections from outside the host
+hostname = "127.0.0.1:3001"
 ## The number of server worker threads.
 runtime_size = 8
 ## The maximum receive message size for gRPC server.
@@ -231,7 +250,6 @@ overwrite_entry_start_id = false
 # secret_access_key = "123456"
 # endpoint = "https://s3.amazonaws.com"
 # region = "us-west-2"
-# enable_virtual_host_style = false

 # Example of using Oss as the storage.
 # [storage]
@@ -498,11 +516,6 @@ aux_path = ""
 ## The max capacity of the staging directory.
 staging_size = "2GB"

-## The TTL of the staging directory.
-## Defaults to 7 days.
-## Setting it to "0s" to disable TTL.
-staging_ttl = "7d"
-
 ## Cache size for inverted index metadata.
 metadata_cache_size = "64MiB"

--- a/config/flownode.example.toml
+++ b/config/flownode.example.toml
@@ -14,10 +14,10 @@ node_id = 14
 ## The gRPC server options.
 [grpc]
 ## The address to bind the gRPC server.
-bind_addr = "127.0.0.1:6800"
-## The address advertised to the metasrv,
+addr = "127.0.0.1:6800"
+## The hostname advertised to the metasrv,
 ## and used for connections from outside the host
-server_addr = "127.0.0.1:6800"
+hostname = "127.0.0.1"
 ## The number of server worker threads.
 runtime_size = 2
 ## The maximum receive message size for gRPC server.
--- a/config/frontend.example.toml
+++ b/config/frontend.example.toml
@@ -41,11 +41,10 @@ cors_allowed_origins = ["https://example.com"]
 ## The gRPC server options.
 [grpc]
 ## The address to bind the gRPC server.
-bind_addr = "127.0.0.1:4001"
-## The address advertised to the metasrv, and used for connections from outside the host.
-## If left empty or unset, the server will automatically use the IP address of the first network interface
-## on the host, with the same port number as the one specified in `grpc.bind_addr`.
-server_addr = "127.0.0.1:4001"
+addr = "127.0.0.1:4001"
+## The hostname advertised to the metasrv,
+## and used for connections from outside the host
+hostname = "127.0.0.1:4001"
 ## The number of server worker threads.
 runtime_size = 8

@@ -74,9 +73,6 @@ enable = true
 addr = "127.0.0.1:4002"
 ## The number of server worker threads.
 runtime_size = 2
-## Server-side keep-alive time.
-## Set to 0 (default) to disable.
-keep_alive = "0s"

 # MySQL server TLS options.
 [mysql.tls]
@@ -108,9 +104,6 @@ enable = true
 addr = "127.0.0.1:4003"
 ## The number of server worker threads.
 runtime_size = 2
-## Server-side keep-alive time.
-## Set to 0 (default) to disable.
-keep_alive = "0s"

 ## PostgresSQL server TLS options, see `mysql.tls` section.
 [postgres.tls]
@@ -138,11 +131,6 @@ enable = true
 ## Whether to enable InfluxDB protocol in HTTP API.
 enable = true

-## Jaeger protocol options.
-[jaeger]
-## Whether to enable Jaeger protocol in HTTP API.
-enable = true
-
 ## Prometheus remote storage options
 [prom_store]
 ## Whether to enable Prometheus remote write and read in HTTP API.
--- a/config/metasrv.example.toml
+++ b/config/metasrv.example.toml
@@ -4,9 +4,7 @@ data_home = "/tmp/metasrv/"
 ## The bind address of metasrv.
 bind_addr = "127.0.0.1:3002"

-## The communication server address for the frontend and datanode to connect to metasrv.
-## If left empty or unset, the server will automatically use the IP address of the first network interface
-## on the host, with the same port number as the one specified in `bind_addr`.
+## The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost.
 server_addr = "127.0.0.1:3002"

 ## Store server address default to etcd store.
@@ -50,9 +48,6 @@ use_memory_store = false
 ## - Using shared storage (e.g., s3).
 enable_region_failover = false

-## Max allowed idle time before removing node info from metasrv memory.
-node_max_idle_time = "24hours"
-
 ## Whether to enable greptimedb telemetry. Enabled by default.
 #+ enable_telemetry = true

--- a/config/standalone.example.toml
+++ b/config/standalone.example.toml
@@ -49,7 +49,7 @@ cors_allowed_origins = ["https://example.com"]
 ## The gRPC server options.
 [grpc]
 ## The address to bind the gRPC server.
-bind_addr = "127.0.0.1:4001"
+addr = "127.0.0.1:4001"
 ## The number of server worker threads.
 runtime_size = 8

@@ -78,9 +78,6 @@ enable = true
 addr = "127.0.0.1:4002"
 ## The number of server worker threads.
 runtime_size = 2
-## Server-side keep-alive time.
-## Set to 0 (default) to disable.
-keep_alive = "0s"

 # MySQL server TLS options.
 [mysql.tls]
@@ -112,9 +109,6 @@ enable = true
 addr = "127.0.0.1:4003"
 ## The number of server worker threads.
 runtime_size = 2
-## Server-side keep-alive time.
-## Set to 0 (default) to disable.
-keep_alive = "0s"

 ## PostgresSQL server TLS options, see `mysql.tls` section.
 [postgres.tls]
@@ -142,11 +136,6 @@ enable = true
 ## Whether to enable InfluxDB protocol in HTTP API.
 enable = true

-## Jaeger protocol options.
-[jaeger]
-## Whether to enable Jaeger protocol in HTTP API.
-enable = true
-
 ## Prometheus remote storage options
 [prom_store]
 ## Whether to enable Prometheus remote write and read in HTTP API.
@@ -170,11 +159,11 @@ dir = "/tmp/greptimedb/wal"
 ## **It's only used when the provider is `raft_engine`**.
 file_size = "128MB"

-## The threshold of the WAL size to trigger a purge.
+## The threshold of the WAL size to trigger a flush.
 ## **It's only used when the provider is `raft_engine`**.
 purge_threshold = "1GB"

-## The interval to trigger a purge.
+## The interval to trigger a flush.
 ## **It's only used when the provider is `raft_engine`**.
 purge_interval = "1m"

@@ -289,12 +278,10 @@ overwrite_entry_start_id = false

 ## Metadata storage options.
 [metadata_store]
-## The size of the metadata store log file.
-file_size = "64MB"
-## The threshold of the metadata store size to trigger a purge.
-purge_threshold = "256MB"
-## The interval of the metadata store to trigger a purge.
-purge_interval = "1m"
+## Kv file size in bytes.
+file_size = "256MB"
+## Kv purge threshold.
+purge_threshold = "4GB"

 ## Procedure storage options.
 [procedure]
@@ -318,7 +305,6 @@ retry_delay = "500ms"
 # secret_access_key = "123456"
 # endpoint = "https://s3.amazonaws.com"
 # region = "us-west-2"
-# enable_virtual_host_style = false

 # Example of using Oss as the storage.
 # [storage]
@@ -585,11 +571,6 @@ aux_path = ""
 ## The max capacity of the staging directory.
 staging_size = "2GB"

-## The TTL of the staging directory.
-## Defaults to 7 days.
-## Setting it to "0s" to disable TTL.
-staging_ttl = "7d"
-
 ## Cache size for inverted index metadata.
 metadata_cache_size = "64MiB"

--- a/docker/buildx/ubuntu/Dockerfile
+++ b/docker/buildx/ubuntu/Dockerfile
@@ -1,4 +1,4 @@
-FROM ubuntu:22.04 as builder
+FROM ubuntu:20.04 as builder

 ARG CARGO_PROFILE
 ARG FEATURES
--- a/docker/ci/ubuntu/Dockerfile.fuzztests
+++ b/docker/ci/ubuntu/Dockerfile.fuzztests
@@ -1,4 +1,4 @@
-FROM ubuntu:latest
+FROM ubuntu:22.04

 # The binary name of GreptimeDB executable.
 # Defaults to "greptime", but sometimes in other projects it might be different.
--- a/docker/dev-builder/ubuntu/Dockerfile
+++ b/docker/dev-builder/ubuntu/Dockerfile
@@ -1,4 +1,4 @@
-FROM ubuntu:22.04
+FROM ubuntu:20.04

 # The root path under which contains all the dependencies to build this Dockerfile.
 ARG DOCKER_BUILD_ROOT=.
@@ -41,7 +41,7 @@ RUN mv protoc3/include/* /usr/local/include/
 # 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
+# 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.
--- a/docker/dev-builder/ubuntu/Dockerfile-18.10
+++ b/docker/dev-builder/ubuntu/Dockerfile-18.10
@@ -0,0 +1,51 @@
+# Use the legacy glibc 2.28.
+FROM ubuntu:18.10
+
+ENV LANG en_US.utf8
+WORKDIR /greptimedb
+
+# Use old-releases.ubuntu.com to avoid 404s: https://help.ubuntu.com/community/EOLUpgrades.
+RUN echo "deb http://old-releases.ubuntu.com/ubuntu/ cosmic main restricted universe multiverse\n\
+deb http://old-releases.ubuntu.com/ubuntu/ cosmic-updates main restricted universe multiverse\n\
+deb http://old-releases.ubuntu.com/ubuntu/ cosmic-security main restricted universe multiverse" > /etc/apt/sources.list
+
+# Install dependencies.
+RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
+    libssl-dev \
+    tzdata \
+    curl \
+    ca-certificates \
+    git \
+    build-essential \
+    unzip \
+    pkg-config
+
+# Install protoc.
+ENV PROTOC_VERSION=29.3
+RUN if [ "$(uname -m)" = "x86_64" ]; then \
+        PROTOC_ZIP=protoc-${PROTOC_VERSION}-linux-x86_64.zip; \
+    elif [ "$(uname -m)" = "aarch64" ]; then \
+        PROTOC_ZIP=protoc-${PROTOC_VERSION}-linux-aarch_64.zip; \
+    else \
+        echo "Unsupported architecture"; exit 1; \
+    fi && \
+    curl -OL https://github.com/protocolbuffers/protobuf/releases/download/v${PROTOC_VERSION}/${PROTOC_ZIP} && \
+    unzip -o ${PROTOC_ZIP} -d /usr/local bin/protoc && \
+    unzip -o ${PROTOC_ZIP} -d /usr/local 'include/*' && \
+    rm -f ${PROTOC_ZIP}
+
+# Install Rust.
+SHELL ["/bin/bash", "-c"]
+RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- --no-modify-path --default-toolchain none -y
+ENV PATH /root/.cargo/bin/:$PATH
+
+# Install Rust toolchains.
+ARG RUST_TOOLCHAIN
+RUN rustup toolchain install ${RUST_TOOLCHAIN}
+
+# Install cargo-binstall with a specific version to adapt the current rust toolchain.
+# Note: if we use the latest version, we may encounter the following `use of unstable library feature 'io_error_downcast'` error.
+RUN cargo install cargo-binstall --version 1.6.6 --locked
+
+# Install nextest.
+RUN cargo binstall cargo-nextest --no-confirm
--- a/docker/dev-builder/ubuntu/Dockerfile-20.04
+++ b/docker/dev-builder/ubuntu/Dockerfile-20.04
@@ -1,66 +0,0 @@
-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
-
-RUN apt-get update && \
-    DEBIAN_FRONTEND=noninteractive apt-get install -y software-properties-common
-# Install dependencies.
-RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
-    libssl-dev \
-    tzdata \
-    curl \
-    unzip \
-    ca-certificates \
-    git \
-    build-essential \
-    pkg-config
-
-ARG TARGETPLATFORM
-RUN echo "target platform: $TARGETPLATFORM"
-
-ARG PROTOBUF_VERSION=29.3
-
-# Install protobuf, because the one in the apt is too old (v3.12).
-RUN if [ "$TARGETPLATFORM" = "linux/arm64" ]; then \
-    curl -OL https://github.com/protocolbuffers/protobuf/releases/download/v${PROTOBUF_VERSION}/protoc-${PROTOBUF_VERSION}-linux-aarch_64.zip && \
-    unzip protoc-${PROTOBUF_VERSION}-linux-aarch_64.zip -d protoc3; \
-elif [ "$TARGETPLATFORM" = "linux/amd64" ]; then \
-    curl -OL https://github.com/protocolbuffers/protobuf/releases/download/v${PROTOBUF_VERSION}/protoc-${PROTOBUF_VERSION}-linux-x86_64.zip && \
-    unzip protoc-${PROTOBUF_VERSION}-linux-x86_64.zip -d protoc3; \
-fi
-RUN mv protoc3/bin/* /usr/local/bin/
-RUN mv protoc3/include/* /usr/local/include/
-
-# Silence all `safe.directory` warnings, to avoid the "detect dubious repository" error when building with submodules.
-# Disabling the safe directory check here won't pose extra security issues, because in our usage for this dev build
-# image, we use it solely on our own environment (that github action's VM, or ECS created dynamically by ourselves),
-# and the repositories are pulled from trusted sources (still us, of course). Doing so does not violate the intention
-# of the Git's addition to the "safe.directory" at the first place (see the commit message here:
-# https://github.com/git/git/commit/8959555cee7ec045958f9b6dd62e541affb7e7d9).
-# There's also another solution to this, that we add the desired submodules to the safe directory, instead of using 
-# wildcard here. However, that requires the git's config files and the submodules all owned by the very same user.
-# It's troublesome to do this since the dev build runs in Docker, which is under user "root"; while outside the Docker,
-# it can be a different user that have prepared the submodules.
-RUN git config --global --add safe.directory '*'
-
-# Install Rust.
-SHELL ["/bin/bash", "-c"]
-RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- --no-modify-path --default-toolchain none -y
-ENV PATH /root/.cargo/bin/:$PATH
-
-# Install Rust toolchains.
-ARG RUST_TOOLCHAIN
-RUN rustup toolchain install ${RUST_TOOLCHAIN}
-
-# Install cargo-binstall with a specific version to adapt the current rust toolchain.
-# Note: if we use the latest version, we may encounter the following `use of unstable library feature 'io_error_downcast'` error.
-# compile from source take too long, so we use the precompiled binary instead
-COPY $DOCKER_BUILD_ROOT/docker/dev-builder/binstall/pull_binstall.sh /usr/local/bin/pull_binstall.sh
-RUN chmod +x /usr/local/bin/pull_binstall.sh && /usr/local/bin/pull_binstall.sh
-
-# Install nextest.
-RUN cargo binstall cargo-nextest --no-confirm
--- a/docker/docker-compose/cluster-with-etcd.yaml
+++ b/docker/docker-compose/cluster-with-etcd.yaml
@@ -43,8 +43,8 @@ services:
    command:
      - metasrv
      - start
-      - --rpc-bind-addr=0.0.0.0:3002
-      - --rpc-server-addr=metasrv:3002
+      - --bind-addr=0.0.0.0:3002
+      - --server-addr=metasrv:3002
      - --store-addrs=etcd0:2379
      - --http-addr=0.0.0.0:3000
    healthcheck:
@@ -68,8 +68,8 @@ services:
      - datanode
      - start
      - --node-id=0
-      - --rpc-bind-addr=0.0.0.0:3001
-      - --rpc-server-addr=datanode0:3001
+      - --rpc-addr=0.0.0.0:3001
+      - --rpc-hostname=datanode0:3001
      - --metasrv-addrs=metasrv:3002
      - --http-addr=0.0.0.0:5000
    volumes:
@@ -98,7 +98,7 @@ services:
      - start
      - --metasrv-addrs=metasrv:3002
      - --http-addr=0.0.0.0:4000
-      - --rpc-bind-addr=0.0.0.0:4001
+      - --rpc-addr=0.0.0.0:4001
      - --mysql-addr=0.0.0.0:4002
      - --postgres-addr=0.0.0.0:4003
    healthcheck:
@@ -123,8 +123,8 @@ services:
      - start
      - --node-id=0
      - --metasrv-addrs=metasrv:3002
-      - --rpc-bind-addr=0.0.0.0:4004
-      - --rpc-server-addr=flownode0:4004
+      - --rpc-addr=0.0.0.0:4004
+      - --rpc-hostname=flownode0:4004
      - --http-addr=0.0.0.0:4005
    depends_on:
      frontend0:
--- a/docs/benchmarks/tsbs/v0.12.0.md
+++ b/docs/benchmarks/tsbs/v0.12.0.md
@@ -1,40 +0,0 @@
-# TSBS benchmark - v0.12.0
-
-## Environment
-
-### Amazon EC2
-
-|         |                         |
-|---------|-------------------------|
-| Machine | c5d.2xlarge             |
-| CPU     | 8 core                  |
-| Memory  | 16GB                    |
-| Disk    | 100GB (GP3)             |
-| OS      | Ubuntu Server 24.04 LTS |
-
-## Write performance
-
-| Environment     | Ingest rate (rows/s) |
-|-----------------|----------------------|
-| EC2 c5d.2xlarge | 326839.28            |
-
-## Query performance
-
-| Query type            | EC2 c5d.2xlarge (ms) |
-|-----------------------|----------------------|
-| cpu-max-all-1         | 12.46                |
-| cpu-max-all-8         | 24.20                |
-| double-groupby-1      | 673.08               |
-| double-groupby-5      | 963.99               |
-| double-groupby-all    | 1330.05              |
-| groupby-orderby-limit | 952.46               |
-| high-cpu-1            | 5.08                 |
-| high-cpu-all          | 4638.57              |
-| lastpoint             | 591.02               |
-| single-groupby-1-1-1  | 4.06                 |
-| single-groupby-1-1-12 | 4.73                 |
-| single-groupby-1-8-1  | 8.23                 |
-| single-groupby-5-1-1  | 4.61                 |
-| single-groupby-5-1-12 | 5.61                 |
-| single-groupby-5-8-1  | 9.74                 |
-
--- a/docs/how-to/how-to-profile-memory.md
+++ b/docs/how-to/how-to-profile-memory.md
@@ -4,16 +4,6 @@ This crate provides an easy approach to dump memory profiling info.

 ## Prerequisites
 ### jemalloc
-jeprof is already compiled in the target directory of GreptimeDB. You can find the binary and use it.
-```
-# find jeprof binary
-find . -name 'jeprof'
-# add executable permission
-chmod +x <path_to_jeprof>
-```
-The path is usually under `./target/${PROFILE}/build/tikv-jemalloc-sys-${HASH}/out/build/bin/jeprof`.
-The default version of jemalloc installed from the package manager may not have the `--collapsed` option.
-You may need to check the whether the `jeprof` version is >= `5.3.0` if you want to install it from the package manager.
 ```bash
 # for macOS
 brew install jemalloc
@@ -33,11 +23,7 @@ curl https://raw.githubusercontent.com/brendangregg/FlameGraph/master/flamegraph
 Start GreptimeDB instance with environment variables:

 ```bash
-# for Linux
 MALLOC_CONF=prof:true ./target/debug/greptime standalone start
-
-# for macOS
-_RJEM_MALLOC_CONF=prof:true ./target/debug/greptime standalone start
 ```

 Dump memory profiling data through HTTP API:
--- a/docs/logo-text-padding-dark.png
+++ b/docs/logo-text-padding-dark.png
--- a/docs/logo-text-padding.png
+++ b/docs/logo-text-padding.png
--- a/grafana/check.sh
+++ b/grafana/check.sh
@@ -1,19 +0,0 @@
-#!/usr/bin/env bash
-
-BASEDIR=$(dirname "$0")
-
-# Use jq to check for panels with empty or missing descriptions
-invalid_panels=$(cat $BASEDIR/greptimedb-cluster.json | jq -r '
-  .panels[]
-  | select((.type == "stats" or .type == "timeseries") and (.description == "" or .description == null))
-')
-
-# Check if any invalid panels were found
-if [[ -n "$invalid_panels" ]]; then
-  echo "Error: The following panels have empty or missing descriptions:"
-  echo "$invalid_panels"
-  exit 1
-else
-  echo "All panels with type 'stats' or 'timeseries' have valid descriptions."
-  exit 0
-fi
--- a/grafana/greptimedb-cluster.json
+++ b/grafana/greptimedb-cluster.json
--- a/grafana/greptimedb.json
+++ b/grafana/greptimedb.json
@@ -384,8 +384,8 @@
        "rowHeight": 0.9,
        "showValue": "auto",
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "none",
+          "sort": "none"
        }
      },
      "targets": [
@@ -483,8 +483,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "pluginVersion": "10.2.3",
@@ -578,8 +578,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "pluginVersion": "10.2.3",
@@ -601,7 +601,7 @@
      "type": "timeseries"
    },
    {
-      "collapsed": false,
+      "collapsed": true,
      "gridPos": {
        "h": 1,
        "w": 24,
@@ -684,8 +684,8 @@
              "showLegend": true
            },
            "tooltip": {
-              "mode": "multi",
-              "sort": "desc"
+              "mode": "single",
+              "sort": "none"
            }
          },
          "targets": [
@@ -878,8 +878,8 @@
              "showLegend": true
            },
            "tooltip": {
-              "mode": "multi",
-              "sort": "desc"
+              "mode": "single",
+              "sort": "none"
            }
          },
          "targets": [
@@ -1124,8 +1124,8 @@
              "showLegend": true
            },
            "tooltip": {
-              "mode": "multi",
-              "sort": "desc"
+              "mode": "single",
+              "sort": "none"
            }
          },
          "targets": [
@@ -1223,8 +1223,8 @@
              "showLegend": true
            },
            "tooltip": {
-              "mode": "multi",
-              "sort": "desc"
+              "mode": "single",
+              "sort": "none"
            }
          },
          "targets": [
@@ -1322,8 +1322,8 @@
              "showLegend": true
            },
            "tooltip": {
-              "mode": "multi",
-              "sort": "desc"
+              "mode": "single",
+              "sort": "none"
            }
          },
          "targets": [
@@ -1456,8 +1456,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -1573,8 +1573,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -1673,8 +1673,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -1773,8 +1773,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -1890,8 +1890,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2002,8 +2002,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2120,8 +2120,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2233,8 +2233,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2334,8 +2334,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2435,8 +2435,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2548,8 +2548,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2661,8 +2661,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2788,8 +2788,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2889,8 +2889,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -2990,8 +2990,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3091,8 +3091,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3191,8 +3191,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3302,8 +3302,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3432,8 +3432,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3543,8 +3543,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3657,8 +3657,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3808,8 +3808,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -3909,8 +3909,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -4011,8 +4011,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
@@ -4113,8 +4113,8 @@
          "showLegend": true
        },
        "tooltip": {
-          "mode": "multi",
-          "sort": "desc"
+          "mode": "single",
+          "sort": "none"
        }
      },
      "targets": [
--- a/grafana/summary.sh
+++ b/grafana/summary.sh
@@ -1,11 +0,0 @@
-#!/usr/bin/env bash
-
-BASEDIR=$(dirname "$0")
-echo '| Title | Description | Expressions |
-|---|---|---|'
-
-cat $BASEDIR/greptimedb-cluster.json | jq -r '
-  .panels |
-  map(select(.type == "stat" or .type == "timeseries")) |
-  .[] | "| \(.title) | \(.description | gsub("\n"; "<br>")) | \(.targets | map(.expr // .rawSql | "`\(.|gsub("\n"; "<br>"))`")  | join("<br>")) |"
-'
--- a/src/api/Cargo.toml
+++ b/src/api/Cargo.toml
@@ -15,10 +15,13 @@ common-macro.workspace = true
 common-time.workspace = true
 datatypes.workspace = true
 greptime-proto.workspace = true
-paste.workspace = true
+paste = "1.0"
 prost.workspace = true
 serde_json.workspace = true
 snafu.workspace = true

 [build-dependencies]
 tonic-build = "0.11"
+
+[dev-dependencies]
+paste = "1.0"
--- a/src/api/src/helper.rs
+++ b/src/api/src/helper.rs
@@ -19,7 +19,9 @@ use common_decimal::decimal128::{DECIMAL128_DEFAULT_SCALE, DECIMAL128_MAX_PRECIS
 use common_decimal::Decimal128;
 use common_time::time::Time;
 use common_time::timestamp::TimeUnit;
-use common_time::{Date, IntervalDayTime, IntervalMonthDayNano, IntervalYearMonth, Timestamp};
+use common_time::{
+    Date, DateTime, IntervalDayTime, IntervalMonthDayNano, IntervalYearMonth, Timestamp,
+};
 use datatypes::prelude::{ConcreteDataType, ValueRef};
 use datatypes::scalars::ScalarVector;
 use datatypes::types::{
@@ -27,8 +29,8 @@ use datatypes::types::{
 };
 use datatypes::value::{OrderedF32, OrderedF64, Value};
 use datatypes::vectors::{
-    BinaryVector, BooleanVector, DateVector, Decimal128Vector, Float32Vector, Float64Vector,
-    Int32Vector, Int64Vector, IntervalDayTimeVector, IntervalMonthDayNanoVector,
+    BinaryVector, BooleanVector, DateTimeVector, DateVector, Decimal128Vector, Float32Vector,
+    Float64Vector, Int32Vector, Int64Vector, IntervalDayTimeVector, IntervalMonthDayNanoVector,
    IntervalYearMonthVector, PrimitiveVector, StringVector, TimeMicrosecondVector,
    TimeMillisecondVector, TimeNanosecondVector, TimeSecondVector, TimestampMicrosecondVector,
    TimestampMillisecondVector, TimestampNanosecondVector, TimestampSecondVector, UInt32Vector,
@@ -116,7 +118,7 @@ impl From<ColumnDataTypeWrapper> for ConcreteDataType {
            ColumnDataType::Json => ConcreteDataType::json_datatype(),
            ColumnDataType::String => ConcreteDataType::string_datatype(),
            ColumnDataType::Date => ConcreteDataType::date_datatype(),
-            ColumnDataType::Datetime => ConcreteDataType::timestamp_microsecond_datatype(),
+            ColumnDataType::Datetime => ConcreteDataType::datetime_datatype(),
            ColumnDataType::TimestampSecond => ConcreteDataType::timestamp_second_datatype(),
            ColumnDataType::TimestampMillisecond => {
                ConcreteDataType::timestamp_millisecond_datatype()
@@ -269,6 +271,7 @@ impl TryFrom<ConcreteDataType> for ColumnDataTypeWrapper {
            ConcreteDataType::Binary(_) => ColumnDataType::Binary,
            ConcreteDataType::String(_) => ColumnDataType::String,
            ConcreteDataType::Date(_) => ColumnDataType::Date,
+            ConcreteDataType::DateTime(_) => ColumnDataType::Datetime,
            ConcreteDataType::Timestamp(t) => match t {
                TimestampType::Second(_) => ColumnDataType::TimestampSecond,
                TimestampType::Millisecond(_) => ColumnDataType::TimestampMillisecond,
@@ -473,6 +476,7 @@ pub fn push_vals(column: &mut Column, origin_count: usize, vector: VectorRef) {
        Value::String(val) => values.string_values.push(val.as_utf8().to_string()),
        Value::Binary(val) => values.binary_values.push(val.to_vec()),
        Value::Date(val) => values.date_values.push(val.val()),
+        Value::DateTime(val) => values.datetime_values.push(val.val()),
        Value::Timestamp(val) => match val.unit() {
            TimeUnit::Second => values.timestamp_second_values.push(val.value()),
            TimeUnit::Millisecond => values.timestamp_millisecond_values.push(val.value()),
@@ -573,11 +577,12 @@ pub fn pb_value_to_value_ref<'a>(
        ValueData::BinaryValue(bytes) => ValueRef::Binary(bytes.as_slice()),
        ValueData::StringValue(string) => ValueRef::String(string.as_str()),
        ValueData::DateValue(d) => ValueRef::Date(Date::from(*d)),
+        ValueData::DatetimeValue(d) => ValueRef::DateTime(DateTime::new(*d)),
        ValueData::TimestampSecondValue(t) => ValueRef::Timestamp(Timestamp::new_second(*t)),
        ValueData::TimestampMillisecondValue(t) => {
            ValueRef::Timestamp(Timestamp::new_millisecond(*t))
        }
-        ValueData::DatetimeValue(t) | ValueData::TimestampMicrosecondValue(t) => {
+        ValueData::TimestampMicrosecondValue(t) => {
            ValueRef::Timestamp(Timestamp::new_microsecond(*t))
        }
        ValueData::TimestampNanosecondValue(t) => {
@@ -646,6 +651,7 @@ pub fn pb_values_to_vector_ref(data_type: &ConcreteDataType, values: Values) ->
        ConcreteDataType::Binary(_) => Arc::new(BinaryVector::from(values.binary_values)),
        ConcreteDataType::String(_) => Arc::new(StringVector::from_vec(values.string_values)),
        ConcreteDataType::Date(_) => Arc::new(DateVector::from_vec(values.date_values)),
+        ConcreteDataType::DateTime(_) => Arc::new(DateTimeVector::from_vec(values.datetime_values)),
        ConcreteDataType::Timestamp(unit) => match unit {
            TimestampType::Second(_) => Arc::new(TimestampSecondVector::from_vec(
                values.timestamp_second_values,
@@ -781,6 +787,11 @@ pub fn pb_values_to_values(data_type: &ConcreteDataType, values: Values) -> Vec<
            .into_iter()
            .map(|val| val.into())
            .collect(),
+        ConcreteDataType::DateTime(_) => values
+            .datetime_values
+            .into_iter()
+            .map(|v| Value::DateTime(v.into()))
+            .collect(),
        ConcreteDataType::Date(_) => values
            .date_values
            .into_iter()
@@ -936,6 +947,9 @@ pub fn to_proto_value(value: Value) -> Option<v1::Value> {
        Value::Date(v) => v1::Value {
            value_data: Some(ValueData::DateValue(v.val())),
        },
+        Value::DateTime(v) => v1::Value {
+            value_data: Some(ValueData::DatetimeValue(v.val())),
+        },
        Value::Timestamp(v) => match v.unit() {
            TimeUnit::Second => v1::Value {
                value_data: Some(ValueData::TimestampSecondValue(v.value())),
@@ -1052,6 +1066,7 @@ pub fn value_to_grpc_value(value: Value) -> GrpcValue {
            Value::String(v) => Some(ValueData::StringValue(v.as_utf8().to_string())),
            Value::Binary(v) => Some(ValueData::BinaryValue(v.to_vec())),
            Value::Date(v) => Some(ValueData::DateValue(v.val())),
+            Value::DateTime(v) => Some(ValueData::DatetimeValue(v.val())),
            Value::Timestamp(v) => Some(match v.unit() {
                TimeUnit::Second => ValueData::TimestampSecondValue(v.value()),
                TimeUnit::Millisecond => ValueData::TimestampMillisecondValue(v.value()),
@@ -1233,7 +1248,7 @@ mod tests {
            ColumnDataTypeWrapper::date_datatype().into()
        );
        assert_eq!(
-            ConcreteDataType::timestamp_microsecond_datatype(),
+            ConcreteDataType::datetime_datatype(),
            ColumnDataTypeWrapper::datetime_datatype().into()
        );
        assert_eq!(
@@ -1324,6 +1339,10 @@ mod tests {
            ColumnDataTypeWrapper::date_datatype(),
            ConcreteDataType::date_datatype().try_into().unwrap()
        );
+        assert_eq!(
+            ColumnDataTypeWrapper::datetime_datatype(),
+            ConcreteDataType::datetime_datatype().try_into().unwrap()
+        );
        assert_eq!(
            ColumnDataTypeWrapper::timestamp_millisecond_datatype(),
            ConcreteDataType::timestamp_millisecond_datatype()
@@ -1811,6 +1830,17 @@ mod tests {
        ]
    );

+    test_convert_values!(
+        datetime,
+        vec![1.into(), 2.into(), 3.into()],
+        datetime,
+        vec![
+            Value::DateTime(1.into()),
+            Value::DateTime(2.into()),
+            Value::DateTime(3.into())
+        ]
+    );
+
    #[test]
    fn test_vectors_to_rows_for_different_types() {
        let boolean_vec = BooleanVector::from_vec(vec![true, false, true]);
--- a/src/api/src/v1/column_def.rs
+++ b/src/api/src/v1/column_def.rs
@@ -15,10 +15,10 @@
 use std::collections::HashMap;

 use datatypes::schema::{
-    ColumnDefaultConstraint, ColumnSchema, FulltextAnalyzer, FulltextOptions, SkippingIndexOptions,
-    SkippingIndexType, COMMENT_KEY, FULLTEXT_KEY, INVERTED_INDEX_KEY, SKIPPING_INDEX_KEY,
+    ColumnDefaultConstraint, ColumnSchema, FulltextAnalyzer, FulltextOptions, COMMENT_KEY,
+    FULLTEXT_KEY, INVERTED_INDEX_KEY, SKIPPING_INDEX_KEY,
 };
-use greptime_proto::v1::{Analyzer, SkippingIndexType as PbSkippingIndexType};
+use greptime_proto::v1::Analyzer;
 use snafu::ResultExt;

 use crate::error::{self, Result};
@@ -103,13 +103,6 @@ pub fn contains_fulltext(options: &Option<ColumnOptions>) -> bool {
        .is_some_and(|o| o.options.contains_key(FULLTEXT_GRPC_KEY))
 }

-/// Checks if the `ColumnOptions` contains skipping index options.
-pub fn contains_skipping(options: &Option<ColumnOptions>) -> bool {
-    options
-        .as_ref()
-        .is_some_and(|o| o.options.contains_key(SKIPPING_INDEX_GRPC_KEY))
-}
-
 /// Tries to construct a `ColumnOptions` from the given `FulltextOptions`.
 pub fn options_from_fulltext(fulltext: &FulltextOptions) -> Result<Option<ColumnOptions>> {
    let mut options = ColumnOptions::default();
@@ -120,27 +113,6 @@ pub fn options_from_fulltext(fulltext: &FulltextOptions) -> Result<Option<Column
    Ok((!options.options.is_empty()).then_some(options))
 }

-/// Tries to construct a `ColumnOptions` from the given `SkippingIndexOptions`.
-pub fn options_from_skipping(skipping: &SkippingIndexOptions) -> Result<Option<ColumnOptions>> {
-    let mut options = ColumnOptions::default();
-
-    let v = serde_json::to_string(skipping).context(error::SerializeJsonSnafu)?;
-    options
-        .options
-        .insert(SKIPPING_INDEX_GRPC_KEY.to_string(), v);
-
-    Ok((!options.options.is_empty()).then_some(options))
-}
-
-/// Tries to construct a `ColumnOptions` for inverted index.
-pub fn options_from_inverted() -> ColumnOptions {
-    let mut options = ColumnOptions::default();
-    options
-        .options
-        .insert(INVERTED_INDEX_GRPC_KEY.to_string(), "true".to_string());
-    options
-}
-
 /// Tries to construct a `FulltextAnalyzer` from the given analyzer.
 pub fn as_fulltext_option(analyzer: Analyzer) -> FulltextAnalyzer {
    match analyzer {
@@ -149,13 +121,6 @@ pub fn as_fulltext_option(analyzer: Analyzer) -> FulltextAnalyzer {
    }
 }

-/// Tries to construct a `SkippingIndexType` from the given skipping index type.
-pub fn as_skipping_index_type(skipping_index_type: PbSkippingIndexType) -> SkippingIndexType {
-    match skipping_index_type {
-        PbSkippingIndexType::BloomFilter => SkippingIndexType::BloomFilter,
-    }
-}
-
 #[cfg(test)]
 mod tests {

--- a/src/catalog/Cargo.toml
+++ b/src/catalog/Cargo.toml
@@ -15,7 +15,7 @@ api.workspace = true
 arrow.workspace = true
 arrow-schema.workspace = true
 async-stream.workspace = true
-async-trait.workspace = true
+async-trait = "0.1"
 bytes.workspace = true
 common-catalog.workspace = true
 common-error.workspace = true
@@ -31,7 +31,7 @@ common-version.workspace = true
 dashmap.workspace = true
 datafusion.workspace = true
 datatypes.workspace = true
-futures.workspace = true
+futures = "0.3"
 futures-util.workspace = true
 humantime.workspace = true
 itertools.workspace = true
@@ -39,7 +39,7 @@ lazy_static.workspace = true
 meta-client.workspace = true
 moka = { workspace = true, features = ["future", "sync"] }
 partition.workspace = true
-paste.workspace = true
+paste = "1.0"
 prometheus.workspace = true
 rustc-hash.workspace = true
 serde_json.workspace = true
@@ -49,7 +49,7 @@ sql.workspace = true
 store-api.workspace = true
 table.workspace = true
 tokio.workspace = true
-tokio-stream.workspace = true
+tokio-stream = "0.1"

 [dev-dependencies]
 cache.workspace = true
--- a/src/catalog/src/kvbackend/manager.rs
+++ b/src/catalog/src/kvbackend/manager.rs
@@ -38,7 +38,6 @@ use partition::manager::{PartitionRuleManager, PartitionRuleManagerRef};
 use session::context::{Channel, QueryContext};
 use snafu::prelude::*;
 use table::dist_table::DistTable;
-use table::metadata::TableId;
 use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
 use table::table_name::TableName;
 use table::TableRef;
@@ -287,28 +286,6 @@ impl CatalogManager for KvBackendCatalogManager {
        return Ok(None);
    }

-    async fn tables_by_ids(
-        &self,
-        catalog: &str,
-        schema: &str,
-        table_ids: &[TableId],
-    ) -> Result<Vec<TableRef>> {
-        let table_info_values = self
-            .table_metadata_manager
-            .table_info_manager()
-            .batch_get(table_ids)
-            .await
-            .context(TableMetadataManagerSnafu)?;
-
-        let tables = table_info_values
-            .into_values()
-            .filter(|t| t.table_info.catalog_name == catalog && t.table_info.schema_name == schema)
-            .map(build_table)
-            .collect::<Result<Vec<_>>>()?;
-
-        Ok(tables)
-    }
-
    fn tables<'a>(
        &'a self,
        catalog: &'a str,
--- a/src/catalog/src/lib.rs
+++ b/src/catalog/src/lib.rs
@@ -87,14 +87,6 @@ pub trait CatalogManager: Send + Sync {
        query_ctx: Option<&QueryContext>,
    ) -> Result<Option<TableRef>>;

-    /// Returns the tables by table ids.
-    async fn tables_by_ids(
-        &self,
-        catalog: &str,
-        schema: &str,
-        table_ids: &[TableId],
-    ) -> Result<Vec<TableRef>>;
-
    /// Returns all tables with a stream by catalog and schema.
    fn tables<'a>(
        &'a self,
--- a/src/catalog/src/memory/manager.rs
+++ b/src/catalog/src/memory/manager.rs
@@ -14,7 +14,7 @@

 use std::any::Any;
 use std::collections::hash_map::Entry;
-use std::collections::{HashMap, HashSet};
+use std::collections::HashMap;
 use std::sync::{Arc, RwLock, Weak};

 use async_stream::{stream, try_stream};
@@ -28,7 +28,6 @@ use common_meta::kv_backend::memory::MemoryKvBackend;
 use futures_util::stream::BoxStream;
 use session::context::QueryContext;
 use snafu::OptionExt;
-use table::metadata::TableId;
 use table::TableRef;

 use crate::error::{CatalogNotFoundSnafu, Result, SchemaNotFoundSnafu, TableExistsSnafu};
@@ -144,33 +143,6 @@ impl CatalogManager for MemoryCatalogManager {
        Ok(result)
    }

-    async fn tables_by_ids(
-        &self,
-        catalog: &str,
-        schema: &str,
-        table_ids: &[TableId],
-    ) -> Result<Vec<TableRef>> {
-        let catalogs = self.catalogs.read().unwrap();
-
-        let schemas = catalogs.get(catalog).context(CatalogNotFoundSnafu {
-            catalog_name: catalog,
-        })?;
-
-        let tables = schemas
-            .get(schema)
-            .context(SchemaNotFoundSnafu { catalog, schema })?;
-
-        let filter_ids: HashSet<_> = table_ids.iter().collect();
-        // It is very inefficient, but we do not need to optimize it since it will not be called in `MemoryCatalogManager`.
-        let tables = tables
-            .values()
-            .filter(|t| filter_ids.contains(&t.table_info().table_id()))
-            .cloned()
-            .collect::<Vec<_>>();
-
-        Ok(tables)
-    }
-
    fn tables<'a>(
        &'a self,
        catalog: &'a str,
--- a/src/catalog/src/system_schema.rs
+++ b/src/catalog/src/system_schema.rs
@@ -77,7 +77,7 @@ trait SystemSchemaProviderInner {
    fn system_table(&self, name: &str) -> Option<SystemTableRef>;

    fn table_info(catalog_name: String, table: &SystemTableRef) -> TableInfoRef {
-        let table_meta = TableMetaBuilder::empty()
+        let table_meta = TableMetaBuilder::default()
            .schema(table.schema())
            .primary_key_indices(vec![])
            .next_column_id(0)
--- a/src/catalog/src/system_schema/information_schema/columns.rs
+++ b/src/catalog/src/system_schema/information_schema/columns.rs
@@ -365,6 +365,10 @@ impl InformationSchemaColumnsBuilder {
            self.numeric_scales.push(None);

            match &column_schema.data_type {
+                ConcreteDataType::DateTime(datetime_type) => {
+                    self.datetime_precisions
+                        .push(Some(datetime_type.precision() as i64));
+                }
                ConcreteDataType::Timestamp(ts_type) => {
                    self.datetime_precisions
                        .push(Some(ts_type.precision() as i64));
--- a/src/catalog/src/system_schema/information_schema/flows.rs
+++ b/src/catalog/src/system_schema/information_schema/flows.rs
@@ -28,19 +28,16 @@ use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
 use datatypes::prelude::ConcreteDataType as CDT;
 use datatypes::scalars::ScalarVectorBuilder;
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
-use datatypes::timestamp::TimestampMillisecond;
 use datatypes::value::Value;
 use datatypes::vectors::{
-    Int64VectorBuilder, StringVectorBuilder, TimestampMillisecondVectorBuilder,
-    UInt32VectorBuilder, UInt64VectorBuilder, VectorRef,
+    Int64VectorBuilder, StringVectorBuilder, UInt32VectorBuilder, UInt64VectorBuilder, VectorRef,
 };
 use futures::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::{ScanRequest, TableId};

 use crate::error::{
-    CreateRecordBatchSnafu, FlowInfoNotFoundSnafu, InternalSnafu, JsonSnafu, ListFlowsSnafu,
-    Result, UpgradeWeakCatalogManagerRefSnafu,
+    CreateRecordBatchSnafu, FlowInfoNotFoundSnafu, InternalSnafu, JsonSnafu, ListFlowsSnafu, Result,
 };
 use crate::information_schema::{Predicates, FLOWS};
 use crate::system_schema::information_schema::InformationTable;
@@ -62,10 +59,6 @@ pub const SOURCE_TABLE_IDS: &str = "source_table_ids";
 pub const SINK_TABLE_NAME: &str = "sink_table_name";
 pub const FLOWNODE_IDS: &str = "flownode_ids";
 pub const OPTIONS: &str = "options";
-pub const CREATED_TIME: &str = "created_time";
-pub const UPDATED_TIME: &str = "updated_time";
-pub const LAST_EXECUTION_TIME: &str = "last_execution_time";
-pub const SOURCE_TABLE_NAMES: &str = "source_table_names";

 /// The `information_schema.flows` to provides information about flows in databases.
 #[derive(Debug)]
@@ -106,14 +99,6 @@ impl InformationSchemaFlows {
                (SINK_TABLE_NAME, CDT::string_datatype(), false),
                (FLOWNODE_IDS, CDT::string_datatype(), true),
                (OPTIONS, CDT::string_datatype(), true),
-                (CREATED_TIME, CDT::timestamp_millisecond_datatype(), false),
-                (UPDATED_TIME, CDT::timestamp_millisecond_datatype(), false),
-                (
-                    LAST_EXECUTION_TIME,
-                    CDT::timestamp_millisecond_datatype(),
-                    true,
-                ),
-                (SOURCE_TABLE_NAMES, CDT::string_datatype(), true),
            ]
            .into_iter()
            .map(|(name, ty, nullable)| ColumnSchema::new(name, ty, nullable))
@@ -185,10 +170,6 @@ struct InformationSchemaFlowsBuilder {
    sink_table_names: StringVectorBuilder,
    flownode_id_groups: StringVectorBuilder,
    option_groups: StringVectorBuilder,
-    created_time: TimestampMillisecondVectorBuilder,
-    updated_time: TimestampMillisecondVectorBuilder,
-    last_execution_time: TimestampMillisecondVectorBuilder,
-    source_table_names: StringVectorBuilder,
 }

 impl InformationSchemaFlowsBuilder {
@@ -215,10 +196,6 @@ impl InformationSchemaFlowsBuilder {
            sink_table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
            flownode_id_groups: StringVectorBuilder::with_capacity(INIT_CAPACITY),
            option_groups: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            created_time: TimestampMillisecondVectorBuilder::with_capacity(INIT_CAPACITY),
-            updated_time: TimestampMillisecondVectorBuilder::with_capacity(INIT_CAPACITY),
-            last_execution_time: TimestampMillisecondVectorBuilder::with_capacity(INIT_CAPACITY),
-            source_table_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
        }
    }

@@ -258,14 +235,13 @@ impl InformationSchemaFlowsBuilder {
                    catalog_name: catalog_name.to_string(),
                    flow_name: flow_name.to_string(),
                })?;
-            self.add_flow(&predicates, flow_id.flow_id(), flow_info, &flow_stat)
-                .await?;
+            self.add_flow(&predicates, flow_id.flow_id(), flow_info, &flow_stat)?;
        }

        self.finish()
    }

-    async fn add_flow(
+    fn add_flow(
        &mut self,
        predicates: &Predicates,
        flow_id: FlowId,
@@ -314,36 +290,6 @@ impl InformationSchemaFlowsBuilder {
                    input: format!("{:?}", flow_info.options()),
                },
            )?));
-        self.created_time
-            .push(Some(flow_info.created_time().timestamp_millis().into()));
-        self.updated_time
-            .push(Some(flow_info.updated_time().timestamp_millis().into()));
-        self.last_execution_time
-            .push(flow_stat.as_ref().and_then(|state| {
-                state
-                    .last_exec_time_map
-                    .get(&flow_id)
-                    .map(|v| TimestampMillisecond::new(*v))
-            }));
-
-        let mut source_table_names = vec![];
-        let catalog_name = self.catalog_name.clone();
-        let catalog_manager = self
-            .catalog_manager
-            .upgrade()
-            .context(UpgradeWeakCatalogManagerRefSnafu)?;
-        for schema_name in catalog_manager.schema_names(&catalog_name, None).await? {
-            source_table_names.extend(
-                catalog_manager
-                    .tables_by_ids(&catalog_name, &schema_name, flow_info.source_table_ids())
-                    .await?
-                    .into_iter()
-                    .map(|table| table.table_info().full_table_name()),
-            );
-        }
-
-        let source_table_names = source_table_names.join(",");
-        self.source_table_names.push(Some(&source_table_names));

        Ok(())
    }
@@ -361,10 +307,6 @@ impl InformationSchemaFlowsBuilder {
            Arc::new(self.sink_table_names.finish()),
            Arc::new(self.flownode_id_groups.finish()),
            Arc::new(self.option_groups.finish()),
-            Arc::new(self.created_time.finish()),
-            Arc::new(self.updated_time.finish()),
-            Arc::new(self.last_execution_time.finish()),
-            Arc::new(self.source_table_names.finish()),
        ];
        RecordBatch::new(self.schema.clone(), columns).context(CreateRecordBatchSnafu)
    }
--- a/src/catalog/src/system_schema/information_schema/information_memory_table.rs
+++ b/src/catalog/src/system_schema/information_schema/information_memory_table.rs
@@ -20,7 +20,7 @@ use datatypes::vectors::{Int64Vector, StringVector, VectorRef};

 use super::table_names::*;
 use crate::system_schema::utils::tables::{
-    bigint_column, string_column, string_columns, timestamp_micro_column,
+    bigint_column, datetime_column, string_column, string_columns,
 };

 const NO_VALUE: &str = "NO";
@@ -163,17 +163,17 @@ pub(super) fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>
                string_column("EVENT_BODY"),
                string_column("EVENT_DEFINITION"),
                string_column("EVENT_TYPE"),
-                timestamp_micro_column("EXECUTE_AT"),
+                datetime_column("EXECUTE_AT"),
                bigint_column("INTERVAL_VALUE"),
                string_column("INTERVAL_FIELD"),
                string_column("SQL_MODE"),
-                timestamp_micro_column("STARTS"),
-                timestamp_micro_column("ENDS"),
+                datetime_column("STARTS"),
+                datetime_column("ENDS"),
                string_column("STATUS"),
                string_column("ON_COMPLETION"),
-                timestamp_micro_column("CREATED"),
-                timestamp_micro_column("LAST_ALTERED"),
-                timestamp_micro_column("LAST_EXECUTED"),
+                datetime_column("CREATED"),
+                datetime_column("LAST_ALTERED"),
+                datetime_column("LAST_EXECUTED"),
                string_column("EVENT_COMMENT"),
                bigint_column("ORIGINATOR"),
                string_column("CHARACTER_SET_CLIENT"),
@@ -204,10 +204,10 @@ pub(super) fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>
                bigint_column("INITIAL_SIZE"),
                bigint_column("MAXIMUM_SIZE"),
                bigint_column("AUTOEXTEND_SIZE"),
-                timestamp_micro_column("CREATION_TIME"),
-                timestamp_micro_column("LAST_UPDATE_TIME"),
-                timestamp_micro_column("LAST_ACCESS_TIME"),
-                timestamp_micro_column("RECOVER_TIME"),
+                datetime_column("CREATION_TIME"),
+                datetime_column("LAST_UPDATE_TIME"),
+                datetime_column("LAST_ACCESS_TIME"),
+                datetime_column("RECOVER_TIME"),
                bigint_column("TRANSACTION_COUNTER"),
                string_column("VERSION"),
                string_column("ROW_FORMAT"),
@@ -217,9 +217,9 @@ pub(super) fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>
                bigint_column("MAX_DATA_LENGTH"),
                bigint_column("INDEX_LENGTH"),
                bigint_column("DATA_FREE"),
-                timestamp_micro_column("CREATE_TIME"),
-                timestamp_micro_column("UPDATE_TIME"),
-                timestamp_micro_column("CHECK_TIME"),
+                datetime_column("CREATE_TIME"),
+                datetime_column("UPDATE_TIME"),
+                datetime_column("CHECK_TIME"),
                string_column("CHECKSUM"),
                string_column("STATUS"),
                string_column("EXTRA"),
@@ -330,8 +330,8 @@ pub(super) fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>
                string_column("SQL_DATA_ACCESS"),
                string_column("SQL_PATH"),
                string_column("SECURITY_TYPE"),
-                timestamp_micro_column("CREATED"),
-                timestamp_micro_column("LAST_ALTERED"),
+                datetime_column("CREATED"),
+                datetime_column("LAST_ALTERED"),
                string_column("SQL_MODE"),
                string_column("ROUTINE_COMMENT"),
                string_column("DEFINER"),
@@ -383,7 +383,7 @@ pub(super) fn get_schema_columns(table_name: &str) -> (SchemaRef, Vec<VectorRef>
                string_column("ACTION_REFERENCE_NEW_TABLE"),
                string_column("ACTION_REFERENCE_OLD_ROW"),
                string_column("ACTION_REFERENCE_NEW_ROW"),
-                timestamp_micro_column("CREATED"),
+                datetime_column("CREATED"),
                string_column("SQL_MODE"),
                string_column("DEFINER"),
                string_column("CHARACTER_SET_CLIENT"),
--- a/src/catalog/src/system_schema/information_schema/key_column_usage.rs
+++ b/src/catalog/src/system_schema/information_schema/key_column_usage.rs
@@ -228,6 +228,12 @@ impl InformationSchemaKeyColumnUsageBuilder {
                let keys = &table_info.meta.primary_key_indices;
                let schema = table.schema();

+                // For compatibility, use primary key columns as inverted index columns.
+                let pk_as_inverted_index = !schema
+                    .column_schemas()
+                    .iter()
+                    .any(|c| c.has_inverted_index_key());
+
                for (idx, column) in schema.column_schemas().iter().enumerate() {
                    let mut constraints = vec![];
                    if column.is_time_index() {
@@ -245,6 +251,10 @@ impl InformationSchemaKeyColumnUsageBuilder {
                    // TODO(dimbtp): foreign key constraint not supported yet
                    if keys.contains(&idx) {
                        constraints.push(PRI_CONSTRAINT_NAME);
+
+                        if pk_as_inverted_index {
+                            constraints.push(INVERTED_INDEX_CONSTRAINT_NAME);
+                        }
                    }
                    if column.is_inverted_indexed() {
                        constraints.push(INVERTED_INDEX_CONSTRAINT_NAME);
--- a/src/catalog/src/system_schema/information_schema/partitions.rs
+++ b/src/catalog/src/system_schema/information_schema/partitions.rs
@@ -20,18 +20,17 @@ use common_catalog::consts::INFORMATION_SCHEMA_PARTITIONS_TABLE_ID;
 use common_error::ext::BoxedError;
 use common_recordbatch::adapter::RecordBatchStreamAdapter;
 use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
+use common_time::datetime::DateTime;
 use datafusion::execution::TaskContext;
 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::timestamp::TimestampMicrosecond;
 use datatypes::value::Value;
 use datatypes::vectors::{
-    ConstantVector, Int64Vector, Int64VectorBuilder, MutableVector, StringVector,
-    StringVectorBuilder, TimestampMicrosecondVector, TimestampMicrosecondVectorBuilder,
-    UInt64VectorBuilder,
+    ConstantVector, DateTimeVector, DateTimeVectorBuilder, Int64Vector, Int64VectorBuilder,
+    MutableVector, StringVector, StringVectorBuilder, UInt64VectorBuilder,
 };
 use futures::{StreamExt, TryStreamExt};
 use partition::manager::PartitionInfo;
@@ -128,21 +127,9 @@ impl InformationSchemaPartitions {
            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::timestamp_microsecond_datatype(),
-                true,
-            ),
-            ColumnSchema::new(
-                "update_time",
-                ConcreteDataType::timestamp_microsecond_datatype(),
-                true,
-            ),
-            ColumnSchema::new(
-                "check_time",
-                ConcreteDataType::timestamp_microsecond_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",
@@ -213,7 +200,7 @@ struct InformationSchemaPartitionsBuilder {
    partition_names: StringVectorBuilder,
    partition_ordinal_positions: Int64VectorBuilder,
    partition_expressions: StringVectorBuilder,
-    create_times: TimestampMicrosecondVectorBuilder,
+    create_times: DateTimeVectorBuilder,
    partition_ids: UInt64VectorBuilder,
 }

@@ -233,7 +220,7 @@ impl InformationSchemaPartitionsBuilder {
            partition_names: StringVectorBuilder::with_capacity(INIT_CAPACITY),
            partition_ordinal_positions: Int64VectorBuilder::with_capacity(INIT_CAPACITY),
            partition_expressions: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            create_times: TimestampMicrosecondVectorBuilder::with_capacity(INIT_CAPACITY),
+            create_times: DateTimeVectorBuilder::with_capacity(INIT_CAPACITY),
            partition_ids: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
        }
    }
@@ -337,7 +324,7 @@ impl InformationSchemaPartitionsBuilder {
            };

            self.partition_expressions.push(expressions.as_deref());
-            self.create_times.push(Some(TimestampMicrosecond::from(
+            self.create_times.push(Some(DateTime::from(
                table_info.meta.created_on.timestamp_millis(),
            )));
            self.partition_ids.push(Some(partition.id.as_u64()));
@@ -355,8 +342,8 @@ impl InformationSchemaPartitionsBuilder {
            Arc::new(Int64Vector::from(vec![None])),
            rows_num,
        ));
-        let null_timestampmicrosecond_vector = Arc::new(ConstantVector::new(
-            Arc::new(TimestampMicrosecondVector::from(vec![None])),
+        let null_datetime_vector = Arc::new(ConstantVector::new(
+            Arc::new(DateTimeVector::from(vec![None])),
            rows_num,
        ));
        let partition_methods = Arc::new(ConstantVector::new(
@@ -386,8 +373,8 @@ impl InformationSchemaPartitionsBuilder {
            null_i64_vector.clone(),
            Arc::new(self.create_times.finish()),
            // TODO(dennis): supports update_time
-            null_timestampmicrosecond_vector.clone(),
-            null_timestampmicrosecond_vector,
+            null_datetime_vector.clone(),
+            null_datetime_vector,
            null_i64_vector,
            null_string_vector.clone(),
            null_string_vector.clone(),
--- a/src/catalog/src/system_schema/information_schema/tables.rs
+++ b/src/catalog/src/system_schema/information_schema/tables.rs
@@ -30,8 +30,7 @@ use datatypes::prelude::{ConcreteDataType, ScalarVectorBuilder, VectorRef};
 use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
 use datatypes::value::Value;
 use datatypes::vectors::{
-    StringVectorBuilder, TimestampMicrosecondVectorBuilder, UInt32VectorBuilder,
-    UInt64VectorBuilder,
+    DateTimeVectorBuilder, StringVectorBuilder, UInt32VectorBuilder, UInt64VectorBuilder,
 };
 use futures::TryStreamExt;
 use snafu::{OptionExt, ResultExt};
@@ -106,21 +105,9 @@ impl InformationSchemaTables {
            ColumnSchema::new(TABLE_ROWS, ConcreteDataType::uint64_datatype(), true),
            ColumnSchema::new(DATA_FREE, ConcreteDataType::uint64_datatype(), true),
            ColumnSchema::new(AUTO_INCREMENT, ConcreteDataType::uint64_datatype(), true),
-            ColumnSchema::new(
-                CREATE_TIME,
-                ConcreteDataType::timestamp_microsecond_datatype(),
-                true,
-            ),
-            ColumnSchema::new(
-                UPDATE_TIME,
-                ConcreteDataType::timestamp_microsecond_datatype(),
-                true,
-            ),
-            ColumnSchema::new(
-                CHECK_TIME,
-                ConcreteDataType::timestamp_microsecond_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(TABLE_COLLATION, ConcreteDataType::string_datatype(), true),
            ColumnSchema::new(CHECKSUM, ConcreteDataType::uint64_datatype(), true),
            ColumnSchema::new(CREATE_OPTIONS, ConcreteDataType::string_datatype(), true),
@@ -195,9 +182,9 @@ struct InformationSchemaTablesBuilder {
    max_index_length: UInt64VectorBuilder,
    data_free: UInt64VectorBuilder,
    auto_increment: UInt64VectorBuilder,
-    create_time: TimestampMicrosecondVectorBuilder,
-    update_time: TimestampMicrosecondVectorBuilder,
-    check_time: TimestampMicrosecondVectorBuilder,
+    create_time: DateTimeVectorBuilder,
+    update_time: DateTimeVectorBuilder,
+    check_time: DateTimeVectorBuilder,
    table_collation: StringVectorBuilder,
    checksum: UInt64VectorBuilder,
    create_options: StringVectorBuilder,
@@ -232,9 +219,9 @@ impl InformationSchemaTablesBuilder {
            max_index_length: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
            data_free: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
            auto_increment: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
-            create_time: TimestampMicrosecondVectorBuilder::with_capacity(INIT_CAPACITY),
-            update_time: TimestampMicrosecondVectorBuilder::with_capacity(INIT_CAPACITY),
-            check_time: TimestampMicrosecondVectorBuilder::with_capacity(INIT_CAPACITY),
+            create_time: DateTimeVectorBuilder::with_capacity(INIT_CAPACITY),
+            update_time: DateTimeVectorBuilder::with_capacity(INIT_CAPACITY),
+            check_time: DateTimeVectorBuilder::with_capacity(INIT_CAPACITY),
            table_collation: StringVectorBuilder::with_capacity(INIT_CAPACITY),
            checksum: UInt64VectorBuilder::with_capacity(INIT_CAPACITY),
            create_options: StringVectorBuilder::with_capacity(INIT_CAPACITY),
--- a/src/catalog/src/system_schema/utils/tables.rs
+++ b/src/catalog/src/system_schema/utils/tables.rs
@@ -51,10 +51,10 @@ pub fn bigint_column(name: &str) -> ColumnSchema {
    )
 }

-pub fn timestamp_micro_column(name: &str) -> ColumnSchema {
+pub fn datetime_column(name: &str) -> ColumnSchema {
    ColumnSchema::new(
        str::to_lowercase(name),
-        ConcreteDataType::timestamp_microsecond_datatype(),
+        ConcreteDataType::datetime_datatype(),
        false,
    )
 }
--- a/src/cli/Cargo.toml
+++ b/src/cli/Cargo.toml
@@ -6,7 +6,6 @@ license.workspace = true

 [features]
 pg_kvbackend = ["common-meta/pg_kvbackend"]
-mysql_kvbackend = ["common-meta/mysql_kvbackend"]

 [lints]
 workspace = true
@@ -44,10 +43,6 @@ futures.workspace = true
 humantime.workspace = true
 meta-client.workspace = true
 nu-ansi-term = "0.46"
-opendal = { version = "0.51.1", features = [
-    "services-fs",
-    "services-s3",
-] }
 query.workspace = true
 rand.workspace = true
 reqwest.workspace = true
--- a/src/cli/src/bench.rs
+++ b/src/cli/src/bench.rs
@@ -23,14 +23,11 @@ use common_error::ext::BoxedError;
 use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
 use common_meta::kv_backend::etcd::EtcdStore;
 use common_meta::kv_backend::memory::MemoryKvBackend;
-#[cfg(feature = "mysql_kvbackend")]
-use common_meta::kv_backend::rds::MySqlStore;
 #[cfg(feature = "pg_kvbackend")]
-use common_meta::kv_backend::rds::PgStore;
+use common_meta::kv_backend::postgres::PgStore;
 use common_meta::peer::Peer;
 use common_meta::rpc::router::{Region, RegionRoute};
 use common_telemetry::info;
-use common_wal::options::WalOptions;
 use datatypes::data_type::ConcreteDataType;
 use datatypes::schema::{ColumnSchema, RawSchema};
 use rand::Rng;
@@ -65,9 +62,6 @@ pub struct BenchTableMetadataCommand {
    #[cfg(feature = "pg_kvbackend")]
    #[clap(long)]
    postgres_addr: Option<String>,
-    #[cfg(feature = "mysql_kvbackend")]
-    #[clap(long)]
-    mysql_addr: Option<String>,
    #[clap(long)]
    count: u32,
 }
@@ -91,16 +85,6 @@ impl BenchTableMetadataCommand {
            kv_backend
        };

-        #[cfg(feature = "mysql_kvbackend")]
-        let kv_backend = if let Some(mysql_addr) = &self.mysql_addr {
-            info!("Using mysql as kv backend");
-            MySqlStore::with_url(mysql_addr, "greptime_metakv", 128)
-                .await
-                .unwrap()
-        } else {
-            kv_backend
-        };
-
        let table_metadata_manager = Arc::new(TableMetadataManager::new(kv_backend));

        let tool = BenchTableMetadata {
@@ -200,7 +184,7 @@ fn create_region_routes(regions: Vec<RegionNumber>) -> Vec<RegionRoute> {
    region_routes
 }

-fn create_region_wal_options(regions: Vec<RegionNumber>) -> HashMap<RegionNumber, WalOptions> {
+fn create_region_wal_options(regions: Vec<RegionNumber>) -> HashMap<RegionNumber, String> {
    // TODO(niebayes): construct region wal options for benchmark.
    let _ = regions;
    HashMap::default()
--- a/src/cli/src/bench/metadata.rs
+++ b/src/cli/src/bench/metadata.rs
@@ -49,12 +49,7 @@ impl TableMetadataBencher {

                let regions: Vec<_> = (0..64).collect();
                let region_routes = create_region_routes(regions.clone());
-                let region_wal_options = create_region_wal_options(regions)
-                    .into_iter()
-                    .map(|(region_id, wal_options)| {
-                        (region_id, serde_json::to_string(&wal_options).unwrap())
-                    })
-                    .collect();
+                let region_wal_options = create_region_wal_options(regions);

                let start = Instant::now();

@@ -114,17 +109,9 @@ impl TableMetadataBencher {
                let table_info = table_info.unwrap();
                let table_route = table_route.unwrap();
                let table_id = table_info.table_info.ident.table_id;
-
-                let regions: Vec<_> = (0..64).collect();
-                let region_wal_options = create_region_wal_options(regions);
                let _ = self
                    .table_metadata_manager
-                    .delete_table_metadata(
-                        table_id,
-                        &table_info.table_name(),
-                        &table_route,
-                        &region_wal_options,
-                    )
+                    .delete_table_metadata(table_id, &table_info.table_name(), &table_route)
                    .await;
                start.elapsed()
            },
--- a/src/cli/src/error.rs
+++ b/src/cli/src/error.rs
@@ -276,24 +276,6 @@ pub enum Error {
        #[snafu(implicit)]
        location: Location,
    },
-
-    #[snafu(display("OpenDAL operator failed"))]
-    OpenDal {
-        #[snafu(implicit)]
-        location: Location,
-        #[snafu(source)]
-        error: opendal::Error,
-    },
-    #[snafu(display("S3 config need be set"))]
-    S3ConfigNotSet {
-        #[snafu(implicit)]
-        location: Location,
-    },
-    #[snafu(display("Output directory not set"))]
-    OutputDirNotSet {
-        #[snafu(implicit)]
-        location: Location,
-    },
 }

 pub type Result<T> = std::result::Result<T, Error>;
@@ -337,9 +319,6 @@ impl ErrorExt for Error {
            | Error::BuildClient { .. } => StatusCode::Unexpected,

            Error::Other { source, .. } => source.status_code(),
-            Error::OpenDal { .. } => StatusCode::Internal,
-            Error::S3ConfigNotSet { .. } => StatusCode::InvalidArguments,
-            Error::OutputDirNotSet { .. } => StatusCode::InvalidArguments,

            Error::BuildRuntime { source, .. } => source.status_code(),

--- a/src/cli/src/export.rs
+++ b/src/cli/src/export.rs
@@ -21,18 +21,15 @@ use async_trait::async_trait;
 use clap::{Parser, ValueEnum};
 use common_error::ext::BoxedError;
 use common_telemetry::{debug, error, info};
-use opendal::layers::LoggingLayer;
-use opendal::{services, Operator};
 use serde_json::Value;
 use snafu::{OptionExt, ResultExt};
+use tokio::fs::File;
+use tokio::io::{AsyncWriteExt, BufWriter};
 use tokio::sync::Semaphore;
 use tokio::time::Instant;

 use crate::database::{parse_proxy_opts, DatabaseClient};
-use crate::error::{
-    EmptyResultSnafu, Error, OpenDalSnafu, OutputDirNotSetSnafu, Result, S3ConfigNotSetSnafu,
-    SchemaNotFoundSnafu,
-};
+use crate::error::{EmptyResultSnafu, Error, FileIoSnafu, Result, SchemaNotFoundSnafu};
 use crate::{database, Tool};

 type TableReference = (String, String, String);
@@ -55,9 +52,8 @@ pub struct ExportCommand {
    addr: String,

    /// Directory to put the exported data. E.g.: /tmp/greptimedb-export
-    /// for local export.
    #[clap(long)]
-    output_dir: Option<String>,
+    output_dir: String,

    /// The name of the catalog to export.
    #[clap(long, default_value = "greptime-*")]
@@ -105,51 +101,10 @@ pub struct ExportCommand {
    /// Disable proxy server, if set, will not use any proxy.
    #[clap(long)]
    no_proxy: bool,
-
-    /// if export data to s3
-    #[clap(long)]
-    s3: bool,
-
-    /// The s3 bucket name
-    /// if s3 is set, this is required
-    #[clap(long)]
-    s3_bucket: Option<String>,
-
-    /// The s3 endpoint
-    /// if s3 is set, this is required
-    #[clap(long)]
-    s3_endpoint: Option<String>,
-
-    /// The s3 access key
-    /// if s3 is set, this is required
-    #[clap(long)]
-    s3_access_key: Option<String>,
-
-    /// The s3 secret key
-    /// if s3 is set, this is required
-    #[clap(long)]
-    s3_secret_key: Option<String>,
-
-    /// The s3 region
-    /// if s3 is set, this is required
-    #[clap(long)]
-    s3_region: Option<String>,
 }

 impl ExportCommand {
    pub async fn build(&self) -> std::result::Result<Box<dyn Tool>, BoxedError> {
-        if self.s3
-            && (self.s3_bucket.is_none()
-                || self.s3_endpoint.is_none()
-                || self.s3_access_key.is_none()
-                || self.s3_secret_key.is_none()
-                || self.s3_region.is_none())
-        {
-            return Err(BoxedError::new(S3ConfigNotSetSnafu {}.build()));
-        }
-        if !self.s3 && self.output_dir.is_none() {
-            return Err(BoxedError::new(OutputDirNotSetSnafu {}.build()));
-        }
        let (catalog, schema) =
            database::split_database(&self.database).map_err(BoxedError::new)?;
        let proxy = parse_proxy_opts(self.proxy.clone(), self.no_proxy)?;
@@ -171,43 +126,24 @@ impl ExportCommand {
            target: self.target.clone(),
            start_time: self.start_time.clone(),
            end_time: self.end_time.clone(),
-            s3: self.s3,
-            s3_bucket: self.s3_bucket.clone(),
-            s3_endpoint: self.s3_endpoint.clone(),
-            s3_access_key: self.s3_access_key.clone(),
-            s3_secret_key: self.s3_secret_key.clone(),
-            s3_region: self.s3_region.clone(),
        }))
    }
 }

-#[derive(Clone)]
 pub struct Export {
    catalog: String,
    schema: Option<String>,
    database_client: DatabaseClient,
-    output_dir: Option<String>,
+    output_dir: String,
    parallelism: usize,
    target: ExportTarget,
    start_time: Option<String>,
    end_time: Option<String>,
-    s3: bool,
-    s3_bucket: Option<String>,
-    s3_endpoint: Option<String>,
-    s3_access_key: Option<String>,
-    s3_secret_key: Option<String>,
-    s3_region: Option<String>,
 }

 impl Export {
    fn catalog_path(&self) -> PathBuf {
-        if self.s3 {
-            PathBuf::from(&self.catalog)
-        } else if let Some(dir) = &self.output_dir {
-            PathBuf::from(dir).join(&self.catalog)
-        } else {
-            unreachable!("catalog_path: output_dir must be set when not using s3")
-        }
+        PathBuf::from(&self.output_dir).join(&self.catalog)
    }

    async fn get_db_names(&self) -> Result<Vec<String>> {
@@ -364,23 +300,19 @@ impl Export {
        let timer = Instant::now();
        let db_names = self.get_db_names().await?;
        let db_count = db_names.len();
-        let operator = self.build_operator().await?;
-
        for schema in db_names {
+            let db_dir = self.catalog_path().join(format!("{schema}/"));
+            tokio::fs::create_dir_all(&db_dir)
+                .await
+                .context(FileIoSnafu)?;
+            let file = db_dir.join("create_database.sql");
+            let mut file = File::create(file).await.context(FileIoSnafu)?;
            let create_database = self
                .show_create("DATABASE", &self.catalog, &schema, None)
                .await?;
-
-            let file_path = self.get_file_path(&schema, "create_database.sql");
-            self.write_to_storage(&operator, &file_path, create_database.into_bytes())
-                .await?;
-
-            info!(
-                "Exported {}.{} database creation SQL to {}",
-                self.catalog,
-                schema,
-                self.format_output_path(&file_path)
-            );
+            file.write_all(create_database.as_bytes())
+                .await
+                .context(FileIoSnafu)?;
        }

        let elapsed = timer.elapsed();
@@ -394,267 +326,149 @@ impl Export {
        let semaphore = Arc::new(Semaphore::new(self.parallelism));
        let db_names = self.get_db_names().await?;
        let db_count = db_names.len();
-        let operator = Arc::new(self.build_operator().await?);
        let mut tasks = Vec::with_capacity(db_names.len());
-
        for schema in db_names {
            let semaphore_moved = semaphore.clone();
-            let export_self = self.clone();
-            let operator = operator.clone();
            tasks.push(async move {
                let _permit = semaphore_moved.acquire().await.unwrap();
-                let (metric_physical_tables, remaining_tables, views) = export_self
-                    .get_table_list(&export_self.catalog, &schema)
-                    .await?;
-
-                // Create directory if needed for file system storage
-                if !export_self.s3 {
-                    let db_dir = format!("{}/{}/", export_self.catalog, schema);
-                    operator.create_dir(&db_dir).await.context(OpenDalSnafu)?;
+                let (metric_physical_tables, remaining_tables, views) =
+                    self.get_table_list(&self.catalog, &schema).await?;
+                let table_count =
+                    metric_physical_tables.len() + remaining_tables.len() + views.len();
+                let db_dir = self.catalog_path().join(format!("{schema}/"));
+                tokio::fs::create_dir_all(&db_dir)
+                    .await
+                    .context(FileIoSnafu)?;
+                let file = db_dir.join("create_tables.sql");
+                let mut file = File::create(file).await.context(FileIoSnafu)?;
+                for (c, s, t) in metric_physical_tables.into_iter().chain(remaining_tables) {
+                    let create_table = self.show_create("TABLE", &c, &s, Some(&t)).await?;
+                    file.write_all(create_table.as_bytes())
+                        .await
+                        .context(FileIoSnafu)?;
                }
-
-                let file_path = export_self.get_file_path(&schema, "create_tables.sql");
-                let mut content = Vec::new();
-
-                // Add table creation SQL
-                for (c, s, t) in metric_physical_tables.iter().chain(&remaining_tables) {
-                    let create_table = export_self.show_create("TABLE", c, s, Some(t)).await?;
-                    content.extend_from_slice(create_table.as_bytes());
+                for (c, s, v) in views {
+                    let create_view = self.show_create("VIEW", &c, &s, Some(&v)).await?;
+                    file.write_all(create_view.as_bytes())
+                        .await
+                        .context(FileIoSnafu)?;
                }

-                // Add view creation SQL
-                for (c, s, v) in &views {
-                    let create_view = export_self.show_create("VIEW", c, s, Some(v)).await?;
-                    content.extend_from_slice(create_view.as_bytes());
-                }
-
-                // Write to storage
-                export_self
-                    .write_to_storage(&operator, &file_path, content)
-                    .await?;
-
                info!(
-                    "Finished exporting {}.{schema} with {} table schemas to path: {}",
-                    export_self.catalog,
-                    metric_physical_tables.len() + remaining_tables.len() + views.len(),
-                    export_self.format_output_path(&file_path)
+                    "Finished exporting {}.{schema} with {table_count} table schemas to path: {}",
+                    self.catalog,
+                    db_dir.to_string_lossy()
                );

                Ok::<(), Error>(())
            });
        }

-        let success = self.execute_tasks(tasks).await;
+        let success = futures::future::join_all(tasks)
+            .await
+            .into_iter()
+            .filter(|r| match r {
+                Ok(_) => true,
+                Err(e) => {
+                    error!(e; "export schema job failed");
+                    false
+                }
+            })
+            .count();
+
        let elapsed = timer.elapsed();
        info!("Success {success}/{db_count} jobs, cost: {elapsed:?}");

        Ok(())
    }

-    async fn build_operator(&self) -> Result<Operator> {
-        if self.s3 {
-            self.build_s3_operator().await
-        } else {
-            self.build_fs_operator().await
-        }
-    }
-
-    async fn build_s3_operator(&self) -> Result<Operator> {
-        let mut builder = services::S3::default().root("").bucket(
-            self.s3_bucket
-                .as_ref()
-                .expect("s3_bucket must be provided when s3 is enabled"),
-        );
-
-        if let Some(endpoint) = self.s3_endpoint.as_ref() {
-            builder = builder.endpoint(endpoint);
-        }
-
-        if let Some(region) = self.s3_region.as_ref() {
-            builder = builder.region(region);
-        }
-
-        if let Some(key_id) = self.s3_access_key.as_ref() {
-            builder = builder.access_key_id(key_id);
-        }
-
-        if let Some(secret_key) = self.s3_secret_key.as_ref() {
-            builder = builder.secret_access_key(secret_key);
-        }
-
-        let op = Operator::new(builder)
-            .context(OpenDalSnafu)?
-            .layer(LoggingLayer::default())
-            .finish();
-        Ok(op)
-    }
-
-    async fn build_fs_operator(&self) -> Result<Operator> {
-        let root = self
-            .output_dir
-            .as_ref()
-            .context(OutputDirNotSetSnafu)?
-            .clone();
-        let op = Operator::new(services::Fs::default().root(&root))
-            .context(OpenDalSnafu)?
-            .layer(LoggingLayer::default())
-            .finish();
-        Ok(op)
-    }
-
    async fn export_database_data(&self) -> Result<()> {
        let timer = Instant::now();
        let semaphore = Arc::new(Semaphore::new(self.parallelism));
        let db_names = self.get_db_names().await?;
        let db_count = db_names.len();
        let mut tasks = Vec::with_capacity(db_count);
-        let operator = Arc::new(self.build_operator().await?);
-        let with_options = build_with_options(&self.start_time, &self.end_time);
-
        for schema in db_names {
            let semaphore_moved = semaphore.clone();
-            let export_self = self.clone();
-            let with_options_clone = with_options.clone();
-            let operator = operator.clone();
-
            tasks.push(async move {
                let _permit = semaphore_moved.acquire().await.unwrap();
+                let db_dir = self.catalog_path().join(format!("{schema}/"));
+                tokio::fs::create_dir_all(&db_dir)
+                    .await
+                    .context(FileIoSnafu)?;

-                // Create directory if not using S3
-                if !export_self.s3 {
-                    let db_dir = format!("{}/{}/", export_self.catalog, schema);
-                    operator.create_dir(&db_dir).await.context(OpenDalSnafu)?;
-                }
+                let with_options = match (&self.start_time, &self.end_time) {
+                    (Some(start_time), Some(end_time)) => {
+                        format!(
+                            "WITH (FORMAT='parquet', start_time='{}', end_time='{}')",
+                            start_time, end_time
+                        )
+                    }
+                    (Some(start_time), None) => {
+                        format!("WITH (FORMAT='parquet', start_time='{}')", start_time)
+                    }
+                    (None, Some(end_time)) => {
+                        format!("WITH (FORMAT='parquet', end_time='{}')", end_time)
+                    }
+                    (None, None) => "WITH (FORMAT='parquet')".to_string(),
+                };

-                let (path, connection_part) = export_self.get_storage_params(&schema);
-
-                // Execute COPY DATABASE TO command
                let sql = format!(
-                    r#"COPY DATABASE "{}"."{}" TO '{}' WITH ({}){};"#,
-                    export_self.catalog, schema, path, with_options_clone, connection_part
-                );
-                info!("Executing sql: {sql}");
-                export_self.database_client.sql_in_public(&sql).await?;
-                info!(
-                    "Finished exporting {}.{} data to {}",
-                    export_self.catalog, schema, path
-                );
-
-                // Create copy_from.sql file
-                let copy_database_from_sql = format!(
-                    r#"COPY DATABASE "{}"."{}" FROM '{}' WITH ({}){};"#,
-                    export_self.catalog, schema, path, with_options_clone, connection_part
-                );
-
-                let copy_from_path = export_self.get_file_path(&schema, "copy_from.sql");
-                export_self
-                    .write_to_storage(
-                        &operator,
-                        &copy_from_path,
-                        copy_database_from_sql.into_bytes(),
-                    )
-                    .await?;
-
-                info!(
-                    "Finished exporting {}.{} copy_from.sql to {}",
-                    export_self.catalog,
+                    r#"COPY DATABASE "{}"."{}" TO '{}' {};"#,
+                    self.catalog,
                    schema,
-                    export_self.format_output_path(&copy_from_path)
+                    db_dir.to_str().unwrap(),
+                    with_options
                );

+                info!("Executing sql: {sql}");
+
+                self.database_client.sql_in_public(&sql).await?;
+
+                info!(
+                    "Finished exporting {}.{schema} data into path: {}",
+                    self.catalog,
+                    db_dir.to_string_lossy()
+                );
+
+                // The export copy from sql
+                let copy_from_file = db_dir.join("copy_from.sql");
+                let mut writer =
+                    BufWriter::new(File::create(copy_from_file).await.context(FileIoSnafu)?);
+                let copy_database_from_sql = format!(
+                    r#"COPY DATABASE "{}"."{}" FROM '{}' WITH (FORMAT='parquet');"#,
+                    self.catalog,
+                    schema,
+                    db_dir.to_str().unwrap()
+                );
+                writer
+                    .write(copy_database_from_sql.as_bytes())
+                    .await
+                    .context(FileIoSnafu)?;
+                writer.flush().await.context(FileIoSnafu)?;
+
+                info!("Finished exporting {}.{schema} copy_from.sql", self.catalog);
+
                Ok::<(), Error>(())
-            });
+            })
        }

-        let success = self.execute_tasks(tasks).await;
-        let elapsed = timer.elapsed();
-        info!("Success {success}/{db_count} jobs, costs: {elapsed:?}");
-
-        Ok(())
-    }
-
-    fn get_file_path(&self, schema: &str, file_name: &str) -> String {
-        format!("{}/{}/{}", self.catalog, schema, file_name)
-    }
-
-    fn format_output_path(&self, file_path: &str) -> String {
-        if self.s3 {
-            format!(
-                "s3://{}/{}",
-                self.s3_bucket.as_ref().unwrap_or(&String::new()),
-                file_path
-            )
-        } else {
-            format!(
-                "{}/{}",
-                self.output_dir.as_ref().unwrap_or(&String::new()),
-                file_path
-            )
-        }
-    }
-
-    async fn write_to_storage(
-        &self,
-        op: &Operator,
-        file_path: &str,
-        content: Vec<u8>,
-    ) -> Result<()> {
-        op.write(file_path, content).await.context(OpenDalSnafu)
-    }
-
-    fn get_storage_params(&self, schema: &str) -> (String, String) {
-        if self.s3 {
-            let s3_path = format!(
-                "s3://{}/{}/{}/",
-                // Safety: s3_bucket is required when s3 is enabled
-                self.s3_bucket.as_ref().unwrap(),
-                self.catalog,
-                schema
-            );
-
-            // endpoint is optional
-            let endpoint_option = if let Some(endpoint) = self.s3_endpoint.as_ref() {
-                format!(", ENDPOINT='{}'", endpoint)
-            } else {
-                String::new()
-            };
-
-            // Safety: All s3 options are required
-            let connection_options = format!(
-                "ACCESS_KEY_ID='{}', SECRET_ACCESS_KEY='{}', REGION='{}'{}",
-                self.s3_access_key.as_ref().unwrap(),
-                self.s3_secret_key.as_ref().unwrap(),
-                self.s3_region.as_ref().unwrap(),
-                endpoint_option
-            );
-
-            (s3_path, format!(" CONNECTION ({})", connection_options))
-        } else {
-            (
-                self.catalog_path()
-                    .join(format!("{schema}/"))
-                    .to_string_lossy()
-                    .to_string(),
-                String::new(),
-            )
-        }
-    }
-
-    async fn execute_tasks(
-        &self,
-        tasks: Vec<impl std::future::Future<Output = Result<()>>>,
-    ) -> usize {
-        futures::future::join_all(tasks)
+        let success = futures::future::join_all(tasks)
            .await
            .into_iter()
            .filter(|r| match r {
                Ok(_) => true,
                Err(e) => {
-                    error!(e; "export job failed");
+                    error!(e; "export database job failed");
                    false
                }
            })
-            .count()
+            .count();
+        let elapsed = timer.elapsed();
+
+        info!("Success {success}/{db_count} jobs, costs: {elapsed:?}");
+
+        Ok(())
    }
 }

@@ -679,15 +493,3 @@ impl Tool for Export {
        }
    }
 }
-
-/// Builds the WITH options string for SQL commands, assuming consistent syntax across S3 and local exports.
-fn build_with_options(start_time: &Option<String>, end_time: &Option<String>) -> String {
-    let mut options = vec!["format = 'parquet'".to_string()];
-    if let Some(start) = start_time {
-        options.push(format!("start_time = '{}'", start));
-    }
-    if let Some(end) = end_time {
-        options.push(format!("end_time = '{}'", end));
-    }
-    options.join(", ")
-}
--- a/src/cmd/src/datanode.rs
+++ b/src/cmd/src/datanode.rs
@@ -126,14 +126,10 @@ impl SubCommand {
 struct StartCommand {
    #[clap(long)]
    node_id: Option<u64>,
-    /// The address to bind the gRPC server.
-    #[clap(long, alias = "rpc-addr")]
-    rpc_bind_addr: Option<String>,
-    /// The address advertised to the metasrv, and used for connections from outside the host.
-    /// If left empty or unset, the server will automatically use the IP address of the first network interface
-    /// on the host, with the same port number as the one specified in `rpc_bind_addr`.
-    #[clap(long, alias = "rpc-hostname")]
-    rpc_server_addr: Option<String>,
+    #[clap(long)]
+    rpc_addr: Option<String>,
+    #[clap(long)]
+    rpc_hostname: Option<String>,
    #[clap(long, value_delimiter = ',', num_args = 1..)]
    metasrv_addrs: Option<Vec<String>>,
    #[clap(short, long)]
@@ -185,18 +181,18 @@ impl StartCommand {
            tokio_console_addr: global_options.tokio_console_addr.clone(),
        };

-        if let Some(addr) = &self.rpc_bind_addr {
-            opts.grpc.bind_addr.clone_from(addr);
+        if let Some(addr) = &self.rpc_addr {
+            opts.grpc.addr.clone_from(addr);
        } else if let Some(addr) = &opts.rpc_addr {
            warn!("Use the deprecated attribute `DatanodeOptions.rpc_addr`, please use `grpc.addr` instead.");
-            opts.grpc.bind_addr.clone_from(addr);
+            opts.grpc.addr.clone_from(addr);
        }

-        if let Some(server_addr) = &self.rpc_server_addr {
-            opts.grpc.server_addr.clone_from(server_addr);
-        } else if let Some(server_addr) = &opts.rpc_hostname {
+        if let Some(hostname) = &self.rpc_hostname {
+            opts.grpc.hostname.clone_from(hostname);
+        } else if let Some(hostname) = &opts.rpc_hostname {
            warn!("Use the deprecated attribute `DatanodeOptions.rpc_hostname`, please use `grpc.hostname` instead.");
-            opts.grpc.server_addr.clone_from(server_addr);
+            opts.grpc.hostname.clone_from(hostname);
        }

        if let Some(runtime_size) = opts.rpc_runtime_size {
@@ -281,12 +277,13 @@ impl StartCommand {

        let plugin_opts = opts.plugins;
        let mut opts = opts.component;
-        opts.grpc.detect_server_addr();
+        opts.grpc.detect_hostname();
        let mut plugins = Plugins::new();
        plugins::setup_datanode_plugins(&mut plugins, &plugin_opts, &opts)
            .await
            .context(StartDatanodeSnafu)?;

+        let cluster_id = 0; // TODO(hl): read from config
        let member_id = opts
            .node_id
            .context(MissingConfigSnafu { msg: "'node_id'" })?;
@@ -295,10 +292,13 @@ impl StartCommand {
            msg: "'meta_client_options'",
        })?;

-        let meta_client =
-            meta_client::create_meta_client(MetaClientType::Datanode { member_id }, meta_config)
-                .await
-                .context(MetaClientInitSnafu)?;
+        let meta_client = meta_client::create_meta_client(
+            cluster_id,
+            MetaClientType::Datanode { member_id },
+            meta_config,
+        )
+        .await
+        .context(MetaClientInitSnafu)?;

        let meta_backend = Arc::new(MetaKvBackend {
            client: meta_client.clone(),
@@ -357,8 +357,8 @@ mod tests {
            rpc_addr = "127.0.0.1:4001"
            rpc_hostname = "192.168.0.1"
            [grpc]
-            bind_addr = "127.0.0.1:3001"
-            server_addr = "127.0.0.1"
+            addr = "127.0.0.1:3001"
+            hostname = "127.0.0.1"
            runtime_size = 8
        "#;
        write!(file, "{}", toml_str).unwrap();
@@ -369,8 +369,8 @@ mod tests {
        };

        let options = cmd.load_options(&Default::default()).unwrap().component;
-        assert_eq!("127.0.0.1:4001".to_string(), options.grpc.bind_addr);
-        assert_eq!("192.168.0.1".to_string(), options.grpc.server_addr);
+        assert_eq!("127.0.0.1:4001".to_string(), options.grpc.addr);
+        assert_eq!("192.168.0.1".to_string(), options.grpc.hostname);
    }

    #[test]
@@ -431,7 +431,7 @@ mod tests {

        let options = cmd.load_options(&Default::default()).unwrap().component;

-        assert_eq!("127.0.0.1:3001".to_string(), options.grpc.bind_addr);
+        assert_eq!("127.0.0.1:3001".to_string(), options.grpc.addr);
        assert_eq!(Some(42), options.node_id);

        let DatanodeWalConfig::RaftEngine(raft_engine_config) = options.wal else {
@@ -645,7 +645,7 @@ mod tests {
                    opts.http.addr,
                    DatanodeOptions::default().component.http.addr
                );
-                assert_eq!(opts.grpc.server_addr, "10.103.174.219");
+                assert_eq!(opts.grpc.hostname, "10.103.174.219");
            },
        );
    }
--- a/src/cmd/src/flownode.rs
+++ b/src/cmd/src/flownode.rs
@@ -129,13 +129,11 @@ struct StartCommand {
    #[clap(long)]
    node_id: Option<u64>,
    /// Bind address for the gRPC server.
-    #[clap(long, alias = "rpc-addr")]
-    rpc_bind_addr: Option<String>,
-    /// The address advertised to the metasrv, and used for connections from outside the host.
-    /// If left empty or unset, the server will automatically use the IP address of the first network interface
-    /// on the host, with the same port number as the one specified in `rpc_bind_addr`.
-    #[clap(long, alias = "rpc-hostname")]
-    rpc_server_addr: Option<String>,
+    #[clap(long)]
+    rpc_addr: Option<String>,
+    /// Hostname for the gRPC server.
+    #[clap(long)]
+    rpc_hostname: Option<String>,
    /// Metasrv address list;
    #[clap(long, value_delimiter = ',', num_args = 1..)]
    metasrv_addrs: Option<Vec<String>>,
@@ -186,12 +184,12 @@ impl StartCommand {
            tokio_console_addr: global_options.tokio_console_addr.clone(),
        };

-        if let Some(addr) = &self.rpc_bind_addr {
-            opts.grpc.bind_addr.clone_from(addr);
+        if let Some(addr) = &self.rpc_addr {
+            opts.grpc.addr.clone_from(addr);
        }

-        if let Some(server_addr) = &self.rpc_server_addr {
-            opts.grpc.server_addr.clone_from(server_addr);
+        if let Some(hostname) = &self.rpc_hostname {
+            opts.grpc.hostname.clone_from(hostname);
        }

        if let Some(node_id) = self.node_id {
@@ -239,7 +237,10 @@ impl StartCommand {
        info!("Flownode options: {:#?}", opts);

        let mut opts = opts.component;
-        opts.grpc.detect_server_addr();
+        opts.grpc.detect_hostname();
+
+        // TODO(discord9): make it not optionale after cluster id is required
+        let cluster_id = opts.cluster_id.unwrap_or(0);

        let member_id = opts
            .node_id
@@ -249,10 +250,13 @@ impl StartCommand {
            msg: "'meta_client_options'",
        })?;

-        let meta_client =
-            meta_client::create_meta_client(MetaClientType::Flownode { member_id }, meta_config)
-                .await
-                .context(MetaClientInitSnafu)?;
+        let meta_client = meta_client::create_meta_client(
+            cluster_id,
+            MetaClientType::Flownode { member_id },
+            meta_config,
+        )
+        .await
+        .context(MetaClientInitSnafu)?;

        let cache_max_capacity = meta_config.metadata_cache_max_capacity;
        let cache_ttl = meta_config.metadata_cache_ttl;
--- a/src/cmd/src/frontend.rs
+++ b/src/cmd/src/frontend.rs
@@ -136,19 +136,13 @@ impl SubCommand {

 #[derive(Debug, Default, Parser)]
 pub struct StartCommand {
-    /// The address to bind the gRPC server.
-    #[clap(long, alias = "rpc-addr")]
-    rpc_bind_addr: Option<String>,
-    /// The address advertised to the metasrv, and used for connections from outside the host.
-    /// If left empty or unset, the server will automatically use the IP address of the first network interface
-    /// on the host, with the same port number as the one specified in `rpc_bind_addr`.
-    #[clap(long, alias = "rpc-hostname")]
-    rpc_server_addr: Option<String>,
    #[clap(long)]
    http_addr: Option<String>,
    #[clap(long)]
    http_timeout: Option<u64>,
    #[clap(long)]
+    rpc_addr: Option<String>,
+    #[clap(long)]
    mysql_addr: Option<String>,
    #[clap(long)]
    postgres_addr: Option<String>,
@@ -224,15 +218,11 @@ impl StartCommand {
            opts.http.disable_dashboard = disable_dashboard;
        }

-        if let Some(addr) = &self.rpc_bind_addr {
-            opts.grpc.bind_addr.clone_from(addr);
+        if let Some(addr) = &self.rpc_addr {
+            opts.grpc.addr.clone_from(addr);
            opts.grpc.tls = tls_opts.clone();
        }

-        if let Some(addr) = &self.rpc_server_addr {
-            opts.grpc.server_addr.clone_from(addr);
-        }
-
        if let Some(addr) = &self.mysql_addr {
            opts.mysql.enable = true;
            opts.mysql.addr.clone_from(addr);
@@ -279,7 +269,7 @@ impl StartCommand {

        let plugin_opts = opts.plugins;
        let mut opts = opts.component;
-        opts.grpc.detect_server_addr();
+        opts.grpc.detect_hostname();
        let mut plugins = Plugins::new();
        plugins::setup_frontend_plugins(&mut plugins, &plugin_opts, &opts)
            .await
@@ -295,10 +285,14 @@ impl StartCommand {
        let cache_ttl = meta_client_options.metadata_cache_ttl;
        let cache_tti = meta_client_options.metadata_cache_tti;

-        let meta_client =
-            meta_client::create_meta_client(MetaClientType::Frontend, meta_client_options)
-                .await
-                .context(MetaClientInitSnafu)?;
+        let cluster_id = 0; // (TODO: jeremy): It is currently a reserved field and has not been enabled.
+        let meta_client = meta_client::create_meta_client(
+            cluster_id,
+            MetaClientType::Frontend,
+            meta_client_options,
+        )
+        .await
+        .context(MetaClientInitSnafu)?;

        // TODO(discord9): add helper function to ease the creation of cache registry&such
        let cached_meta_backend =
@@ -419,7 +413,7 @@ mod tests {

        let default_opts = FrontendOptions::default().component;

-        assert_eq!(opts.grpc.bind_addr, default_opts.grpc.bind_addr);
+        assert_eq!(opts.grpc.addr, default_opts.grpc.addr);
        assert!(opts.mysql.enable);
        assert_eq!(opts.mysql.runtime_size, default_opts.mysql.runtime_size);
        assert!(opts.postgres.enable);
@@ -610,7 +604,7 @@ mod tests {
                assert_eq!(fe_opts.http.addr, "127.0.0.1:14000");

                // Should be default value.
-                assert_eq!(fe_opts.grpc.bind_addr, GrpcOptions::default().bind_addr);
+                assert_eq!(fe_opts.grpc.addr, GrpcOptions::default().addr);
            },
        );
    }
--- a/src/cmd/src/metasrv.rs
+++ b/src/cmd/src/metasrv.rs
@@ -42,7 +42,7 @@ pub struct Instance {
 }

 impl Instance {
-    pub fn new(instance: MetasrvInstance, guard: Vec<WorkerGuard>) -> Self {
+    fn new(instance: MetasrvInstance, guard: Vec<WorkerGuard>) -> Self {
        Self {
            instance,
            _guard: guard,
@@ -133,15 +133,11 @@ impl SubCommand {

 #[derive(Debug, Default, Parser)]
 struct StartCommand {
-    /// The address to bind the gRPC server.
-    #[clap(long, alias = "bind-addr")]
-    rpc_bind_addr: Option<String>,
-    /// The communication server address for the frontend and datanode to connect to metasrv.
-    /// If left empty or unset, the server will automatically use the IP address of the first network interface
-    /// on the host, with the same port number as the one specified in `rpc_bind_addr`.
-    #[clap(long, alias = "server-addr")]
-    rpc_server_addr: Option<String>,
-    #[clap(long, alias = "store-addr", value_delimiter = ',', num_args = 1..)]
+    #[clap(long)]
+    bind_addr: Option<String>,
+    #[clap(long)]
+    server_addr: Option<String>,
+    #[clap(long, aliases = ["store-addr"], value_delimiter = ',', num_args = 1..)]
    store_addrs: Option<Vec<String>>,
    #[clap(short, long)]
    config_file: Option<String>,
@@ -205,11 +201,11 @@ impl StartCommand {
            tokio_console_addr: global_options.tokio_console_addr.clone(),
        };

-        if let Some(addr) = &self.rpc_bind_addr {
+        if let Some(addr) = &self.bind_addr {
            opts.bind_addr.clone_from(addr);
        }

-        if let Some(addr) = &self.rpc_server_addr {
+        if let Some(addr) = &self.server_addr {
            opts.server_addr.clone_from(addr);
        }

@@ -273,13 +269,11 @@ impl StartCommand {
        log_versions(version(), short_version(), APP_NAME);

        info!("Metasrv start command: {:#?}", self);
+        info!("Metasrv options: {:#?}", opts);

        let plugin_opts = opts.plugins;
        let mut opts = opts.component;
        opts.detect_server_addr();
-
-        info!("Metasrv options: {:#?}", opts);
-
        let mut plugins = Plugins::new();
        plugins::setup_metasrv_plugins(&mut plugins, &plugin_opts, &opts)
            .await
@@ -312,8 +306,8 @@ mod tests {
    #[test]
    fn test_read_from_cmd() {
        let cmd = StartCommand {
-            rpc_bind_addr: Some("127.0.0.1:3002".to_string()),
-            rpc_server_addr: Some("127.0.0.1:3002".to_string()),
+            bind_addr: Some("127.0.0.1:3002".to_string()),
+            server_addr: Some("127.0.0.1:3002".to_string()),
            store_addrs: Some(vec!["127.0.0.1:2380".to_string()]),
            selector: Some("LoadBased".to_string()),
            ..Default::default()
@@ -387,8 +381,8 @@ mod tests {
    #[test]
    fn test_load_log_options_from_cli() {
        let cmd = StartCommand {
-            rpc_bind_addr: Some("127.0.0.1:3002".to_string()),
-            rpc_server_addr: Some("127.0.0.1:3002".to_string()),
+            bind_addr: Some("127.0.0.1:3002".to_string()),
+            server_addr: Some("127.0.0.1:3002".to_string()),
            store_addrs: Some(vec!["127.0.0.1:2380".to_string()]),
            selector: Some("LoadBased".to_string()),
            ..Default::default()
--- a/src/cmd/src/standalone.rs
+++ b/src/cmd/src/standalone.rs
@@ -60,8 +60,7 @@ use frontend::instance::builder::FrontendBuilder;
 use frontend::instance::{FrontendInstance, Instance as FeInstance, StandaloneDatanodeManager};
 use frontend::server::Services;
 use frontend::service_config::{
-    InfluxdbOptions, JaegerOptions, MysqlOptions, OpentsdbOptions, PostgresOptions,
-    PromStoreOptions,
+    InfluxdbOptions, MysqlOptions, OpentsdbOptions, PostgresOptions, PromStoreOptions,
 };
 use meta_srv::metasrv::{FLOW_ID_SEQ, TABLE_ID_SEQ};
 use mito2::config::MitoConfig;
@@ -141,7 +140,6 @@ pub struct StandaloneOptions {
    pub postgres: PostgresOptions,
    pub opentsdb: OpentsdbOptions,
    pub influxdb: InfluxdbOptions,
-    pub jaeger: JaegerOptions,
    pub prom_store: PromStoreOptions,
    pub wal: DatanodeWalConfig,
    pub storage: StorageConfig,
@@ -171,7 +169,6 @@ impl Default for StandaloneOptions {
            postgres: PostgresOptions::default(),
            opentsdb: OpentsdbOptions::default(),
            influxdb: InfluxdbOptions::default(),
-            jaeger: JaegerOptions::default(),
            prom_store: PromStoreOptions::default(),
            wal: DatanodeWalConfig::default(),
            storage: StorageConfig::default(),
@@ -220,7 +217,6 @@ impl StandaloneOptions {
            postgres: cloned_opts.postgres,
            opentsdb: cloned_opts.opentsdb,
            influxdb: cloned_opts.influxdb,
-            jaeger: cloned_opts.jaeger,
            prom_store: cloned_opts.prom_store,
            meta_client: None,
            logging: cloned_opts.logging,
@@ -333,8 +329,8 @@ impl App for Instance {
 pub struct StartCommand {
    #[clap(long)]
    http_addr: Option<String>,
-    #[clap(long, alias = "rpc-addr")]
-    rpc_bind_addr: Option<String>,
+    #[clap(long)]
+    rpc_addr: Option<String>,
    #[clap(long)]
    mysql_addr: Option<String>,
    #[clap(long)]
@@ -411,9 +407,9 @@ impl StartCommand {
            opts.storage.data_home.clone_from(data_home);
        }

-        if let Some(addr) = &self.rpc_bind_addr {
+        if let Some(addr) = &self.rpc_addr {
            // frontend grpc addr conflict with datanode default grpc addr
-            let datanode_grpc_addr = DatanodeOptions::default().grpc.bind_addr;
+            let datanode_grpc_addr = DatanodeOptions::default().grpc.addr;
            if addr.eq(&datanode_grpc_addr) {
                return IllegalConfigSnafu {
                    msg: format!(
@@ -421,7 +417,7 @@ impl StartCommand {
                    ),
                }.fail();
            }
-            opts.grpc.bind_addr.clone_from(addr)
+            opts.grpc.addr.clone_from(addr)
        }

        if let Some(addr) = &self.mysql_addr {
@@ -468,7 +464,7 @@ impl StartCommand {
        let mut plugins = Plugins::new();
        let plugin_opts = opts.plugins;
        let mut opts = opts.component;
-        opts.grpc.detect_server_addr();
+        opts.grpc.detect_hostname();
        let fe_opts = opts.frontend_options();
        let dn_opts = opts.datanode_options();

@@ -490,8 +486,8 @@ impl StartCommand {
        let metadata_dir = metadata_store_dir(data_home);
        let (kv_backend, procedure_manager) = FeInstance::try_build_standalone_components(
            metadata_dir,
-            opts.metadata_store,
-            opts.procedure,
+            opts.metadata_store.clone(),
+            opts.procedure.clone(),
        )
        .await
        .context(StartFrontendSnafu)?;
@@ -911,7 +907,7 @@ mod tests {
        assert_eq!("127.0.0.1:4000".to_string(), fe_opts.http.addr);
        assert_eq!(Duration::from_secs(33), fe_opts.http.timeout);
        assert_eq!(ReadableSize::mb(128), fe_opts.http.body_limit);
-        assert_eq!("127.0.0.1:4001".to_string(), fe_opts.grpc.bind_addr);
+        assert_eq!("127.0.0.1:4001".to_string(), fe_opts.grpc.addr);
        assert!(fe_opts.mysql.enable);
        assert_eq!("127.0.0.1:4002", fe_opts.mysql.addr);
        assert_eq!(2, fe_opts.mysql.runtime_size);
@@ -1041,7 +1037,7 @@ mod tests {
                assert_eq!(ReadableSize::mb(64), fe_opts.http.body_limit);

                // Should be default value.
-                assert_eq!(fe_opts.grpc.bind_addr, GrpcOptions::default().bind_addr);
+                assert_eq!(fe_opts.grpc.addr, GrpcOptions::default().addr);
            },
        );
    }
--- a/src/cmd/tests/cli.rs
+++ b/src/cmd/tests/cli.rs
@@ -63,7 +63,7 @@ mod tests {
            .args([
                "datanode",
                "start",
-                "--rpc-bind-addr=0.0.0.0:4321",
+                "--rpc-addr=0.0.0.0:4321",
                "--node-id=1",
                &format!("--data-home={}", data_home.path().display()),
                &format!("--wal-dir={}", wal_dir.path().display()),
@@ -80,7 +80,7 @@ mod tests {
            "--log-level=off",
            "cli",
            "attach",
-            "--grpc-bind-addr=0.0.0.0:4321",
+            "--grpc-addr=0.0.0.0:4321",
            // history commands can sneaky into stdout and mess up our tests, so disable it
            "--disable-helper",
        ]);
--- a/src/cmd/tests/load_config_test.rs
+++ b/src/cmd/tests/load_config_test.rs
@@ -17,6 +17,9 @@ use std::time::Duration;
 use cmd::options::GreptimeOptions;
 use cmd::standalone::StandaloneOptions;
 use common_config::Configurable;
+use common_grpc::channel_manager::{
+    DEFAULT_MAX_GRPC_RECV_MESSAGE_SIZE, DEFAULT_MAX_GRPC_SEND_MESSAGE_SIZE,
+};
 use common_options::datanode::{ClientOptions, DatanodeClientOptions};
 use common_telemetry::logging::{LoggingOptions, SlowQueryOptions, DEFAULT_OTLP_ENDPOINT};
 use common_wal::config::raft_engine::RaftEngineConfig;
@@ -88,8 +91,13 @@ fn test_load_datanode_example_config() {
                ..Default::default()
            },
            grpc: GrpcOptions::default()
-                .with_bind_addr("127.0.0.1:3001")
-                .with_server_addr("127.0.0.1:3001"),
+                .with_addr("127.0.0.1:3001")
+                .with_hostname("127.0.0.1:3001"),
+            rpc_addr: Some("127.0.0.1:3001".to_string()),
+            rpc_hostname: Some("127.0.0.1".to_string()),
+            rpc_runtime_size: Some(8),
+            rpc_max_recv_message_size: Some(DEFAULT_MAX_GRPC_RECV_MESSAGE_SIZE),
+            rpc_max_send_message_size: Some(DEFAULT_MAX_GRPC_SEND_MESSAGE_SIZE),
            ..Default::default()
        },
        ..Default::default()
@@ -136,9 +144,7 @@ fn test_load_frontend_example_config() {
                remote_write: Some(Default::default()),
                ..Default::default()
            },
-            grpc: GrpcOptions::default()
-                .with_bind_addr("127.0.0.1:4001")
-                .with_server_addr("127.0.0.1:4001"),
+            grpc: GrpcOptions::default().with_hostname("127.0.0.1:4001"),
            http: HttpOptions {
                cors_allowed_origins: vec!["https://example.com".to_string()],
                ..Default::default()
--- a/src/common/base/Cargo.toml
+++ b/src/common/base/Cargo.toml
@@ -18,7 +18,7 @@ bytes.workspace = true
 common-error.workspace = true
 common-macro.workspace = true
 futures.workspace = true
-paste.workspace = true
+paste = "1.0"
 pin-project.workspace = true
 rand.workspace = true
 serde = { version = "1.0", features = ["derive"] }
--- a/src/common/catalog/src/consts.rs
+++ b/src/common/catalog/src/consts.rs
@@ -130,10 +130,3 @@ pub const SEMANTIC_TYPE_TIME_INDEX: &str = "TIMESTAMP";
 pub fn is_readonly_schema(schema: &str) -> bool {
    matches!(schema, INFORMATION_SCHEMA_NAME)
 }
-
-// ---- special table and fields ----
-pub const TRACE_ID_COLUMN: &str = "trace_id";
-pub const SPAN_ID_COLUMN: &str = "span_id";
-pub const SPAN_NAME_COLUMN: &str = "span_name";
-pub const PARENT_SPAN_ID_COLUMN: &str = "parent_span_id";
-// ---- End of special table and fields ----
--- a/src/common/config/Cargo.toml
+++ b/src/common/config/Cargo.toml
@@ -12,11 +12,9 @@ common-base.workspace = true
 common-error.workspace = true
 common-macro.workspace = true
 config.workspace = true
-humantime-serde.workspace = true
 num_cpus.workspace = true
 serde.workspace = true
 serde_json.workspace = true
-serde_with.workspace = true
 snafu.workspace = true
 sysinfo.workspace = true
 toml.workspace = true
--- a/src/common/config/src/lib.rs
+++ b/src/common/config/src/lib.rs
@@ -16,8 +16,6 @@ pub mod config;
 pub mod error;
 pub mod utils;

-use std::time::Duration;
-
 use common_base::readable_size::ReadableSize;
 pub use config::*;
 use serde::{Deserialize, Serialize};
@@ -36,27 +34,22 @@ pub enum Mode {
    Distributed,
 }

-#[derive(Debug, Clone, Copy, Serialize, Deserialize, PartialEq, Eq)]
+#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq)]
 #[serde(default)]
 pub struct KvBackendConfig {
-    /// The size of the metadata store backend log file.
+    // Kv file size in bytes
    pub file_size: ReadableSize,
-    /// The threshold of the metadata store size to trigger a purge.
+    // Kv purge threshold in bytes
    pub purge_threshold: ReadableSize,
-    /// The interval of the metadata store to trigger a purge.
-    #[serde(with = "humantime_serde")]
-    pub purge_interval: Duration,
 }

 impl Default for KvBackendConfig {
    fn default() -> Self {
        Self {
-            // The log file size 64MB
-            file_size: ReadableSize::mb(64),
-            // The log purge threshold 256MB
-            purge_threshold: ReadableSize::mb(256),
-            // The log purge interval 1m
-            purge_interval: Duration::from_secs(60),
+            // log file size 256MB
+            file_size: ReadableSize::mb(256),
+            // purge threshold 4GB
+            purge_threshold: ReadableSize::gb(4),
        }
    }
 }
--- a/src/common/datasource/Cargo.toml
+++ b/src/common/datasource/Cargo.toml
@@ -35,7 +35,7 @@ orc-rust = { version = "0.5", default-features = false, features = [
    "async",
 ] }
 parquet.workspace = true
-paste.workspace = true
+paste = "1.0"
 rand.workspace = true
 regex = "1.7"
 serde.workspace = true
--- a/src/common/function/Cargo.toml
+++ b/src/common/function/Cargo.toml
@@ -12,12 +12,9 @@ default = ["geo"]
 geo = ["geohash", "h3o", "s2", "wkt", "geo-types", "dep:geo"]

 [dependencies]
-ahash = "0.8"
 api.workspace = true
 arc-swap = "1.0"
 async-trait.workspace = true
-bincode = "1.3"
-chrono.workspace = true
 common-base.workspace = true
 common-catalog.workspace = true
 common-error.workspace = true
@@ -29,21 +26,18 @@ common-telemetry.workspace = true
 common-time.workspace = true
 common-version.workspace = true
 datafusion.workspace = true
-datafusion-common.workspace = true
-datafusion-expr.workspace = true
 datatypes.workspace = true
 derive_more = { version = "1", default-features = false, features = ["display"] }
 geo = { version = "0.29", optional = true }
 geo-types = { version = "0.7", optional = true }
 geohash = { version = "0.13", optional = true }
 h3o = { version = "0.6", optional = true }
-hyperloglogplus = "0.4"
 jsonb.workspace = true
 nalgebra.workspace = true
 num = "0.4"
 num-traits = "0.2"
 once_cell.workspace = true
-paste.workspace = true
+paste = "1.0"
 s2 = { version = "0.0.12", optional = true }
 serde.workspace = true
 serde_json.workspace = true
@@ -53,7 +47,6 @@ sql.workspace = true
 statrs = "0.16"
 store-api.workspace = true
 table.workspace = true
-uddsketch = { git = "https://github.com/GreptimeTeam/timescaledb-toolkit.git", rev = "84828fe8fb494a6a61412a3da96517fc80f7bb20" }
 wkt = { version = "0.11", optional = true }

 [dev-dependencies]
--- a/src/common/function/src/aggr/geo_path.rs
+++ b/src/common/function/src/aggr/geo_path.rs
@@ -1,433 +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::sync::Arc;
-
-use datafusion::arrow::array::{Array, ArrayRef};
-use datafusion::common::cast::as_primitive_array;
-use datafusion::error::{DataFusionError, Result as DfResult};
-use datafusion::logical_expr::{Accumulator as DfAccumulator, AggregateUDF, Volatility};
-use datafusion::prelude::create_udaf;
-use datafusion_common::cast::{as_list_array, as_struct_array};
-use datafusion_common::utils::SingleRowListArrayBuilder;
-use datafusion_common::ScalarValue;
-use datatypes::arrow::array::{Float64Array, Int64Array, ListArray, StructArray};
-use datatypes::arrow::datatypes::{
-    DataType, Field, Float64Type, Int64Type, TimeUnit, TimestampNanosecondType,
-};
-use datatypes::compute::{self, sort_to_indices};
-
-pub const GEO_PATH_NAME: &str = "geo_path";
-
-const LATITUDE_FIELD: &str = "lat";
-const LONGITUDE_FIELD: &str = "lng";
-const TIMESTAMP_FIELD: &str = "timestamp";
-const DEFAULT_LIST_FIELD_NAME: &str = "item";
-
-#[derive(Debug, Default)]
-pub struct GeoPathAccumulator {
-    lat: Vec<Option<f64>>,
-    lng: Vec<Option<f64>>,
-    timestamp: Vec<Option<i64>>,
-}
-
-impl GeoPathAccumulator {
-    pub fn new() -> Self {
-        Self::default()
-    }
-
-    pub fn udf_impl() -> AggregateUDF {
-        create_udaf(
-            GEO_PATH_NAME,
-            // Input types: lat, lng, timestamp
-            vec![
-                DataType::Float64,
-                DataType::Float64,
-                DataType::Timestamp(TimeUnit::Nanosecond, None),
-            ],
-            // Output type: list of points {[lat], [lng]}
-            Arc::new(DataType::Struct(
-                vec![
-                    Field::new(
-                        LATITUDE_FIELD,
-                        DataType::List(Arc::new(Field::new(
-                            DEFAULT_LIST_FIELD_NAME,
-                            DataType::Float64,
-                            true,
-                        ))),
-                        false,
-                    ),
-                    Field::new(
-                        LONGITUDE_FIELD,
-                        DataType::List(Arc::new(Field::new(
-                            DEFAULT_LIST_FIELD_NAME,
-                            DataType::Float64,
-                            true,
-                        ))),
-                        false,
-                    ),
-                ]
-                .into(),
-            )),
-            Volatility::Immutable,
-            // Create the accumulator
-            Arc::new(|_| Ok(Box::new(GeoPathAccumulator::new()))),
-            // Intermediate state types
-            Arc::new(vec![DataType::Struct(
-                vec![
-                    Field::new(
-                        LATITUDE_FIELD,
-                        DataType::List(Arc::new(Field::new(
-                            DEFAULT_LIST_FIELD_NAME,
-                            DataType::Float64,
-                            true,
-                        ))),
-                        false,
-                    ),
-                    Field::new(
-                        LONGITUDE_FIELD,
-                        DataType::List(Arc::new(Field::new(
-                            DEFAULT_LIST_FIELD_NAME,
-                            DataType::Float64,
-                            true,
-                        ))),
-                        false,
-                    ),
-                    Field::new(
-                        TIMESTAMP_FIELD,
-                        DataType::List(Arc::new(Field::new(
-                            DEFAULT_LIST_FIELD_NAME,
-                            DataType::Int64,
-                            true,
-                        ))),
-                        false,
-                    ),
-                ]
-                .into(),
-            )]),
-        )
-    }
-}
-
-impl DfAccumulator for GeoPathAccumulator {
-    fn update_batch(&mut self, values: &[ArrayRef]) -> datafusion::error::Result<()> {
-        if values.len() != 3 {
-            return Err(DataFusionError::Internal(format!(
-                "Expected 3 columns for geo_path, got {}",
-                values.len()
-            )));
-        }
-
-        let lat_array = as_primitive_array::<Float64Type>(&values[0])?;
-        let lng_array = as_primitive_array::<Float64Type>(&values[1])?;
-        let ts_array = as_primitive_array::<TimestampNanosecondType>(&values[2])?;
-
-        let size = lat_array.len();
-        self.lat.reserve(size);
-        self.lng.reserve(size);
-
-        for idx in 0..size {
-            self.lat.push(if lat_array.is_null(idx) {
-                None
-            } else {
-                Some(lat_array.value(idx))
-            });
-
-            self.lng.push(if lng_array.is_null(idx) {
-                None
-            } else {
-                Some(lng_array.value(idx))
-            });
-
-            self.timestamp.push(if ts_array.is_null(idx) {
-                None
-            } else {
-                Some(ts_array.value(idx))
-            });
-        }
-
-        Ok(())
-    }
-
-    fn evaluate(&mut self) -> DfResult<ScalarValue> {
-        let unordered_lng_array = Float64Array::from(self.lng.clone());
-        let unordered_lat_array = Float64Array::from(self.lat.clone());
-        let ts_array = Int64Array::from(self.timestamp.clone());
-
-        let ordered_indices = sort_to_indices(&ts_array, None, None)?;
-        let lat_array = compute::take(&unordered_lat_array, &ordered_indices, None)?;
-        let lng_array = compute::take(&unordered_lng_array, &ordered_indices, None)?;
-
-        let lat_list = Arc::new(SingleRowListArrayBuilder::new(lat_array).build_list_array());
-        let lng_list = Arc::new(SingleRowListArrayBuilder::new(lng_array).build_list_array());
-
-        let result = ScalarValue::Struct(Arc::new(StructArray::new(
-            vec![
-                Field::new(
-                    LATITUDE_FIELD,
-                    DataType::List(Arc::new(Field::new("item", DataType::Float64, true))),
-                    false,
-                ),
-                Field::new(
-                    LONGITUDE_FIELD,
-                    DataType::List(Arc::new(Field::new("item", DataType::Float64, true))),
-                    false,
-                ),
-            ]
-            .into(),
-            vec![lat_list, lng_list],
-            None,
-        )));
-
-        Ok(result)
-    }
-
-    fn size(&self) -> usize {
-        // Base size of GeoPathAccumulator struct fields
-        let mut total_size = std::mem::size_of::<Self>();
-
-        // Size of vectors (approximation)
-        total_size += self.lat.capacity() * std::mem::size_of::<Option<f64>>();
-        total_size += self.lng.capacity() * std::mem::size_of::<Option<f64>>();
-        total_size += self.timestamp.capacity() * std::mem::size_of::<Option<i64>>();
-
-        total_size
-    }
-
-    fn state(&mut self) -> datafusion::error::Result<Vec<ScalarValue>> {
-        let lat_array = Arc::new(ListArray::from_iter_primitive::<Float64Type, _, _>(vec![
-            Some(self.lat.clone()),
-        ]));
-        let lng_array = Arc::new(ListArray::from_iter_primitive::<Float64Type, _, _>(vec![
-            Some(self.lng.clone()),
-        ]));
-        let ts_array = Arc::new(ListArray::from_iter_primitive::<Int64Type, _, _>(vec![
-            Some(self.timestamp.clone()),
-        ]));
-
-        let state_struct = StructArray::new(
-            vec![
-                Field::new(
-                    LATITUDE_FIELD,
-                    DataType::List(Arc::new(Field::new("item", DataType::Float64, true))),
-                    false,
-                ),
-                Field::new(
-                    LONGITUDE_FIELD,
-                    DataType::List(Arc::new(Field::new("item", DataType::Float64, true))),
-                    false,
-                ),
-                Field::new(
-                    TIMESTAMP_FIELD,
-                    DataType::List(Arc::new(Field::new("item", DataType::Int64, true))),
-                    false,
-                ),
-            ]
-            .into(),
-            vec![lat_array, lng_array, ts_array],
-            None,
-        );
-
-        Ok(vec![ScalarValue::Struct(Arc::new(state_struct))])
-    }
-
-    fn merge_batch(&mut self, states: &[ArrayRef]) -> datafusion::error::Result<()> {
-        if states.len() != 1 {
-            return Err(DataFusionError::Internal(format!(
-                "Expected 1 states for geo_path, got {}",
-                states.len()
-            )));
-        }
-
-        for state in states {
-            let state = as_struct_array(state)?;
-            let lat_list = as_list_array(state.column(0))?.value(0);
-            let lat_array = as_primitive_array::<Float64Type>(&lat_list)?;
-            let lng_list = as_list_array(state.column(1))?.value(0);
-            let lng_array = as_primitive_array::<Float64Type>(&lng_list)?;
-            let ts_list = as_list_array(state.column(2))?.value(0);
-            let ts_array = as_primitive_array::<Int64Type>(&ts_list)?;
-
-            self.lat.extend(lat_array);
-            self.lng.extend(lng_array);
-            self.timestamp.extend(ts_array);
-        }
-
-        Ok(())
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use datafusion::arrow::array::{Float64Array, TimestampNanosecondArray};
-    use datafusion::scalar::ScalarValue;
-
-    use super::*;
-
-    #[test]
-    fn test_geo_path_basic() {
-        let mut accumulator = GeoPathAccumulator::new();
-
-        // Create test data
-        let lat_array = Arc::new(Float64Array::from(vec![1.0, 2.0, 3.0]));
-        let lng_array = Arc::new(Float64Array::from(vec![4.0, 5.0, 6.0]));
-        let ts_array = Arc::new(TimestampNanosecondArray::from(vec![100, 200, 300]));
-
-        // Update batch
-        accumulator
-            .update_batch(&[lat_array, lng_array, ts_array])
-            .unwrap();
-
-        // Evaluate
-        let result = accumulator.evaluate().unwrap();
-        if let ScalarValue::Struct(struct_array) = result {
-            // Verify structure
-            let fields = struct_array.fields().clone();
-            assert_eq!(fields.len(), 2);
-            assert_eq!(fields[0].name(), LATITUDE_FIELD);
-            assert_eq!(fields[1].name(), LONGITUDE_FIELD);
-
-            // Verify data
-            let columns = struct_array.columns();
-            assert_eq!(columns.len(), 2);
-
-            // Check latitude values
-            let lat_list = as_list_array(&columns[0]).unwrap().value(0);
-            let lat_array = as_primitive_array::<Float64Type>(&lat_list).unwrap();
-            assert_eq!(lat_array.len(), 3);
-            assert_eq!(lat_array.value(0), 1.0);
-            assert_eq!(lat_array.value(1), 2.0);
-            assert_eq!(lat_array.value(2), 3.0);
-
-            // Check longitude values
-            let lng_list = as_list_array(&columns[1]).unwrap().value(0);
-            let lng_array = as_primitive_array::<Float64Type>(&lng_list).unwrap();
-            assert_eq!(lng_array.len(), 3);
-            assert_eq!(lng_array.value(0), 4.0);
-            assert_eq!(lng_array.value(1), 5.0);
-            assert_eq!(lng_array.value(2), 6.0);
-        } else {
-            panic!("Expected Struct scalar value");
-        }
-    }
-
-    #[test]
-    fn test_geo_path_sort_by_timestamp() {
-        let mut accumulator = GeoPathAccumulator::new();
-
-        // Create test data with unordered timestamps
-        let lat_array = Arc::new(Float64Array::from(vec![1.0, 2.0, 3.0]));
-        let lng_array = Arc::new(Float64Array::from(vec![4.0, 5.0, 6.0]));
-        let ts_array = Arc::new(TimestampNanosecondArray::from(vec![300, 100, 200]));
-
-        // Update batch
-        accumulator
-            .update_batch(&[lat_array, lng_array, ts_array])
-            .unwrap();
-
-        // Evaluate
-        let result = accumulator.evaluate().unwrap();
-        if let ScalarValue::Struct(struct_array) = result {
-            // Extract arrays
-            let columns = struct_array.columns();
-
-            // Check latitude values
-            let lat_list = as_list_array(&columns[0]).unwrap().value(0);
-            let lat_array = as_primitive_array::<Float64Type>(&lat_list).unwrap();
-            assert_eq!(lat_array.len(), 3);
-            assert_eq!(lat_array.value(0), 2.0); // timestamp 100
-            assert_eq!(lat_array.value(1), 3.0); // timestamp 200
-            assert_eq!(lat_array.value(2), 1.0); // timestamp 300
-
-            // Check longitude values (should be sorted by timestamp)
-            let lng_list = as_list_array(&columns[1]).unwrap().value(0);
-            let lng_array = as_primitive_array::<Float64Type>(&lng_list).unwrap();
-            assert_eq!(lng_array.len(), 3);
-            assert_eq!(lng_array.value(0), 5.0); // timestamp 100
-            assert_eq!(lng_array.value(1), 6.0); // timestamp 200
-            assert_eq!(lng_array.value(2), 4.0); // timestamp 300
-        } else {
-            panic!("Expected Struct scalar value");
-        }
-    }
-
-    #[test]
-    fn test_geo_path_merge() {
-        let mut accumulator1 = GeoPathAccumulator::new();
-        let mut accumulator2 = GeoPathAccumulator::new();
-
-        // Create test data for first accumulator
-        let lat_array1 = Arc::new(Float64Array::from(vec![1.0]));
-        let lng_array1 = Arc::new(Float64Array::from(vec![4.0]));
-        let ts_array1 = Arc::new(TimestampNanosecondArray::from(vec![100]));
-
-        // Create test data for second accumulator
-        let lat_array2 = Arc::new(Float64Array::from(vec![2.0]));
-        let lng_array2 = Arc::new(Float64Array::from(vec![5.0]));
-        let ts_array2 = Arc::new(TimestampNanosecondArray::from(vec![200]));
-
-        // Update batches
-        accumulator1
-            .update_batch(&[lat_array1, lng_array1, ts_array1])
-            .unwrap();
-        accumulator2
-            .update_batch(&[lat_array2, lng_array2, ts_array2])
-            .unwrap();
-
-        // Get states
-        let state1 = accumulator1.state().unwrap();
-        let state2 = accumulator2.state().unwrap();
-
-        // Create a merged accumulator
-        let mut merged = GeoPathAccumulator::new();
-
-        // Extract the struct arrays from the states
-        let state_array1 = match &state1[0] {
-            ScalarValue::Struct(array) => array.clone(),
-            _ => panic!("Expected Struct scalar value"),
-        };
-
-        let state_array2 = match &state2[0] {
-            ScalarValue::Struct(array) => array.clone(),
-            _ => panic!("Expected Struct scalar value"),
-        };
-
-        // Merge state arrays
-        merged.merge_batch(&[state_array1]).unwrap();
-        merged.merge_batch(&[state_array2]).unwrap();
-
-        // Evaluate merged result
-        let result = merged.evaluate().unwrap();
-        if let ScalarValue::Struct(struct_array) = result {
-            // Extract arrays
-            let columns = struct_array.columns();
-
-            // Check latitude values
-            let lat_list = as_list_array(&columns[0]).unwrap().value(0);
-            let lat_array = as_primitive_array::<Float64Type>(&lat_list).unwrap();
-            assert_eq!(lat_array.len(), 2);
-            assert_eq!(lat_array.value(0), 1.0); // timestamp 100
-            assert_eq!(lat_array.value(1), 2.0); // timestamp 200
-
-            // Check longitude values (should be sorted by timestamp)
-            let lng_list = as_list_array(&columns[1]).unwrap().value(0);
-            let lng_array = as_primitive_array::<Float64Type>(&lng_list).unwrap();
-            assert_eq!(lng_array.len(), 2);
-            assert_eq!(lng_array.value(0), 4.0); // timestamp 100
-            assert_eq!(lng_array.value(1), 5.0); // timestamp 200
-        } else {
-            panic!("Expected Struct scalar value");
-        }
-    }
-}
--- a/src/common/function/src/aggr/hll.rs
+++ b/src/common/function/src/aggr/hll.rs
@@ -1,329 +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.
-
-//! Two UDAFs are implemented for HyperLogLog:
-//!
-//! - `hll`: Accepts a string column and aggregates the values into a
-//!   HyperLogLog state.
-//! - `hll_merge`: Accepts a binary column of states generated by `hll`
-//!   and merges them into a single state.
-//!
-//! The states can be then used to estimate the cardinality of the
-//! values in the column by `hll_count` UDF.
-
-use std::sync::Arc;
-
-use common_query::prelude::*;
-use common_telemetry::trace;
-use datafusion::arrow::array::ArrayRef;
-use datafusion::common::cast::{as_binary_array, as_string_array};
-use datafusion::common::not_impl_err;
-use datafusion::error::{DataFusionError, Result as DfResult};
-use datafusion::logical_expr::function::AccumulatorArgs;
-use datafusion::logical_expr::{Accumulator as DfAccumulator, AggregateUDF};
-use datafusion::prelude::create_udaf;
-use datatypes::arrow::datatypes::DataType;
-use hyperloglogplus::{HyperLogLog, HyperLogLogPlus};
-
-use crate::utils::FixedRandomState;
-
-pub const HLL_NAME: &str = "hll";
-pub const HLL_MERGE_NAME: &str = "hll_merge";
-
-const DEFAULT_PRECISION: u8 = 14;
-
-pub(crate) type HllStateType = HyperLogLogPlus<String, FixedRandomState>;
-
-pub struct HllState {
-    hll: HllStateType,
-}
-
-impl std::fmt::Debug for HllState {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(f, "HllState<Opaque>")
-    }
-}
-
-impl Default for HllState {
-    fn default() -> Self {
-        Self::new()
-    }
-}
-
-impl HllState {
-    pub fn new() -> Self {
-        Self {
-            // Safety: the DEFAULT_PRECISION is fixed and valid
-            hll: HllStateType::new(DEFAULT_PRECISION, FixedRandomState::new()).unwrap(),
-        }
-    }
-
-    /// Create a UDF for the `hll` function.
-    ///
-    /// `hll` accepts a string column and aggregates the
-    /// values into a HyperLogLog state.
-    pub fn state_udf_impl() -> AggregateUDF {
-        create_udaf(
-            HLL_NAME,
-            vec![DataType::Utf8],
-            Arc::new(DataType::Binary),
-            Volatility::Immutable,
-            Arc::new(Self::create_accumulator),
-            Arc::new(vec![DataType::Binary]),
-        )
-    }
-
-    /// Create a UDF for the `hll_merge` function.
-    ///
-    /// `hll_merge` accepts a binary column of states generated by `hll`
-    /// and merges them into a single state.
-    pub fn merge_udf_impl() -> AggregateUDF {
-        create_udaf(
-            HLL_MERGE_NAME,
-            vec![DataType::Binary],
-            Arc::new(DataType::Binary),
-            Volatility::Immutable,
-            Arc::new(Self::create_merge_accumulator),
-            Arc::new(vec![DataType::Binary]),
-        )
-    }
-
-    fn update(&mut self, value: &str) {
-        self.hll.insert(value);
-    }
-
-    fn merge(&mut self, raw: &[u8]) {
-        if let Ok(serialized) = bincode::deserialize::<HllStateType>(raw) {
-            if let Ok(()) = self.hll.merge(&serialized) {
-                return;
-            }
-        }
-        trace!("Warning: Failed to merge HyperLogLog from {:?}", raw);
-    }
-
-    fn create_accumulator(acc_args: AccumulatorArgs) -> DfResult<Box<dyn DfAccumulator>> {
-        let data_type = acc_args.exprs[0].data_type(acc_args.schema)?;
-
-        match data_type {
-            DataType::Utf8 => Ok(Box::new(HllState::new())),
-            other => not_impl_err!("{HLL_NAME} does not support data type: {other}"),
-        }
-    }
-
-    fn create_merge_accumulator(acc_args: AccumulatorArgs) -> DfResult<Box<dyn DfAccumulator>> {
-        let data_type = acc_args.exprs[0].data_type(acc_args.schema)?;
-
-        match data_type {
-            DataType::Binary => Ok(Box::new(HllState::new())),
-            other => not_impl_err!("{HLL_MERGE_NAME} does not support data type: {other}"),
-        }
-    }
-}
-
-impl DfAccumulator for HllState {
-    fn update_batch(&mut self, values: &[ArrayRef]) -> DfResult<()> {
-        let array = &values[0];
-
-        match array.data_type() {
-            DataType::Utf8 => {
-                let string_array = as_string_array(array)?;
-                for value in string_array.iter().flatten() {
-                    self.update(value);
-                }
-            }
-            DataType::Binary => {
-                let binary_array = as_binary_array(array)?;
-                for v in binary_array.iter().flatten() {
-                    self.merge(v);
-                }
-            }
-            _ => {
-                return not_impl_err!(
-                    "HLL functions do not support data type: {}",
-                    array.data_type()
-                )
-            }
-        }
-
-        Ok(())
-    }
-
-    fn evaluate(&mut self) -> DfResult<ScalarValue> {
-        Ok(ScalarValue::Binary(Some(
-            bincode::serialize(&self.hll).map_err(|e| {
-                DataFusionError::Internal(format!("Failed to serialize HyperLogLog: {}", e))
-            })?,
-        )))
-    }
-
-    fn size(&self) -> usize {
-        std::mem::size_of_val(&self.hll)
-    }
-
-    fn state(&mut self) -> DfResult<Vec<ScalarValue>> {
-        Ok(vec![ScalarValue::Binary(Some(
-            bincode::serialize(&self.hll).map_err(|e| {
-                DataFusionError::Internal(format!("Failed to serialize HyperLogLog: {}", e))
-            })?,
-        ))])
-    }
-
-    fn merge_batch(&mut self, states: &[ArrayRef]) -> DfResult<()> {
-        let array = &states[0];
-        let binary_array = as_binary_array(array)?;
-        for v in binary_array.iter().flatten() {
-            self.merge(v);
-        }
-
-        Ok(())
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use datafusion::arrow::array::{BinaryArray, StringArray};
-
-    use super::*;
-
-    #[test]
-    fn test_hll_basic() {
-        let mut state = HllState::new();
-        state.update("1");
-        state.update("2");
-        state.update("3");
-
-        let result = state.evaluate().unwrap();
-        if let ScalarValue::Binary(Some(bytes)) = result {
-            let mut hll: HllStateType = bincode::deserialize(&bytes).unwrap();
-            assert_eq!(hll.count().trunc() as u32, 3);
-        } else {
-            panic!("Expected binary scalar value");
-        }
-    }
-
-    #[test]
-    fn test_hll_roundtrip() {
-        let mut state = HllState::new();
-        state.update("1");
-        state.update("2");
-
-        // Serialize
-        let serialized = state.evaluate().unwrap();
-
-        // Create new state and merge the serialized data
-        let mut new_state = HllState::new();
-        if let ScalarValue::Binary(Some(bytes)) = &serialized {
-            new_state.merge(bytes);
-
-            // Verify the merged state matches original
-            let result = new_state.evaluate().unwrap();
-            if let ScalarValue::Binary(Some(new_bytes)) = result {
-                let mut original: HllStateType = bincode::deserialize(bytes).unwrap();
-                let mut merged: HllStateType = bincode::deserialize(&new_bytes).unwrap();
-                assert_eq!(original.count(), merged.count());
-            } else {
-                panic!("Expected binary scalar value");
-            }
-        } else {
-            panic!("Expected binary scalar value");
-        }
-    }
-
-    #[test]
-    fn test_hll_batch_update() {
-        let mut state = HllState::new();
-
-        // Test string values
-        let str_values = vec!["a", "b", "c", "d", "e", "f", "g", "h", "i"];
-        let str_array = Arc::new(StringArray::from(str_values)) as ArrayRef;
-        state.update_batch(&[str_array]).unwrap();
-
-        let result = state.evaluate().unwrap();
-        if let ScalarValue::Binary(Some(bytes)) = result {
-            let mut hll: HllStateType = bincode::deserialize(&bytes).unwrap();
-            assert_eq!(hll.count().trunc() as u32, 9);
-        } else {
-            panic!("Expected binary scalar value");
-        }
-    }
-
-    #[test]
-    fn test_hll_merge_batch() {
-        let mut state1 = HllState::new();
-        state1.update("1");
-        let state1_binary = state1.evaluate().unwrap();
-
-        let mut state2 = HllState::new();
-        state2.update("2");
-        let state2_binary = state2.evaluate().unwrap();
-
-        let mut merged_state = HllState::new();
-        if let (ScalarValue::Binary(Some(bytes1)), ScalarValue::Binary(Some(bytes2))) =
-            (&state1_binary, &state2_binary)
-        {
-            let binary_array = Arc::new(BinaryArray::from(vec![
-                bytes1.as_slice(),
-                bytes2.as_slice(),
-            ])) as ArrayRef;
-            merged_state.merge_batch(&[binary_array]).unwrap();
-
-            let result = merged_state.evaluate().unwrap();
-            if let ScalarValue::Binary(Some(bytes)) = result {
-                let mut hll: HllStateType = bincode::deserialize(&bytes).unwrap();
-                assert_eq!(hll.count().trunc() as u32, 2);
-            } else {
-                panic!("Expected binary scalar value");
-            }
-        } else {
-            panic!("Expected binary scalar values");
-        }
-    }
-
-    #[test]
-    fn test_hll_merge_function() {
-        // Create two HLL states with different values
-        let mut state1 = HllState::new();
-        state1.update("1");
-        state1.update("2");
-        let state1_binary = state1.evaluate().unwrap();
-
-        let mut state2 = HllState::new();
-        state2.update("2");
-        state2.update("3");
-        let state2_binary = state2.evaluate().unwrap();
-
-        // Create a merge state and merge both states
-        let mut merge_state = HllState::new();
-        if let (ScalarValue::Binary(Some(bytes1)), ScalarValue::Binary(Some(bytes2))) =
-            (&state1_binary, &state2_binary)
-        {
-            let binary_array = Arc::new(BinaryArray::from(vec![
-                bytes1.as_slice(),
-                bytes2.as_slice(),
-            ])) as ArrayRef;
-            merge_state.update_batch(&[binary_array]).unwrap();
-
-            let result = merge_state.evaluate().unwrap();
-            if let ScalarValue::Binary(Some(bytes)) = result {
-                let mut hll: HllStateType = bincode::deserialize(&bytes).unwrap();
-                // Should have 3 unique values: "1", "2", "3"
-                assert_eq!(hll.count().trunc() as u32, 3);
-            } else {
-                panic!("Expected binary scalar value");
-            }
-        } else {
-            panic!("Expected binary scalar values");
-        }
-    }
-}
--- a/src/common/function/src/aggr/uddsketch_state.rs
+++ b/src/common/function/src/aggr/uddsketch_state.rs
@@ -1,313 +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.
-
-//! Implementation of the `uddsketch_state` UDAF that generate the state of
-//! UDDSketch for a given set of values.
-//!
-//! The generated state can be used to compute approximate quantiles using
-//! `uddsketch_calc` UDF.
-
-use std::sync::Arc;
-
-use common_query::prelude::*;
-use common_telemetry::trace;
-use datafusion::common::cast::{as_binary_array, as_primitive_array};
-use datafusion::common::not_impl_err;
-use datafusion::error::{DataFusionError, Result as DfResult};
-use datafusion::logical_expr::function::AccumulatorArgs;
-use datafusion::logical_expr::{Accumulator as DfAccumulator, AggregateUDF};
-use datafusion::physical_plan::expressions::Literal;
-use datafusion::prelude::create_udaf;
-use datatypes::arrow::array::ArrayRef;
-use datatypes::arrow::datatypes::{DataType, Float64Type};
-use uddsketch::{SketchHashKey, UDDSketch};
-
-pub const UDDSKETCH_STATE_NAME: &str = "uddsketch_state";
-
-#[derive(Debug)]
-pub struct UddSketchState {
-    uddsketch: UDDSketch,
-}
-
-impl UddSketchState {
-    pub fn new(bucket_size: u64, error_rate: f64) -> Self {
-        Self {
-            uddsketch: UDDSketch::new(bucket_size, error_rate),
-        }
-    }
-
-    pub fn udf_impl() -> AggregateUDF {
-        create_udaf(
-            UDDSKETCH_STATE_NAME,
-            vec![DataType::Int64, DataType::Float64, DataType::Float64],
-            Arc::new(DataType::Binary),
-            Volatility::Immutable,
-            Arc::new(|args| {
-                let (bucket_size, error_rate) = downcast_accumulator_args(args)?;
-                Ok(Box::new(UddSketchState::new(bucket_size, error_rate)))
-            }),
-            Arc::new(vec![DataType::Binary]),
-        )
-    }
-
-    fn update(&mut self, value: f64) {
-        self.uddsketch.add_value(value);
-    }
-
-    fn merge(&mut self, raw: &[u8]) {
-        if let Ok(uddsketch) = bincode::deserialize::<UDDSketch>(raw) {
-            if uddsketch.count() != 0 {
-                self.uddsketch.merge_sketch(&uddsketch);
-            }
-        } else {
-            trace!("Warning: Failed to deserialize UDDSketch from {:?}", raw);
-        }
-    }
-}
-
-fn downcast_accumulator_args(args: AccumulatorArgs) -> DfResult<(u64, f64)> {
-    let bucket_size = match args.exprs[0]
-        .as_any()
-        .downcast_ref::<Literal>()
-        .map(|lit| lit.value())
-    {
-        Some(ScalarValue::Int64(Some(value))) => *value as u64,
-        _ => {
-            return not_impl_err!(
-                "{} not supported for bucket size: {}",
-                UDDSKETCH_STATE_NAME,
-                &args.exprs[0]
-            )
-        }
-    };
-
-    let error_rate = match args.exprs[1]
-        .as_any()
-        .downcast_ref::<Literal>()
-        .map(|lit| lit.value())
-    {
-        Some(ScalarValue::Float64(Some(value))) => *value,
-        _ => {
-            return not_impl_err!(
-                "{} not supported for error rate: {}",
-                UDDSKETCH_STATE_NAME,
-                &args.exprs[1]
-            )
-        }
-    };
-
-    Ok((bucket_size, error_rate))
-}
-
-impl DfAccumulator for UddSketchState {
-    fn update_batch(&mut self, values: &[ArrayRef]) -> DfResult<()> {
-        let array = &values[2]; // the third column is data value
-        let f64_array = as_primitive_array::<Float64Type>(array)?;
-        for v in f64_array.iter().flatten() {
-            self.update(v);
-        }
-
-        Ok(())
-    }
-
-    fn evaluate(&mut self) -> DfResult<ScalarValue> {
-        Ok(ScalarValue::Binary(Some(
-            bincode::serialize(&self.uddsketch).map_err(|e| {
-                DataFusionError::Internal(format!("Failed to serialize UDDSketch: {}", e))
-            })?,
-        )))
-    }
-
-    fn size(&self) -> usize {
-        // Base size of UDDSketch struct fields
-        let mut total_size = std::mem::size_of::<f64>() * 3 + // alpha, gamma, values_sum
-                            std::mem::size_of::<u32>() +      // compactions
-                            std::mem::size_of::<u64>() * 2; // max_buckets, num_values
-
-        // Size of buckets (SketchHashMap)
-        // Each bucket entry contains:
-        // - SketchHashKey (enum with i64/Zero/Invalid variants)
-        // - SketchHashEntry (count: u64, next: SketchHashKey)
-        let bucket_entry_size = std::mem::size_of::<SketchHashKey>() + // key
-                               std::mem::size_of::<u64>() +            // count
-                               std::mem::size_of::<SketchHashKey>(); // next
-
-        total_size += self.uddsketch.current_buckets_count() * bucket_entry_size;
-
-        total_size
-    }
-
-    fn state(&mut self) -> DfResult<Vec<ScalarValue>> {
-        Ok(vec![ScalarValue::Binary(Some(
-            bincode::serialize(&self.uddsketch).map_err(|e| {
-                DataFusionError::Internal(format!("Failed to serialize UDDSketch: {}", e))
-            })?,
-        ))])
-    }
-
-    fn merge_batch(&mut self, states: &[ArrayRef]) -> DfResult<()> {
-        let array = &states[0];
-        let binary_array = as_binary_array(array)?;
-        for v in binary_array.iter().flatten() {
-            self.merge(v);
-        }
-
-        Ok(())
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use datafusion::arrow::array::{BinaryArray, Float64Array};
-
-    use super::*;
-
-    #[test]
-    fn test_uddsketch_state_basic() {
-        let mut state = UddSketchState::new(10, 0.01);
-        state.update(1.0);
-        state.update(2.0);
-        state.update(3.0);
-
-        let result = state.evaluate().unwrap();
-        if let ScalarValue::Binary(Some(bytes)) = result {
-            let deserialized: UDDSketch = bincode::deserialize(&bytes).unwrap();
-            assert_eq!(deserialized.count(), 3);
-        } else {
-            panic!("Expected binary scalar value");
-        }
-    }
-
-    #[test]
-    fn test_uddsketch_state_roundtrip() {
-        let mut state = UddSketchState::new(10, 0.01);
-        state.update(1.0);
-        state.update(2.0);
-
-        // Serialize
-        let serialized = state.evaluate().unwrap();
-
-        // Create new state and merge the serialized data
-        let mut new_state = UddSketchState::new(10, 0.01);
-        if let ScalarValue::Binary(Some(bytes)) = &serialized {
-            new_state.merge(bytes);
-
-            // Verify the merged state matches original by comparing deserialized values
-            let original_sketch: UDDSketch = bincode::deserialize(bytes).unwrap();
-            let new_result = new_state.evaluate().unwrap();
-            if let ScalarValue::Binary(Some(new_bytes)) = new_result {
-                let new_sketch: UDDSketch = bincode::deserialize(&new_bytes).unwrap();
-                assert_eq!(original_sketch.count(), new_sketch.count());
-                assert_eq!(original_sketch.sum(), new_sketch.sum());
-                assert_eq!(original_sketch.mean(), new_sketch.mean());
-                assert_eq!(original_sketch.max_error(), new_sketch.max_error());
-                // Compare a few quantiles to ensure statistical equivalence
-                for q in [0.1, 0.5, 0.9].iter() {
-                    assert!(
-                        (original_sketch.estimate_quantile(*q) - new_sketch.estimate_quantile(*q))
-                            .abs()
-                            < 1e-10,
-                        "Quantile {} mismatch: original={}, new={}",
-                        q,
-                        original_sketch.estimate_quantile(*q),
-                        new_sketch.estimate_quantile(*q)
-                    );
-                }
-            } else {
-                panic!("Expected binary scalar value");
-            }
-        } else {
-            panic!("Expected binary scalar value");
-        }
-    }
-
-    #[test]
-    fn test_uddsketch_state_batch_update() {
-        let mut state = UddSketchState::new(10, 0.01);
-        let values = vec![1.0f64, 2.0, 3.0];
-        let array = Arc::new(Float64Array::from(values)) as ArrayRef;
-
-        state
-            .update_batch(&[array.clone(), array.clone(), array])
-            .unwrap();
-
-        let result = state.evaluate().unwrap();
-        if let ScalarValue::Binary(Some(bytes)) = result {
-            let deserialized: UDDSketch = bincode::deserialize(&bytes).unwrap();
-            assert_eq!(deserialized.count(), 3);
-        } else {
-            panic!("Expected binary scalar value");
-        }
-    }
-
-    #[test]
-    fn test_uddsketch_state_merge_batch() {
-        let mut state1 = UddSketchState::new(10, 0.01);
-        state1.update(1.0);
-        let state1_binary = state1.evaluate().unwrap();
-
-        let mut state2 = UddSketchState::new(10, 0.01);
-        state2.update(2.0);
-        let state2_binary = state2.evaluate().unwrap();
-
-        let mut merged_state = UddSketchState::new(10, 0.01);
-        if let (ScalarValue::Binary(Some(bytes1)), ScalarValue::Binary(Some(bytes2))) =
-            (&state1_binary, &state2_binary)
-        {
-            let binary_array = Arc::new(BinaryArray::from(vec![
-                bytes1.as_slice(),
-                bytes2.as_slice(),
-            ])) as ArrayRef;
-            merged_state.merge_batch(&[binary_array]).unwrap();
-
-            let result = merged_state.evaluate().unwrap();
-            if let ScalarValue::Binary(Some(bytes)) = result {
-                let deserialized: UDDSketch = bincode::deserialize(&bytes).unwrap();
-                assert_eq!(deserialized.count(), 2);
-            } else {
-                panic!("Expected binary scalar value");
-            }
-        } else {
-            panic!("Expected binary scalar values");
-        }
-    }
-
-    #[test]
-    fn test_uddsketch_state_size() {
-        let mut state = UddSketchState::new(10, 0.01);
-        let initial_size = state.size();
-
-        // Add some values to create buckets
-        state.update(1.0);
-        state.update(2.0);
-        state.update(3.0);
-
-        let size_with_values = state.size();
-        assert!(
-            size_with_values > initial_size,
-            "Size should increase after adding values: initial={}, with_values={}",
-            initial_size,
-            size_with_values
-        );
-
-        // Verify size increases with more buckets
-        state.update(10.0); // This should create a new bucket
-        assert!(
-            state.size() > size_with_values,
-            "Size should increase after adding new bucket: prev={}, new={}",
-            size_with_values,
-            state.size()
-        );
-    }
-}
--- a/src/common/function/src/function.rs
+++ b/src/common/function/src/function.rs
@@ -63,7 +63,7 @@ pub trait Function: fmt::Display + Sync + Send {
    fn signature(&self) -> Signature;

    /// Evaluate the function, e.g. run/execute the function.
-    fn eval(&self, ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef>;
+    fn eval(&self, _func_ctx: FunctionContext, _columns: &[VectorRef]) -> Result<VectorRef>;
 }

 pub type FunctionRef = Arc<dyn Function>;
--- a/src/common/function/src/function_registry.rs
+++ b/src/common/function/src/function_registry.rs
@@ -18,20 +18,17 @@ use std::sync::{Arc, RwLock};

 use once_cell::sync::Lazy;

-use crate::admin::AdminFunction;
 use crate::function::{AsyncFunctionRef, FunctionRef};
 use crate::scalars::aggregate::{AggregateFunctionMetaRef, AggregateFunctions};
 use crate::scalars::date::DateFunction;
 use crate::scalars::expression::ExpressionFunction;
-use crate::scalars::hll_count::HllCalcFunction;
-use crate::scalars::ip::IpFunctions;
 use crate::scalars::json::JsonFunction;
 use crate::scalars::matches::MatchesFunction;
 use crate::scalars::math::MathFunction;
 use crate::scalars::timestamp::TimestampFunction;
-use crate::scalars::uddsketch_calc::UddSketchCalcFunction;
 use crate::scalars::vector::VectorFunction;
 use crate::system::SystemFunction;
+use crate::table::TableFunction;

 #[derive(Default)]
 pub struct FunctionRegistry {
@@ -108,8 +105,6 @@ pub static FUNCTION_REGISTRY: Lazy<Arc<FunctionRegistry>> = Lazy::new(|| {
    TimestampFunction::register(&function_registry);
    DateFunction::register(&function_registry);
    ExpressionFunction::register(&function_registry);
-    UddSketchCalcFunction::register(&function_registry);
-    HllCalcFunction::register(&function_registry);

    // Aggregate functions
    AggregateFunctions::register(&function_registry);
@@ -119,7 +114,7 @@ pub static FUNCTION_REGISTRY: Lazy<Arc<FunctionRegistry>> = Lazy::new(|| {

    // System and administration functions
    SystemFunction::register(&function_registry);
-    AdminFunction::register(&function_registry);
+    TableFunction::register(&function_registry);

    // Json related functions
    JsonFunction::register(&function_registry);
@@ -131,9 +126,6 @@ pub static FUNCTION_REGISTRY: Lazy<Arc<FunctionRegistry>> = Lazy::new(|| {
    #[cfg(feature = "geo")]
    crate::scalars::geo::GeoFunctions::register(&function_registry);

-    // Ip functions
-    IpFunctions::register(&function_registry);
-
    Arc::new(function_registry)
 });

--- a/src/common/function/src/lib.rs
+++ b/src/common/function/src/lib.rs
@@ -15,13 +15,12 @@
 #![feature(let_chains)]
 #![feature(try_blocks)]

-mod admin;
 mod flush_flow;
 mod macros;
 pub mod scalars;
 mod system;
+mod table;

-pub mod aggr;
 pub mod function;
 pub mod function_registry;
 pub mod handlers;
--- a/src/common/function/src/scalars.rs
+++ b/src/common/function/src/scalars.rs
@@ -22,10 +22,7 @@ pub mod matches;
 pub mod math;
 pub mod vector;

-pub(crate) mod hll_count;
-pub mod ip;
 #[cfg(test)]
 pub(crate) mod test;
 pub(crate) mod timestamp;
-pub(crate) mod uddsketch_calc;
 pub mod udf;
--- a/src/common/function/src/scalars/aggregate.rs
+++ b/src/common/function/src/scalars/aggregate.rs
@@ -12,16 +12,24 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.

-//! # Deprecate Warning:
-//!
-//! This module is deprecated and will be removed in the future.
-//! All UDAF implementation here are not maintained and should
-//! not be used before they are refactored into the `src/aggr`
-//! version.
+mod argmax;
+mod argmin;
+mod diff;
+mod mean;
+mod polyval;
+mod scipy_stats_norm_cdf;
+mod scipy_stats_norm_pdf;

 use std::sync::Arc;

+pub use argmax::ArgmaxAccumulatorCreator;
+pub use argmin::ArgminAccumulatorCreator;
 use common_query::logical_plan::AggregateFunctionCreatorRef;
+pub use diff::DiffAccumulatorCreator;
+pub use mean::MeanAccumulatorCreator;
+pub use polyval::PolyvalAccumulatorCreator;
+pub use scipy_stats_norm_cdf::ScipyStatsNormCdfAccumulatorCreator;
+pub use scipy_stats_norm_pdf::ScipyStatsNormPdfAccumulatorCreator;

 use crate::function_registry::FunctionRegistry;
 use crate::scalars::vector::product::VectorProductCreator;
@@ -68,22 +76,31 @@ pub(crate) struct AggregateFunctions;

 impl AggregateFunctions {
    pub fn register(registry: &FunctionRegistry) {
-        registry.register_aggregate_function(Arc::new(AggregateFunctionMeta::new(
-            "vec_sum",
-            1,
-            Arc::new(|| Arc::new(VectorSumCreator::default())),
-        )));
-        registry.register_aggregate_function(Arc::new(AggregateFunctionMeta::new(
-            "vec_product",
-            1,
-            Arc::new(|| Arc::new(VectorProductCreator::default())),
-        )));
+        macro_rules! register_aggr_func {
+            ($name :expr, $arg_count :expr, $creator :ty) => {
+                registry.register_aggregate_function(Arc::new(AggregateFunctionMeta::new(
+                    $name,
+                    $arg_count,
+                    Arc::new(|| Arc::new(<$creator>::default())),
+                )));
+            };
+        }
+
+        register_aggr_func!("diff", 1, DiffAccumulatorCreator);
+        register_aggr_func!("mean", 1, MeanAccumulatorCreator);
+        register_aggr_func!("polyval", 2, PolyvalAccumulatorCreator);
+        register_aggr_func!("argmax", 1, ArgmaxAccumulatorCreator);
+        register_aggr_func!("argmin", 1, ArgminAccumulatorCreator);
+        register_aggr_func!("scipystatsnormcdf", 2, ScipyStatsNormCdfAccumulatorCreator);
+        register_aggr_func!("scipystatsnormpdf", 2, ScipyStatsNormPdfAccumulatorCreator);
+        register_aggr_func!("vec_sum", 1, VectorSumCreator);
+        register_aggr_func!("vec_product", 1, VectorProductCreator);

        #[cfg(feature = "geo")]
-        registry.register_aggregate_function(Arc::new(AggregateFunctionMeta::new(
+        register_aggr_func!(
            "json_encode_path",
            3,
-            Arc::new(|| Arc::new(super::geo::encoding::JsonPathEncodeFunctionCreator::default())),
-        )));
+            super::geo::encoding::JsonPathEncodeFunctionCreator
+        );
    }
 }
--- a/src/common/function/src/scalars/aggregate/argmax.rs
+++ b/src/common/function/src/scalars/aggregate/argmax.rs
@@ -0,0 +1,208 @@
+// 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::cmp::Ordering;
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    BadAccumulatorImplSnafu, CreateAccumulatorSnafu, InvalidInputStateSnafu, Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::types::{LogicalPrimitiveType, WrapperType};
+use datatypes::vectors::{ConstantVector, Helper};
+use datatypes::with_match_primitive_type_id;
+use snafu::ensure;
+
+// https://numpy.org/doc/stable/reference/generated/numpy.argmax.html
+// return the index of the max value
+#[derive(Debug, Default)]
+pub struct Argmax<T> {
+    max: Option<T>,
+    n: u64,
+}
+
+impl<T> Argmax<T>
+where
+    T: PartialOrd + Copy,
+{
+    fn update(&mut self, value: T, index: u64) {
+        if let Some(Ordering::Less) = self.max.partial_cmp(&Some(value)) {
+            self.max = Some(value);
+            self.n = index;
+        }
+    }
+}
+
+impl<T> Accumulator for Argmax<T>
+where
+    T: WrapperType + PartialOrd,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        match self.max {
+            Some(max) => Ok(vec![max.into(), self.n.into()]),
+            _ => Ok(vec![Value::Null, self.n.into()]),
+        }
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        let column = &values[0];
+        let column: &<T as Scalar>::VectorType = if column.is_const() {
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+        for (i, v) in column.iter_data().enumerate() {
+            if let Some(value) = v {
+                self.update(value, i as u64);
+            }
+        }
+        Ok(())
+    }
+
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(
+            states.len() == 2,
+            BadAccumulatorImplSnafu {
+                err_msg: "expect 2 states in `merge_batch`",
+            }
+        );
+
+        let max = &states[0];
+        let index = &states[1];
+        let max: &<T as Scalar>::VectorType = unsafe { Helper::static_cast(max) };
+        let index: &<u64 as Scalar>::VectorType = unsafe { Helper::static_cast(index) };
+        index
+            .iter_data()
+            .flatten()
+            .zip(max.iter_data().flatten())
+            .for_each(|(i, max)| self.update(max, i));
+        Ok(())
+    }
+
+    fn evaluate(&self) -> Result<Value> {
+        match self.max {
+            Some(_) => Ok(self.n.into()),
+            _ => Ok(Value::Null),
+        }
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct ArgmaxAccumulatorCreator {}
+
+impl AggregateFunctionCreator for ArgmaxAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(Argmax::<<$S as LogicalPrimitiveType>::Wrapper>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"ARGMAX\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::uint64_datatype())
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+
+        ensure!(input_types.len() == 1, InvalidInputStateSnafu);
+
+        Ok(vec![
+            input_types.into_iter().next().unwrap(),
+            ConcreteDataType::uint64_datatype(),
+        ])
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::Int32Vector;
+
+    use super::*;
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut argmax = Argmax::<i32>::default();
+        argmax.update_batch(&[]).unwrap();
+        assert_eq!(Value::Null, argmax.evaluate().unwrap());
+
+        // test update one not-null value
+        let mut argmax = Argmax::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Some(42)]))];
+        argmax.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0_u64), argmax.evaluate().unwrap());
+
+        // test update one null value
+        let mut argmax = Argmax::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Option::<i32>::None]))];
+        argmax.update_batch(&v).unwrap();
+        assert_eq!(Value::Null, argmax.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut argmax = Argmax::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-1i32),
+            Some(1),
+            Some(3),
+        ]))];
+        argmax.update_batch(&v).unwrap();
+        assert_eq!(Value::from(2_u64), argmax.evaluate().unwrap());
+
+        // test update null-value batch
+        let mut argmax = Argmax::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-2i32),
+            None,
+            Some(4),
+        ]))];
+        argmax.update_batch(&v).unwrap();
+        assert_eq!(Value::from(2_u64), argmax.evaluate().unwrap());
+
+        // test update with constant vector
+        let mut argmax = Argmax::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(ConstantVector::new(
+            Arc::new(Int32Vector::from_vec(vec![4])),
+            10,
+        ))];
+        argmax.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0_u64), argmax.evaluate().unwrap());
+    }
+}
--- a/src/common/function/src/scalars/aggregate/argmin.rs
+++ b/src/common/function/src/scalars/aggregate/argmin.rs
@@ -0,0 +1,216 @@
+// 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::cmp::Ordering;
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    BadAccumulatorImplSnafu, CreateAccumulatorSnafu, InvalidInputStateSnafu, Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::vectors::{ConstantVector, Helper};
+use datatypes::with_match_primitive_type_id;
+use snafu::ensure;
+
+// // https://numpy.org/doc/stable/reference/generated/numpy.argmin.html
+#[derive(Debug, Default)]
+pub struct Argmin<T> {
+    min: Option<T>,
+    n: u32,
+}
+
+impl<T> Argmin<T>
+where
+    T: Copy + PartialOrd,
+{
+    fn update(&mut self, value: T, index: u32) {
+        match self.min {
+            Some(min) => {
+                if let Some(Ordering::Greater) = min.partial_cmp(&value) {
+                    self.min = Some(value);
+                    self.n = index;
+                }
+            }
+            None => {
+                self.min = Some(value);
+                self.n = index;
+            }
+        }
+    }
+}
+
+impl<T> Accumulator for Argmin<T>
+where
+    T: WrapperType + PartialOrd,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        match self.min {
+            Some(min) => Ok(vec![min.into(), self.n.into()]),
+            _ => Ok(vec![Value::Null, self.n.into()]),
+        }
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(values.len() == 1, InvalidInputStateSnafu);
+
+        let column = &values[0];
+        let column: &<T as Scalar>::VectorType = if column.is_const() {
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+        for (i, v) in column.iter_data().enumerate() {
+            if let Some(value) = v {
+                self.update(value, i as u32);
+            }
+        }
+        Ok(())
+    }
+
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(
+            states.len() == 2,
+            BadAccumulatorImplSnafu {
+                err_msg: "expect 2 states in `merge_batch`",
+            }
+        );
+
+        let min = &states[0];
+        let index = &states[1];
+        let min: &<T as Scalar>::VectorType = unsafe { Helper::static_cast(min) };
+        let index: &<u32 as Scalar>::VectorType = unsafe { Helper::static_cast(index) };
+        index
+            .iter_data()
+            .flatten()
+            .zip(min.iter_data().flatten())
+            .for_each(|(i, min)| self.update(min, i));
+        Ok(())
+    }
+
+    fn evaluate(&self) -> Result<Value> {
+        match self.min {
+            Some(_) => Ok(self.n.into()),
+            _ => Ok(Value::Null),
+        }
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct ArgminAccumulatorCreator {}
+
+impl AggregateFunctionCreator for ArgminAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(Argmin::<<$S as LogicalPrimitiveType>::Wrapper>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"ARGMIN\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::uint32_datatype())
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+
+        ensure!(input_types.len() == 1, InvalidInputStateSnafu);
+
+        Ok(vec![
+            input_types.into_iter().next().unwrap(),
+            ConcreteDataType::uint32_datatype(),
+        ])
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::Int32Vector;
+
+    use super::*;
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut argmin = Argmin::<i32>::default();
+        argmin.update_batch(&[]).unwrap();
+        assert_eq!(Value::Null, argmin.evaluate().unwrap());
+
+        // test update one not-null value
+        let mut argmin = Argmin::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Some(42)]))];
+        argmin.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0_u32), argmin.evaluate().unwrap());
+
+        // test update one null value
+        let mut argmin = Argmin::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Option::<i32>::None]))];
+        argmin.update_batch(&v).unwrap();
+        assert_eq!(Value::Null, argmin.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut argmin = Argmin::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-1i32),
+            Some(1),
+            Some(3),
+        ]))];
+        argmin.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0_u32), argmin.evaluate().unwrap());
+
+        // test update null-value batch
+        let mut argmin = Argmin::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-2i32),
+            None,
+            Some(4),
+        ]))];
+        argmin.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0_u32), argmin.evaluate().unwrap());
+
+        // test update with constant vector
+        let mut argmin = Argmin::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(ConstantVector::new(
+            Arc::new(Int32Vector::from_vec(vec![4])),
+            10,
+        ))];
+        argmin.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0_u32), argmin.evaluate().unwrap());
+    }
+}
--- a/src/common/function/src/scalars/aggregate/diff.rs
+++ b/src/common/function/src/scalars/aggregate/diff.rs
@@ -0,0 +1,252 @@
+// 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::marker::PhantomData;
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    CreateAccumulatorSnafu, DowncastVectorSnafu, FromScalarValueSnafu, InvalidInputStateSnafu,
+    Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::value::ListValue;
+use datatypes::vectors::{ConstantVector, Helper, ListVector};
+use datatypes::with_match_primitive_type_id;
+use num_traits::AsPrimitive;
+use snafu::{ensure, OptionExt, ResultExt};
+
+// https://numpy.org/doc/stable/reference/generated/numpy.diff.html
+// I is the input type, O is the output type.
+#[derive(Debug, Default)]
+pub struct Diff<I, O> {
+    values: Vec<I>,
+    _phantom: PhantomData<O>,
+}
+
+impl<I, O> Diff<I, O> {
+    fn push(&mut self, value: I) {
+        self.values.push(value);
+    }
+}
+
+impl<I, O> Accumulator for Diff<I, O>
+where
+    I: WrapperType,
+    O: WrapperType,
+    I::Native: AsPrimitive<O::Native>,
+    O::Native: std::ops::Sub<Output = O::Native>,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        let nums = self
+            .values
+            .iter()
+            .map(|&n| n.into())
+            .collect::<Vec<Value>>();
+        Ok(vec![Value::List(ListValue::new(
+            nums,
+            I::LogicalType::build_data_type(),
+        ))])
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(values.len() == 1, InvalidInputStateSnafu);
+
+        let column = &values[0];
+        let mut len = 1;
+        let column: &<I as Scalar>::VectorType = if column.is_const() {
+            len = column.len();
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+        (0..len).for_each(|_| {
+            for v in column.iter_data().flatten() {
+                self.push(v);
+            }
+        });
+        Ok(())
+    }
+
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        let states = &states[0];
+        let states = states
+            .as_any()
+            .downcast_ref::<ListVector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect ListVector, got vector type {}",
+                    states.vector_type_name()
+                ),
+            })?;
+        for state in states.values_iter() {
+            if let Some(state) = state.context(FromScalarValueSnafu)? {
+                self.update_batch(&[state])?;
+            }
+        }
+        Ok(())
+    }
+
+    fn evaluate(&self) -> Result<Value> {
+        if self.values.is_empty() || self.values.len() == 1 {
+            return Ok(Value::Null);
+        }
+        let diff = self
+            .values
+            .windows(2)
+            .map(|x| {
+                let native = x[1].into_native().as_() - x[0].into_native().as_();
+                O::from_native(native).into()
+            })
+            .collect::<Vec<Value>>();
+        let diff = Value::List(ListValue::new(diff, O::LogicalType::build_data_type()));
+        Ok(diff)
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct DiffAccumulatorCreator {}
+
+impl AggregateFunctionCreator for DiffAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(Diff::<<$S as LogicalPrimitiveType>::Wrapper, <<$S as LogicalPrimitiveType>::LargestType as LogicalPrimitiveType>::Wrapper>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"DIFF\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 1, InvalidInputStateSnafu);
+        with_match_primitive_type_id!(
+            input_types[0].logical_type_id(),
+            |$S| {
+                Ok(ConcreteDataType::list_datatype($S::default().into()))
+            },
+            {
+                unreachable!()
+            }
+        )
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 1, InvalidInputStateSnafu);
+        with_match_primitive_type_id!(
+            input_types[0].logical_type_id(),
+            |$S| {
+                Ok(vec![ConcreteDataType::list_datatype($S::default().into())])
+            },
+            {
+                unreachable!()
+            }
+        )
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::Int32Vector;
+
+    use super::*;
+
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut diff = Diff::<i32, i64>::default();
+        diff.update_batch(&[]).unwrap();
+        assert!(diff.values.is_empty());
+        assert_eq!(Value::Null, diff.evaluate().unwrap());
+
+        // test update one not-null value
+        let mut diff = Diff::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Some(42)]))];
+        diff.update_batch(&v).unwrap();
+        assert_eq!(Value::Null, diff.evaluate().unwrap());
+
+        // test update one null value
+        let mut diff = Diff::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Option::<i32>::None]))];
+        diff.update_batch(&v).unwrap();
+        assert_eq!(Value::Null, diff.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut diff = Diff::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-1i32),
+            Some(1),
+            Some(2),
+        ]))];
+        let values = vec![Value::from(2_i64), Value::from(1_i64)];
+        diff.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::List(ListValue::new(values, ConcreteDataType::int64_datatype())),
+            diff.evaluate().unwrap()
+        );
+
+        // test update null-value batch
+        let mut diff = Diff::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-2i32),
+            None,
+            Some(3),
+            Some(4),
+        ]))];
+        let values = vec![Value::from(5_i64), Value::from(1_i64)];
+        diff.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::List(ListValue::new(values, ConcreteDataType::int64_datatype())),
+            diff.evaluate().unwrap()
+        );
+
+        // test update with constant vector
+        let mut diff = Diff::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(ConstantVector::new(
+            Arc::new(Int32Vector::from_vec(vec![4])),
+            4,
+        ))];
+        let values = vec![Value::from(0_i64), Value::from(0_i64), Value::from(0_i64)];
+        diff.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::List(ListValue::new(values, ConcreteDataType::int64_datatype())),
+            diff.evaluate().unwrap()
+        );
+    }
+}
--- a/src/common/function/src/scalars/aggregate/mean.rs
+++ b/src/common/function/src/scalars/aggregate/mean.rs
@@ -0,0 +1,238 @@
+// 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::marker::PhantomData;
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    BadAccumulatorImplSnafu, CreateAccumulatorSnafu, DowncastVectorSnafu, InvalidInputStateSnafu,
+    Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::types::WrapperType;
+use datatypes::vectors::{ConstantVector, Float64Vector, Helper, UInt64Vector};
+use datatypes::with_match_primitive_type_id;
+use num_traits::AsPrimitive;
+use snafu::{ensure, OptionExt};
+
+#[derive(Debug, Default)]
+pub struct Mean<T> {
+    sum: f64,
+    n: u64,
+    _phantom: PhantomData<T>,
+}
+
+impl<T> Mean<T>
+where
+    T: WrapperType,
+    T::Native: AsPrimitive<f64>,
+{
+    #[inline(always)]
+    fn push(&mut self, value: T) {
+        self.sum += value.into_native().as_();
+        self.n += 1;
+    }
+
+    #[inline(always)]
+    fn update(&mut self, sum: f64, n: u64) {
+        self.sum += sum;
+        self.n += n;
+    }
+}
+
+impl<T> Accumulator for Mean<T>
+where
+    T: WrapperType,
+    T::Native: AsPrimitive<f64>,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        Ok(vec![self.sum.into(), self.n.into()])
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(values.len() == 1, InvalidInputStateSnafu);
+        let column = &values[0];
+        let mut len = 1;
+        let column: &<T as Scalar>::VectorType = if column.is_const() {
+            len = column.len();
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+        (0..len).for_each(|_| {
+            for v in column.iter_data().flatten() {
+                self.push(v);
+            }
+        });
+
+        Ok(())
+    }
+
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(
+            states.len() == 2,
+            BadAccumulatorImplSnafu {
+                err_msg: "expect 2 states in `merge_batch`",
+            }
+        );
+
+        let sum = &states[0];
+        let n = &states[1];
+
+        let sum = sum
+            .as_any()
+            .downcast_ref::<Float64Vector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect Float64Vector, got vector type {}",
+                    sum.vector_type_name()
+                ),
+            })?;
+
+        let n = n
+            .as_any()
+            .downcast_ref::<UInt64Vector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect UInt64Vector, got vector type {}",
+                    sum.vector_type_name()
+                ),
+            })?;
+
+        sum.iter_data().zip(n.iter_data()).for_each(|(sum, n)| {
+            if let (Some(sum), Some(n)) = (sum, n) {
+                self.update(sum, n);
+            }
+        });
+        Ok(())
+    }
+
+    fn evaluate(&self) -> Result<Value> {
+        if self.n == 0 {
+            return Ok(Value::Null);
+        }
+        let values = self.sum / self.n as f64;
+        Ok(values.into())
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct MeanAccumulatorCreator {}
+
+impl AggregateFunctionCreator for MeanAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(Mean::<<$S as LogicalPrimitiveType>::Native>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"MEAN\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 1, InvalidInputStateSnafu);
+        Ok(ConcreteDataType::float64_datatype())
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 1, InvalidInputStateSnafu);
+        Ok(vec![
+            ConcreteDataType::float64_datatype(),
+            ConcreteDataType::uint64_datatype(),
+        ])
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::Int32Vector;
+
+    use super::*;
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut mean = Mean::<i32>::default();
+        mean.update_batch(&[]).unwrap();
+        assert_eq!(Value::Null, mean.evaluate().unwrap());
+
+        // test update one not-null value
+        let mut mean = Mean::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Some(42)]))];
+        mean.update_batch(&v).unwrap();
+        assert_eq!(Value::from(42.0_f64), mean.evaluate().unwrap());
+
+        // test update one null value
+        let mut mean = Mean::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![Option::<i32>::None]))];
+        mean.update_batch(&v).unwrap();
+        assert_eq!(Value::Null, mean.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut mean = Mean::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-1i32),
+            Some(1),
+            Some(2),
+        ]))];
+        mean.update_batch(&v).unwrap();
+        assert_eq!(Value::from(0.6666666666666666), mean.evaluate().unwrap());
+
+        // test update null-value batch
+        let mut mean = Mean::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(Int32Vector::from(vec![
+            Some(-2i32),
+            None,
+            Some(3),
+            Some(4),
+        ]))];
+        mean.update_batch(&v).unwrap();
+        assert_eq!(Value::from(1.6666666666666667), mean.evaluate().unwrap());
+
+        // test update with constant vector
+        let mut mean = Mean::<i32>::default();
+        let v: Vec<VectorRef> = vec![Arc::new(ConstantVector::new(
+            Arc::new(Int32Vector::from_vec(vec![4])),
+            10,
+        ))];
+        mean.update_batch(&v).unwrap();
+        assert_eq!(Value::from(4.0), mean.evaluate().unwrap());
+    }
+}
--- a/src/common/function/src/scalars/aggregate/polyval.rs
+++ b/src/common/function/src/scalars/aggregate/polyval.rs
@@ -0,0 +1,329 @@
+// 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::marker::PhantomData;
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    self, BadAccumulatorImplSnafu, CreateAccumulatorSnafu, DowncastVectorSnafu,
+    FromScalarValueSnafu, InvalidInputColSnafu, InvalidInputStateSnafu, Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::types::{LogicalPrimitiveType, WrapperType};
+use datatypes::value::ListValue;
+use datatypes::vectors::{ConstantVector, Helper, Int64Vector, ListVector};
+use datatypes::with_match_primitive_type_id;
+use num_traits::AsPrimitive;
+use snafu::{ensure, OptionExt, ResultExt};
+
+// https://numpy.org/doc/stable/reference/generated/numpy.polyval.html
+#[derive(Debug, Default)]
+pub struct Polyval<T, PolyT>
+where
+    T: WrapperType,
+    T::Native: AsPrimitive<PolyT::Native>,
+    PolyT: WrapperType,
+    PolyT::Native: std::ops::Mul<Output = PolyT::Native>,
+{
+    values: Vec<T>,
+    // DataFusion casts constant in into i64 type.
+    x: Option<i64>,
+    _phantom: PhantomData<PolyT>,
+}
+
+impl<T, PolyT> Polyval<T, PolyT>
+where
+    T: WrapperType,
+    T::Native: AsPrimitive<PolyT::Native>,
+    PolyT: WrapperType,
+    PolyT::Native: std::ops::Mul<Output = PolyT::Native>,
+{
+    fn push(&mut self, value: T) {
+        self.values.push(value);
+    }
+}
+
+impl<T, PolyT> Accumulator for Polyval<T, PolyT>
+where
+    T: WrapperType,
+    T::Native: AsPrimitive<PolyT::Native>,
+    PolyT: WrapperType + std::iter::Sum<<PolyT as WrapperType>::Native>,
+    PolyT::Native: std::ops::Mul<Output = PolyT::Native> + std::iter::Sum<PolyT::Native>,
+    i64: AsPrimitive<<PolyT as WrapperType>::Native>,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        let nums = self
+            .values
+            .iter()
+            .map(|&n| n.into())
+            .collect::<Vec<Value>>();
+        Ok(vec![
+            Value::List(ListValue::new(nums, T::LogicalType::build_data_type())),
+            self.x.into(),
+        ])
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(values.len() == 2, InvalidInputStateSnafu);
+        ensure!(values[0].len() == values[1].len(), InvalidInputStateSnafu);
+        if values[0].len() == 0 {
+            return Ok(());
+        }
+        // This is a unary accumulator, so only one column is provided.
+        let column = &values[0];
+        let mut len = 1;
+        let column: &<T as Scalar>::VectorType = if column.is_const() {
+            len = column.len();
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+        (0..len).for_each(|_| {
+            for v in column.iter_data().flatten() {
+                self.push(v);
+            }
+        });
+
+        let x = &values[1];
+        let x = Helper::check_get_scalar::<i64>(x).context(error::InvalidInputTypeSnafu {
+            err_msg: "expecting \"POLYVAL\" function's second argument to be a positive integer",
+        })?;
+        // `get(0)` is safe because we have checked `values[1].len() == values[0].len() != 0`
+        let first = x.get(0);
+        ensure!(!first.is_null(), InvalidInputColSnafu);
+
+        for i in 1..x.len() {
+            ensure!(first == x.get(i), InvalidInputColSnafu);
+        }
+
+        let first = match first {
+            Value::Int64(v) => v,
+            // unreachable because we have checked `first` is not null and is i64 above
+            _ => unreachable!(),
+        };
+        if let Some(x) = self.x {
+            ensure!(x == first, InvalidInputColSnafu);
+        } else {
+            self.x = Some(first);
+        };
+        Ok(())
+    }
+
+    // DataFusion executes accumulators in partitions. In some execution stage, DataFusion will
+    // merge states from other accumulators (returned by `state()` method).
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(
+            states.len() == 2,
+            BadAccumulatorImplSnafu {
+                err_msg: "expect 2 states in `merge_batch`",
+            }
+        );
+
+        let x = &states[1];
+        let x = x
+            .as_any()
+            .downcast_ref::<Int64Vector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect Int64Vector, got vector type {}",
+                    x.vector_type_name()
+                ),
+            })?;
+        let x = x.get(0);
+        if x.is_null() {
+            return Ok(());
+        }
+        let x = match x {
+            Value::Int64(x) => x,
+            _ => unreachable!(),
+        };
+        self.x = Some(x);
+
+        let values = &states[0];
+        let values = values
+            .as_any()
+            .downcast_ref::<ListVector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect ListVector, got vector type {}",
+                    values.vector_type_name()
+                ),
+            })?;
+        for value in values.values_iter() {
+            if let Some(value) = value.context(FromScalarValueSnafu)? {
+                let column: &<T as Scalar>::VectorType = unsafe { Helper::static_cast(&value) };
+                for v in column.iter_data().flatten() {
+                    self.push(v);
+                }
+            }
+        }
+
+        Ok(())
+    }
+
+    // DataFusion expects this function to return the final value of this aggregator.
+    fn evaluate(&self) -> Result<Value> {
+        if self.values.is_empty() {
+            return Ok(Value::Null);
+        }
+        let x = if let Some(x) = self.x {
+            x
+        } else {
+            return Ok(Value::Null);
+        };
+        let len = self.values.len();
+        let polyval: PolyT = self
+            .values
+            .iter()
+            .enumerate()
+            .map(|(i, &value)| value.into_native().as_() * x.pow((len - 1 - i) as u32).as_())
+            .sum();
+        Ok(polyval.into())
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct PolyvalAccumulatorCreator {}
+
+impl AggregateFunctionCreator for PolyvalAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(Polyval::<<$S as LogicalPrimitiveType>::Wrapper, <<$S as LogicalPrimitiveType>::LargestType as LogicalPrimitiveType>::Wrapper>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"POLYVAL\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 2, InvalidInputStateSnafu);
+        let input_type = self.input_types()?[0].logical_type_id();
+        with_match_primitive_type_id!(
+            input_type,
+            |$S| {
+                Ok(<<$S as LogicalPrimitiveType>::LargestType as LogicalPrimitiveType>::build_data_type())
+            },
+            {
+                unreachable!()
+            }
+        )
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 2, InvalidInputStateSnafu);
+        Ok(vec![
+            ConcreteDataType::list_datatype(input_types.into_iter().next().unwrap()),
+            ConcreteDataType::int64_datatype(),
+        ])
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::Int32Vector;
+
+    use super::*;
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut polyval = Polyval::<i32, i64>::default();
+        polyval.update_batch(&[]).unwrap();
+        assert!(polyval.values.is_empty());
+        assert_eq!(Value::Null, polyval.evaluate().unwrap());
+
+        // test update one not-null value
+        let mut polyval = Polyval::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(3)])),
+            Arc::new(Int64Vector::from(vec![Some(2_i64)])),
+        ];
+        polyval.update_batch(&v).unwrap();
+        assert_eq!(Value::Int64(3), polyval.evaluate().unwrap());
+
+        // test update one null value
+        let mut polyval = Polyval::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Option::<i32>::None])),
+            Arc::new(Int64Vector::from(vec![Some(2_i64)])),
+        ];
+        polyval.update_batch(&v).unwrap();
+        assert_eq!(Value::Null, polyval.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut polyval = Polyval::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(3), Some(0), Some(1)])),
+            Arc::new(Int64Vector::from(vec![
+                Some(2_i64),
+                Some(2_i64),
+                Some(2_i64),
+            ])),
+        ];
+        polyval.update_batch(&v).unwrap();
+        assert_eq!(Value::Int64(13), polyval.evaluate().unwrap());
+
+        // test update null-value batch
+        let mut polyval = Polyval::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(3), Some(0), None, Some(1)])),
+            Arc::new(Int64Vector::from(vec![
+                Some(2_i64),
+                Some(2_i64),
+                Some(2_i64),
+                Some(2_i64),
+            ])),
+        ];
+        polyval.update_batch(&v).unwrap();
+        assert_eq!(Value::Int64(13), polyval.evaluate().unwrap());
+
+        // test update with constant vector
+        let mut polyval = Polyval::<i32, i64>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(ConstantVector::new(
+                Arc::new(Int32Vector::from_vec(vec![4])),
+                2,
+            )),
+            Arc::new(Int64Vector::from(vec![Some(5_i64), Some(5_i64)])),
+        ];
+        polyval.update_batch(&v).unwrap();
+        assert_eq!(Value::Int64(24), polyval.evaluate().unwrap());
+    }
+}
--- a/src/common/function/src/scalars/aggregate/scipy_stats_norm_cdf.rs
+++ b/src/common/function/src/scalars/aggregate/scipy_stats_norm_cdf.rs
@@ -0,0 +1,270 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    self, BadAccumulatorImplSnafu, CreateAccumulatorSnafu, DowncastVectorSnafu,
+    FromScalarValueSnafu, GenerateFunctionSnafu, InvalidInputColSnafu, InvalidInputStateSnafu,
+    Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::value::{ListValue, OrderedFloat};
+use datatypes::vectors::{ConstantVector, Float64Vector, Helper, ListVector};
+use datatypes::with_match_primitive_type_id;
+use num_traits::AsPrimitive;
+use snafu::{ensure, OptionExt, ResultExt};
+use statrs::distribution::{ContinuousCDF, Normal};
+use statrs::statistics::Statistics;
+
+// https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.norm.html
+
+#[derive(Debug, Default)]
+pub struct ScipyStatsNormCdf<T> {
+    values: Vec<T>,
+    x: Option<f64>,
+}
+
+impl<T> ScipyStatsNormCdf<T> {
+    fn push(&mut self, value: T) {
+        self.values.push(value);
+    }
+}
+
+impl<T> Accumulator for ScipyStatsNormCdf<T>
+where
+    T: WrapperType + std::iter::Sum<T>,
+    T::Native: AsPrimitive<f64>,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        let nums = self
+            .values
+            .iter()
+            .map(|&x| x.into())
+            .collect::<Vec<Value>>();
+        Ok(vec![
+            Value::List(ListValue::new(nums, T::LogicalType::build_data_type())),
+            self.x.into(),
+        ])
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(values.len() == 2, InvalidInputStateSnafu);
+        ensure!(values[1].len() == values[0].len(), InvalidInputStateSnafu);
+
+        if values[0].len() == 0 {
+            return Ok(());
+        }
+
+        let column = &values[0];
+        let mut len = 1;
+        let column: &<T as Scalar>::VectorType = if column.is_const() {
+            len = column.len();
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+
+        let x = &values[1];
+        let x = Helper::check_get_scalar::<f64>(x).context(error::InvalidInputTypeSnafu {
+            err_msg: "expecting \"SCIPYSTATSNORMCDF\" function's second argument to be a positive integer",
+        })?;
+        let first = x.get(0);
+        ensure!(!first.is_null(), InvalidInputColSnafu);
+        let first = match first {
+            Value::Float64(OrderedFloat(v)) => v,
+            // unreachable because we have checked `first` is not null and is i64 above
+            _ => unreachable!(),
+        };
+        if let Some(x) = self.x {
+            ensure!(x == first, InvalidInputColSnafu);
+        } else {
+            self.x = Some(first);
+        };
+
+        (0..len).for_each(|_| {
+            for v in column.iter_data().flatten() {
+                self.push(v);
+            }
+        });
+        Ok(())
+    }
+
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(
+            states.len() == 2,
+            BadAccumulatorImplSnafu {
+                err_msg: "expect 2 states in `merge_batch`",
+            }
+        );
+
+        let x = &states[1];
+        let x = x
+            .as_any()
+            .downcast_ref::<Float64Vector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect Float64Vector, got vector type {}",
+                    x.vector_type_name()
+                ),
+            })?;
+        let x = x.get(0);
+        if x.is_null() {
+            return Ok(());
+        }
+        let x = match x {
+            Value::Float64(OrderedFloat(x)) => x,
+            _ => unreachable!(),
+        };
+        self.x = Some(x);
+
+        let values = &states[0];
+        let values = values
+            .as_any()
+            .downcast_ref::<ListVector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect ListVector, got vector type {}",
+                    values.vector_type_name()
+                ),
+            })?;
+        for value in values.values_iter() {
+            if let Some(value) = value.context(FromScalarValueSnafu)? {
+                let column: &<T as Scalar>::VectorType = unsafe { Helper::static_cast(&value) };
+                for v in column.iter_data().flatten() {
+                    self.push(v);
+                }
+            }
+        }
+        Ok(())
+    }
+
+    fn evaluate(&self) -> Result<Value> {
+        let mean = self.values.iter().map(|v| v.into_native().as_()).mean();
+        let std_dev = self.values.iter().map(|v| v.into_native().as_()).std_dev();
+        if mean.is_nan() || std_dev.is_nan() {
+            Ok(Value::Null)
+        } else {
+            let x = if let Some(x) = self.x {
+                x
+            } else {
+                return Ok(Value::Null);
+            };
+            let n = Normal::new(mean, std_dev).context(GenerateFunctionSnafu)?;
+            Ok(n.cdf(x).into())
+        }
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct ScipyStatsNormCdfAccumulatorCreator {}
+
+impl AggregateFunctionCreator for ScipyStatsNormCdfAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(ScipyStatsNormCdf::<<$S as LogicalPrimitiveType>::Wrapper>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"SCIPYSTATSNORMCDF\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 2, InvalidInputStateSnafu);
+        Ok(ConcreteDataType::float64_datatype())
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 2, InvalidInputStateSnafu);
+        Ok(vec![
+            ConcreteDataType::list_datatype(input_types[0].clone()),
+            ConcreteDataType::float64_datatype(),
+        ])
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::{Float64Vector, Int32Vector};
+
+    use super::*;
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut scipy_stats_norm_cdf = ScipyStatsNormCdf::<i32>::default();
+        scipy_stats_norm_cdf.update_batch(&[]).unwrap();
+        assert!(scipy_stats_norm_cdf.values.is_empty());
+        assert_eq!(Value::Null, scipy_stats_norm_cdf.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut scipy_stats_norm_cdf = ScipyStatsNormCdf::<i32>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(-1i32), Some(1), Some(2)])),
+            Arc::new(Float64Vector::from(vec![
+                Some(2.0_f64),
+                Some(2.0_f64),
+                Some(2.0_f64),
+            ])),
+        ];
+        scipy_stats_norm_cdf.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::from(0.8086334555398362),
+            scipy_stats_norm_cdf.evaluate().unwrap()
+        );
+
+        // test update null-value batch
+        let mut scipy_stats_norm_cdf = ScipyStatsNormCdf::<i32>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(-2i32), None, Some(3), Some(4)])),
+            Arc::new(Float64Vector::from(vec![
+                Some(2.0_f64),
+                None,
+                Some(2.0_f64),
+                Some(2.0_f64),
+            ])),
+        ];
+        scipy_stats_norm_cdf.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::from(0.5412943699039795),
+            scipy_stats_norm_cdf.evaluate().unwrap()
+        );
+    }
+}
--- a/src/common/function/src/scalars/aggregate/scipy_stats_norm_pdf.rs
+++ b/src/common/function/src/scalars/aggregate/scipy_stats_norm_pdf.rs
@@ -0,0 +1,271 @@
+// Copyright 2023 Greptime Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+
+use common_macro::{as_aggr_func_creator, AggrFuncTypeStore};
+use common_query::error::{
+    self, BadAccumulatorImplSnafu, CreateAccumulatorSnafu, DowncastVectorSnafu,
+    FromScalarValueSnafu, GenerateFunctionSnafu, InvalidInputColSnafu, InvalidInputStateSnafu,
+    Result,
+};
+use common_query::logical_plan::accumulator::AggrFuncTypeStore;
+use common_query::logical_plan::{Accumulator, AggregateFunctionCreator};
+use common_query::prelude::*;
+use datatypes::prelude::*;
+use datatypes::value::{ListValue, OrderedFloat};
+use datatypes::vectors::{ConstantVector, Float64Vector, Helper, ListVector};
+use datatypes::with_match_primitive_type_id;
+use num_traits::AsPrimitive;
+use snafu::{ensure, OptionExt, ResultExt};
+use statrs::distribution::{Continuous, Normal};
+use statrs::statistics::Statistics;
+
+// https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.norm.html
+
+#[derive(Debug, Default)]
+pub struct ScipyStatsNormPdf<T> {
+    values: Vec<T>,
+    x: Option<f64>,
+}
+
+impl<T> ScipyStatsNormPdf<T> {
+    fn push(&mut self, value: T) {
+        self.values.push(value);
+    }
+}
+
+impl<T> Accumulator for ScipyStatsNormPdf<T>
+where
+    T: WrapperType,
+    T::Native: AsPrimitive<f64> + std::iter::Sum<T>,
+{
+    fn state(&self) -> Result<Vec<Value>> {
+        let nums = self
+            .values
+            .iter()
+            .map(|&x| x.into())
+            .collect::<Vec<Value>>();
+        Ok(vec![
+            Value::List(ListValue::new(nums, T::LogicalType::build_data_type())),
+            self.x.into(),
+        ])
+    }
+
+    fn update_batch(&mut self, values: &[VectorRef]) -> Result<()> {
+        if values.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(values.len() == 2, InvalidInputStateSnafu);
+        ensure!(values[1].len() == values[0].len(), InvalidInputStateSnafu);
+
+        if values[0].len() == 0 {
+            return Ok(());
+        }
+
+        let column = &values[0];
+        let mut len = 1;
+        let column: &<T as Scalar>::VectorType = if column.is_const() {
+            len = column.len();
+            let column: &ConstantVector = unsafe { Helper::static_cast(column) };
+            unsafe { Helper::static_cast(column.inner()) }
+        } else {
+            unsafe { Helper::static_cast(column) }
+        };
+
+        let x = &values[1];
+        let x = Helper::check_get_scalar::<f64>(x).context(error::InvalidInputTypeSnafu {
+            err_msg: "expecting \"SCIPYSTATSNORMPDF\" function's second argument to be a positive integer",
+        })?;
+        let first = x.get(0);
+        ensure!(!first.is_null(), InvalidInputColSnafu);
+        let first = match first {
+            Value::Float64(OrderedFloat(v)) => v,
+            // unreachable because we have checked `first` is not null and is i64 above
+            _ => unreachable!(),
+        };
+        if let Some(x) = self.x {
+            ensure!(x == first, InvalidInputColSnafu);
+        } else {
+            self.x = Some(first);
+        };
+
+        (0..len).for_each(|_| {
+            for v in column.iter_data().flatten() {
+                self.push(v);
+            }
+        });
+        Ok(())
+    }
+
+    fn merge_batch(&mut self, states: &[VectorRef]) -> Result<()> {
+        if states.is_empty() {
+            return Ok(());
+        }
+
+        ensure!(
+            states.len() == 2,
+            BadAccumulatorImplSnafu {
+                err_msg: "expect 2 states in `merge_batch`",
+            }
+        );
+
+        let x = &states[1];
+        let x = x
+            .as_any()
+            .downcast_ref::<Float64Vector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect Float64Vector, got vector type {}",
+                    x.vector_type_name()
+                ),
+            })?;
+        let x = x.get(0);
+        if x.is_null() {
+            return Ok(());
+        }
+        let x = match x {
+            Value::Float64(OrderedFloat(x)) => x,
+            _ => unreachable!(),
+        };
+        self.x = Some(x);
+
+        let values = &states[0];
+        let values = values
+            .as_any()
+            .downcast_ref::<ListVector>()
+            .with_context(|| DowncastVectorSnafu {
+                err_msg: format!(
+                    "expect ListVector, got vector type {}",
+                    values.vector_type_name()
+                ),
+            })?;
+        for value in values.values_iter() {
+            if let Some(value) = value.context(FromScalarValueSnafu)? {
+                let column: &<T as Scalar>::VectorType = unsafe { Helper::static_cast(&value) };
+                for v in column.iter_data().flatten() {
+                    self.push(v);
+                }
+            }
+        }
+        Ok(())
+    }
+
+    fn evaluate(&self) -> Result<Value> {
+        let mean = self.values.iter().map(|v| v.into_native().as_()).mean();
+        let std_dev = self.values.iter().map(|v| v.into_native().as_()).std_dev();
+
+        if mean.is_nan() || std_dev.is_nan() {
+            Ok(Value::Null)
+        } else {
+            let x = if let Some(x) = self.x {
+                x
+            } else {
+                return Ok(Value::Null);
+            };
+            let n = Normal::new(mean, std_dev).context(GenerateFunctionSnafu)?;
+            Ok(n.pdf(x).into())
+        }
+    }
+}
+
+#[as_aggr_func_creator]
+#[derive(Debug, Default, AggrFuncTypeStore)]
+pub struct ScipyStatsNormPdfAccumulatorCreator {}
+
+impl AggregateFunctionCreator for ScipyStatsNormPdfAccumulatorCreator {
+    fn creator(&self) -> AccumulatorCreatorFunction {
+        let creator: AccumulatorCreatorFunction = Arc::new(move |types: &[ConcreteDataType]| {
+            let input_type = &types[0];
+            with_match_primitive_type_id!(
+                input_type.logical_type_id(),
+                |$S| {
+                    Ok(Box::new(ScipyStatsNormPdf::<<$S as LogicalPrimitiveType>::Wrapper>::default()))
+                },
+                {
+                    let err_msg = format!(
+                        "\"SCIPYSTATSNORMpdf\" aggregate function not support data type {:?}",
+                        input_type.logical_type_id(),
+                    );
+                    CreateAccumulatorSnafu { err_msg }.fail()?
+                }
+            )
+        });
+        creator
+    }
+
+    fn output_type(&self) -> Result<ConcreteDataType> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 2, InvalidInputStateSnafu);
+        Ok(ConcreteDataType::float64_datatype())
+    }
+
+    fn state_types(&self) -> Result<Vec<ConcreteDataType>> {
+        let input_types = self.input_types()?;
+        ensure!(input_types.len() == 2, InvalidInputStateSnafu);
+        Ok(vec![
+            ConcreteDataType::list_datatype(input_types[0].clone()),
+            ConcreteDataType::float64_datatype(),
+        ])
+    }
+}
+
+#[cfg(test)]
+mod test {
+    use datatypes::vectors::{Float64Vector, Int32Vector};
+
+    use super::*;
+    #[test]
+    fn test_update_batch() {
+        // test update empty batch, expect not updating anything
+        let mut scipy_stats_norm_pdf = ScipyStatsNormPdf::<i32>::default();
+        scipy_stats_norm_pdf.update_batch(&[]).unwrap();
+        assert!(scipy_stats_norm_pdf.values.is_empty());
+        assert_eq!(Value::Null, scipy_stats_norm_pdf.evaluate().unwrap());
+
+        // test update no null-value batch
+        let mut scipy_stats_norm_pdf = ScipyStatsNormPdf::<i32>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(-1i32), Some(1), Some(2)])),
+            Arc::new(Float64Vector::from(vec![
+                Some(2.0_f64),
+                Some(2.0_f64),
+                Some(2.0_f64),
+            ])),
+        ];
+        scipy_stats_norm_pdf.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::from(0.17843340219081558),
+            scipy_stats_norm_pdf.evaluate().unwrap()
+        );
+
+        // test update null-value batch
+        let mut scipy_stats_norm_pdf = ScipyStatsNormPdf::<i32>::default();
+        let v: Vec<VectorRef> = vec![
+            Arc::new(Int32Vector::from(vec![Some(-2i32), None, Some(3), Some(4)])),
+            Arc::new(Float64Vector::from(vec![
+                Some(2.0_f64),
+                None,
+                Some(2.0_f64),
+                Some(2.0_f64),
+            ])),
+        ];
+        scipy_stats_norm_pdf.update_batch(&v).unwrap();
+        assert_eq!(
+            Value::from(0.12343972049858312),
+            scipy_stats_norm_pdf.evaluate().unwrap()
+        );
+    }
+}
--- a/src/common/function/src/scalars/date/date_add.rs
+++ b/src/common/function/src/scalars/date/date_add.rs
@@ -58,7 +58,7 @@ impl Function for DateAddFunction {
        )
    }

-    fn eval(&self, _func_ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
+    fn eval(&self, _func_ctx: FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
        ensure!(
            columns.len() == 2,
            InvalidFuncArgsSnafu {
@@ -146,7 +146,7 @@ mod tests {
        let time_vector = TimestampSecondVector::from(times.clone());
        let interval_vector = IntervalDayTimeVector::from_vec(intervals);
        let args: Vec<VectorRef> = vec![Arc::new(time_vector), Arc::new(interval_vector)];
-        let vector = f.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();

        assert_eq!(4, vector.len());
        for (i, _t) in times.iter().enumerate() {
@@ -178,7 +178,7 @@ mod tests {
        let date_vector = DateVector::from(dates.clone());
        let interval_vector = IntervalYearMonthVector::from_vec(intervals);
        let args: Vec<VectorRef> = vec![Arc::new(date_vector), Arc::new(interval_vector)];
-        let vector = f.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();

        assert_eq!(4, vector.len());
        for (i, _t) in dates.iter().enumerate() {
--- a/src/common/function/src/scalars/date/date_format.rs
+++ b/src/common/function/src/scalars/date/date_format.rs
@@ -43,6 +43,7 @@ impl Function for DateFormatFunction {
        helper::one_of_sigs2(
            vec![
                ConcreteDataType::date_datatype(),
+                ConcreteDataType::datetime_datatype(),
                ConcreteDataType::timestamp_second_datatype(),
                ConcreteDataType::timestamp_millisecond_datatype(),
                ConcreteDataType::timestamp_microsecond_datatype(),
@@ -52,7 +53,7 @@ impl Function for DateFormatFunction {
        )
    }

-    fn eval(&self, func_ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
+    fn eval(&self, func_ctx: FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
        ensure!(
            columns.len() == 2,
            InvalidFuncArgsSnafu {
@@ -104,6 +105,22 @@ impl Function for DateFormatFunction {
                    results.push(result.as_deref());
                }
            }
+            ConcreteDataType::DateTime(_) => {
+                for i in 0..size {
+                    let datetime = left.get(i).as_datetime();
+                    let format = formats.get(i).as_string();
+
+                    let result = match (datetime, format) {
+                        (Some(datetime), Some(fmt)) => datetime
+                            .as_formatted_string(&fmt, Some(&func_ctx.query_ctx.timezone()))
+                            .map_err(BoxedError::new)
+                            .context(error::ExecuteSnafu)?,
+                        _ => None,
+                    };
+
+                    results.push(result.as_deref());
+                }
+            }
            _ => {
                return UnsupportedInputDataTypeSnafu {
                    function: NAME,
@@ -130,7 +147,7 @@ mod tests {
    use common_query::prelude::{TypeSignature, Volatility};
    use datatypes::prelude::{ConcreteDataType, ScalarVector};
    use datatypes::value::Value;
-    use datatypes::vectors::{DateVector, StringVector, TimestampSecondVector};
+    use datatypes::vectors::{DateTimeVector, DateVector, StringVector, TimestampSecondVector};

    use super::{DateFormatFunction, *};

@@ -152,11 +169,16 @@ mod tests {
            ConcreteDataType::string_datatype(),
            f.return_type(&[ConcreteDataType::date_datatype()]).unwrap()
        );
+        assert_eq!(
+            ConcreteDataType::string_datatype(),
+            f.return_type(&[ConcreteDataType::datetime_datatype()])
+                .unwrap()
+        );
        assert!(matches!(f.signature(),
                         Signature {
                             type_signature: TypeSignature::OneOf(sigs),
                             volatility: Volatility::Immutable
-                         } if  sigs.len() == 5));
+                         } if  sigs.len() == 6));
    }

    #[test]
@@ -180,7 +202,7 @@ mod tests {
        let time_vector = TimestampSecondVector::from(times.clone());
        let interval_vector = StringVector::from_vec(formats);
        let args: Vec<VectorRef> = vec![Arc::new(time_vector), Arc::new(interval_vector)];
-        let vector = f.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();

        assert_eq!(4, vector.len());
        for (i, _t) in times.iter().enumerate() {
@@ -221,7 +243,48 @@ mod tests {
        let date_vector = DateVector::from(dates.clone());
        let interval_vector = StringVector::from_vec(formats);
        let args: Vec<VectorRef> = vec![Arc::new(date_vector), Arc::new(interval_vector)];
-        let vector = f.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();
+
+        assert_eq!(4, vector.len());
+        for (i, _t) in dates.iter().enumerate() {
+            let v = vector.get(i);
+            let result = results.get(i).unwrap();
+
+            if result.is_none() {
+                assert_eq!(Value::Null, v);
+                continue;
+            }
+            match v {
+                Value::String(s) => {
+                    assert_eq!(s.as_utf8(), result.unwrap());
+                }
+                _ => unreachable!(),
+            }
+        }
+    }
+
+    #[test]
+    fn test_datetime_date_format() {
+        let f = DateFormatFunction;
+
+        let dates = vec![Some(123), None, Some(42), None];
+        let formats = vec![
+            "%Y-%m-%d %T.%3f",
+            "%Y-%m-%d %T.%3f",
+            "%Y-%m-%d %T.%3f",
+            "%Y-%m-%d %T.%3f",
+        ];
+        let results = [
+            Some("1970-01-01 00:00:00.123"),
+            None,
+            Some("1970-01-01 00:00:00.042"),
+            None,
+        ];
+
+        let date_vector = DateTimeVector::from(dates.clone());
+        let interval_vector = StringVector::from_vec(formats);
+        let args: Vec<VectorRef> = vec![Arc::new(date_vector), Arc::new(interval_vector)];
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();

        assert_eq!(4, vector.len());
        for (i, _t) in dates.iter().enumerate() {
--- a/src/common/function/src/scalars/date/date_sub.rs
+++ b/src/common/function/src/scalars/date/date_sub.rs
@@ -58,7 +58,7 @@ impl Function for DateSubFunction {
        )
    }

-    fn eval(&self, _func_ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
+    fn eval(&self, _func_ctx: FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
        ensure!(
            columns.len() == 2,
            InvalidFuncArgsSnafu {
@@ -118,6 +118,11 @@ mod tests {
            ConcreteDataType::date_datatype(),
            f.return_type(&[ConcreteDataType::date_datatype()]).unwrap()
        );
+        assert_eq!(
+            ConcreteDataType::datetime_datatype(),
+            f.return_type(&[ConcreteDataType::datetime_datatype()])
+                .unwrap()
+        );
        assert!(
            matches!(f.signature(),
                         Signature {
@@ -146,7 +151,7 @@ mod tests {
        let time_vector = TimestampSecondVector::from(times.clone());
        let interval_vector = IntervalDayTimeVector::from_vec(intervals);
        let args: Vec<VectorRef> = vec![Arc::new(time_vector), Arc::new(interval_vector)];
-        let vector = f.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();

        assert_eq!(4, vector.len());
        for (i, _t) in times.iter().enumerate() {
@@ -184,7 +189,7 @@ mod tests {
        let date_vector = DateVector::from(dates.clone());
        let interval_vector = IntervalYearMonthVector::from_vec(intervals);
        let args: Vec<VectorRef> = vec![Arc::new(date_vector), Arc::new(interval_vector)];
-        let vector = f.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = f.eval(FunctionContext::default(), &args).unwrap();

        assert_eq!(4, vector.len());
        for (i, _t) in dates.iter().enumerate() {
--- a/src/common/function/src/scalars/expression/is_null.rs
+++ b/src/common/function/src/scalars/expression/is_null.rs
@@ -55,7 +55,7 @@ impl Function for IsNullFunction {

    fn eval(
        &self,
-        _func_ctx: &FunctionContext,
+        _func_ctx: FunctionContext,
        columns: &[VectorRef],
    ) -> common_query::error::Result<VectorRef> {
        ensure!(
@@ -102,7 +102,7 @@ mod tests {
        let values = vec![None, Some(3.0), None];

        let args: Vec<VectorRef> = vec![Arc::new(Float32Vector::from(values))];
-        let vector = is_null.eval(&FunctionContext::default(), &args).unwrap();
+        let vector = is_null.eval(FunctionContext::default(), &args).unwrap();
        let expect: VectorRef = Arc::new(BooleanVector::from_vec(vec![true, false, true]));
        assert_eq!(expect, vector);
    }
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Lei, HUANG	68593ae92a	### Added Deduplication and Merge Functionality - Introduced `DedupReader` and `MergeReader` in `src/mito2/src/read/sync/dedup.rs` and `src/mito2/src/read/sync/merge.rs` to handle deduplication and merging of sorted batches. ### Enhanced `BulkMemtable` Iteration - Updated `BulkMemtable` in `src/mito2/src/memtable/bulk.rs` to support deduplication and merge modes during iteration. - Added `BulkIterContext` to manage iteration context. ### Testing Enhancements - Added comprehensive tests for `BulkMemtable` and `BulkPart` in `src/mito2/src/memtable/bulk.rs` and `src/mito2/src/memtable/bulk/part.rs`. ### Code Refactoring - Made `BulkPart` and `BulkPartMeta` cloneable in `src/mito2/src/memtable/bulk/part.rs`. - Exposed internal test modules for better test coverage in `src/mito2/src/memtable/time_series.rs` and `src/mito2/src/read/merge.rs`. ### New Modules - Created `sync` module in `src/mito2/src/read.rs` to organize synchronous read operations.	2025-02-20 08:54:44 +00:00
evenyag	91d755d9b5	feat: use spawn	2025-02-13 22:49:02 +08:00
Lei, HUANG	2566d254ad	poc-write-path: Enhance BulkPart Metadata Handling - Updated `BulkMemtable` in `bulk.rs` to track and update `max_sequence`, `max_timestamp`, `min_timestamp`, and `num_rows` using `BulkPart` metadata. - Extended `BulkPartMeta` in `bulk/part.rs` to include `max_sequence`. - Modified `mutations_to_record_batch` function to return `max_sequence` along with timestamps in `bulk/part.rs`. - Adjusted `BulkPartEncoder` to handle the new `max_sequence` metadata in `bulk/part.rs`.	2025-02-10 15:14:45 +00:00
WenyXu	0ec4ed804d	fix: fix alter logic bug	2025-02-10 14:24:13 +00:00
evenyag	cc435234a4	chore: error handling in handle_batch_body	2025-02-10 21:38:10 +08:00
WenyXu	9c4aa81f85	fix: fix add column validation	2025-02-10 13:28:30 +00:00
evenyag	bdbb5435ea	feat: region server use handle_batch_body	2025-02-10 21:20:14 +08:00
WenyXu	fd9940a253	fix: fix compile	2025-02-10 13:18:01 +00:00
WenyXu	e0bafd661c	chore: unused error	2025-02-10 12:57:55 +00:00
WenyXu	99baa86b6a	feat(metric-engine): introduce batch alter request handling	2025-02-10 12:57:54 +00:00
WenyXu	76d69901ea	feat(metric-engine): introduce batch create request handling	2025-02-10 12:57:04 +00:00
WenyXu	764a57b80a	refactor: introduce RegionRequestBundle::Vector	2025-02-10 12:55:30 +00:00
WenyXu	95b388d819	refactor: refactor region server requests handling	2025-02-10 12:55:28 +00:00
evenyag	c2b556e321	feat: batch put logical regions	2025-02-10 19:32:21 +08:00
evenyag	06ebe6b3fb	feat: handle body	2025-02-10 19:32:14 +08:00
Lei, HUANG	bec8245e75	poc-write-path: Enhance Memtable Handling with Primary Key Encoding • Introduced PrimaryKeyEncoding to differentiate between dense and sparse primary key encodings. • Updated BulkMemtableBuilder to conditionally create memtables based on primary key encoding. • Integrated PartitionTreeMemtableBuilder as a fallback for dense encodings. • Modified RegionWriteCtx to handle mutations differently based on primary key encoding. • Adjusted RegionWorkerLoop to skip bulk encoding for dense primary key mutations. • Refactored SparseEncoder to support conditional compilation for testing purposes.	2025-02-10 14:38:11 +08:00
Lei, HUANG	3cb2343f7f	poc-write-path: ### Implement Sparse Primary Key Encoding - Added `SparseEncoder`: Introduced a new module `encoder.rs` to implement sparse primary key encoding, replacing the previous dense encoding approach. - Updated `BulkPartEncoder`: Modified `BulkPartEncoder` in `bulk/part.rs` to utilize `SparseEncoder` for encoding primary keys. - Refactored `PartitionTree`: Updated `partition_tree/tree.rs` to use the new `SparseEncoder` for primary key encoding. - Code Adjustments: Removed redundant code and adjusted imports in `key_values.rs` and `partition_tree/tree.rs` to align with the new encoding strategy.	2025-02-08 08:10:16 +00:00
Lei, HUANG	d10c207371	poc-write-path: Add allocation tracking to `BulkMemtable` methods - Updated `write_bulk` in `bulk.rs` to track memory allocation using `alloc_tracker.on_allocation`. - Modified `freeze` in `bulk.rs` to signal completion of allocation with `alloc_tracker.done_allocating`.	2025-02-06 09:17:37 +00:00
Lei, HUANG	1a73a40bd9	Merge remote-tracking branch 'GreptimeTeam/poc-write-path' into poc-write-path	2025-02-06 08:55:39 +00:00
evenyag	713a73e9b2	feat: encode bulk before writing wal * store bulk in wal * write the BulkPart to the memtable directly	2025-02-06 16:36:41 +08:00
Lei, HUANG	65a88a63db	poc-write-path: feat(memtable): Add BulkMemtableBuilder and BulkMemtable - Introduced `BulkMemtableBuilder` and `BulkMemtable` in `memtable.rs` and `bulk.rs` to support bulk operations. - Added environment variable check for `enable_bulk_memtable` to conditionally use `BulkMemtableBuilder`. - Implemented `MemtableBuilder` for `BulkMemtableBuilder` and `Memtable` for `BulkMemtable`. - Included new fields `dedup` and `merge_mode` in `BulkMemtable` to handle deduplication and merge operations. - Temporarily disabled reads in `BulkMemtable` with `EmptyIter` as a placeholder iterator.	2025-02-06 07:36:37 +00:00
Lei, HUANG	5ad1436a8f	Merge remote-tracking branch 'GreptimeTeam/poc-write-path' into poc-write-path	2025-02-06 06:44:33 +00:00
Lei, HUANG	ae59206caf	feat: impl stats	2025-02-06 06:38:44 +00:00
evenyag	094d0fcdf5	feat: add bulk to wal Mutation	2025-02-05 21:16:18 +08:00
evenyag	7170120de6	chore: update proto (add bulk to Mutation)	2025-02-05 21:05:01 +08:00