fix: last non null iter not init (#4687)

feat: invalidate table via invalidator on region migration

.env.example

@@ -14,10 +14,11 @@ GT_AZBLOB_CONTAINER=AZBLOB container
 GT_AZBLOB_ACCOUNT_NAME=AZBLOB account name
 GT_AZBLOB_ACCOUNT_KEY=AZBLOB account key
 GT_AZBLOB_ENDPOINT=AZBLOB endpoint
-# Settings for gcs test 
-GT_GCS_BUCKET = GCS bucket 
+# Settings for gcs test
+GT_GCS_BUCKET = GCS bucket
 GT_GCS_SCOPE  = GCS scope
-GT_GCS_CREDENTIAL_PATH = GCS credential path 
+GT_GCS_CREDENTIAL_PATH = GCS credential path
+GT_GCS_CREDENTIAL = GCS credential
 GT_GCS_ENDPOINT = GCS end point
 # Settings for kafka wal test
 GT_KAFKA_ENDPOINTS = localhost:9092

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

@@ -17,6 +17,12 @@ inputs:
     description: Enable dev mode, only build standard greptime
     required: false
     default: "false"
+  image-namespace:
+    description: Image Namespace
+    required: true
+  image-registry:
+    description: Image Registry
+    required: true
   working-dir:
     description: Working directory to build the artifacts
     required: false
@@ -31,8 +37,8 @@ runs:
       run: |
         cd ${{ inputs.working-dir }} && \
         make run-it-in-container BUILD_JOBS=4 \
-        IMAGE_NAMESPACE=i8k6a5e1/greptime \
-        IMAGE_REGISTRY=public.ecr.aws
+        IMAGE_NAMESPACE=${{ inputs.image-namespace }} \
+        IMAGE_REGISTRY=${{ inputs.image-registry }}
 
     - name: Upload sqlness logs
       if: ${{ failure() && inputs.disable-run-tests == 'false' }} # Only upload logs when the integration tests failed.
@@ -51,8 +57,8 @@ runs:
         artifacts-dir: greptime-linux-${{ inputs.arch }}-pyo3-${{ inputs.version }}
         version: ${{ inputs.version }}
         working-dir: ${{ inputs.working-dir }}
-        image-registry: public.ecr.aws
-        image-namespace: i8k6a5e1/greptime        
+        image-registry: ${{ inputs.image-registry }}
+        image-namespace: ${{ inputs.image-namespace }}
 
     - name: Build greptime without pyo3
       if: ${{ inputs.dev-mode == 'false' }}
@@ -64,8 +70,8 @@ runs:
         artifacts-dir: greptime-linux-${{ inputs.arch }}-${{ inputs.version }}
         version: ${{ inputs.version }}
         working-dir: ${{ inputs.working-dir }}
-        image-registry: public.ecr.aws
-        image-namespace: i8k6a5e1/greptime
+        image-registry: ${{ inputs.image-registry }}
+        image-namespace: ${{ inputs.image-namespace }}
 
     - name: Clean up the target directory # Clean up the target directory for the centos7 base image, or it will still use the objects of last build.
       shell: bash
@@ -82,8 +88,8 @@ runs:
         artifacts-dir: greptime-linux-${{ inputs.arch }}-centos-${{ inputs.version }}
         version: ${{ inputs.version }}
         working-dir: ${{ inputs.working-dir }}
-        image-registry: public.ecr.aws
-        image-namespace: i8k6a5e1/greptime
+        image-registry: ${{ inputs.image-registry }}
+        image-namespace: ${{ inputs.image-namespace }}
 
     - name: Build greptime on android base image
       uses: ./.github/actions/build-greptime-binary
@@ -94,5 +100,5 @@ runs:
         version: ${{ inputs.version }}
         working-dir: ${{ inputs.working-dir }}
         build-android-artifacts: true
-        image-registry: public.ecr.aws
-        image-namespace: i8k6a5e1/greptime
+        image-registry: ${{ inputs.image-registry }}
+        image-namespace: ${{ inputs.image-namespace }}

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

@@ -4,9 +4,6 @@ inputs:
   arch:
     description: Architecture to build
     required: true
-  rust-toolchain:
-    description: Rust toolchain to use
-    required: true
   cargo-profile:
     description: Cargo profile to build
     required: true
@@ -43,10 +40,9 @@ runs:
         brew install protobuf
 
     - name: Install rust toolchain
-      uses: dtolnay/rust-toolchain@master
+      uses: actions-rust-lang/setup-rust-toolchain@v1
       with:
-        toolchain: ${{ inputs.rust-toolchain }}
-        targets: ${{ inputs.arch }}
+        target: ${{ inputs.arch }}
 
     - name: Start etcd # For integration tests.
       if: ${{ inputs.disable-run-tests == 'false' }}

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

@@ -4,9 +4,6 @@ inputs:
   arch:
     description: Architecture to build
     required: true
-  rust-toolchain:
-    description: Rust toolchain to use
-    required: true
   cargo-profile:
     description: Cargo profile to build
     required: true
@@ -28,10 +25,9 @@ runs:
     - uses: arduino/setup-protoc@v3
 
     - name: Install rust toolchain
-      uses: dtolnay/rust-toolchain@master
+      uses: actions-rust-lang/setup-rust-toolchain@v1
       with:
-        toolchain: ${{ inputs.rust-toolchain }}
-        targets: ${{ inputs.arch }}
+        target: ${{ inputs.arch }}
         components: llvm-tools-preview
 
     - name: Rust Cache
@@ -69,7 +65,7 @@ runs:
       uses: actions/upload-artifact@v4
       with:
         name: sqlness-logs
-        path: C:\tmp\greptime-*.log
+        path: C:\Users\RUNNER~1\AppData\Local\Temp\sqlness*
         retention-days: 3
 
     - name: Build greptime binary

.github/actions/setup-etcd-cluster/action.yml

@@ -18,6 +18,8 @@ runs:
         --set replicaCount=${{ inputs.etcd-replicas }} \
         --set resources.requests.cpu=50m \
         --set resources.requests.memory=128Mi \
+        --set resources.limits.cpu=1000m \
+        --set resources.limits.memory=2Gi \
         --set auth.rbac.create=false \
         --set auth.rbac.token.enabled=false \
         --set persistence.size=2Gi \

.github/actions/setup-greptimedb-cluster/with-disk.yaml

@@ -1,13 +1,13 @@
 meta:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
 datanode:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
     compact_rt_size = 2
 frontend:
-  config: |-
+  configData: |-
     [runtime]
-    global_rt_size = 4
+    global_rt_size = 4

.github/actions/setup-greptimedb-cluster/with-minio-and-cache.yaml

@@ -1,5 +1,5 @@
 meta:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
 
@@ -7,7 +7,7 @@ meta:
     [datanode.client]
     timeout = "60s"
 datanode:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
     compact_rt_size = 2
@@ -16,7 +16,7 @@ datanode:
     cache_path = "/data/greptimedb/s3cache"
     cache_capacity = "256MB"
 frontend:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
 

.github/actions/setup-greptimedb-cluster/with-minio.yaml

@@ -1,5 +1,5 @@
 meta:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
     
@@ -7,12 +7,12 @@ meta:
     [datanode.client]
     timeout = "60s"
 datanode:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
     compact_rt_size = 2
 frontend:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
 

.github/actions/setup-greptimedb-cluster/with-remote-wal.yaml

@@ -1,5 +1,5 @@
 meta:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
     
@@ -13,7 +13,7 @@ meta:
     [datanode.client]
     timeout = "60s"
 datanode:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
     compact_rt_size = 2
@@ -23,7 +23,7 @@ datanode:
     broker_endpoints = ["kafka.kafka-cluster.svc.cluster.local:9092"]
     linger = "2ms"
 frontend:
-  config: |-
+  configData: |-
     [runtime]
     global_rt_size = 4
 

.github/actions/setup-postgres-cluster/action.yml

@@ -0,0 +1,30 @@
+name: Setup PostgreSQL
+description: Deploy PostgreSQL on Kubernetes
+inputs:
+  postgres-replicas:
+    default: 1
+    description: "Number of PostgreSQL replicas"
+  namespace:
+    default: "postgres-namespace"
+  postgres-version:
+    default: "14.2"
+    description: "PostgreSQL version"
+  storage-size:
+    default: "1Gi"
+    description: "Storage size for PostgreSQL"
+
+runs:
+  using: composite
+  steps:
+  - name: Install PostgreSQL
+    shell: bash
+    run: |
+      helm upgrade \
+        --install postgresql oci://registry-1.docker.io/bitnamicharts/postgresql \
+        --set replicaCount=${{ inputs.postgres-replicas }} \
+        --set image.tag=${{ inputs.postgres-version }} \
+        --set persistence.size=${{ inputs.storage-size }} \
+        --set postgresql.username=greptimedb \
+        --set postgresql.password=admin \
+        --create-namespace \
+        -n ${{ inputs.namespace }}

.github/actions/start-runner/action.yml

@@ -38,7 +38,7 @@ runs:
   steps:
     - name: Configure AWS credentials
       if: startsWith(inputs.runner, 'ec2')
-      uses: aws-actions/configure-aws-credentials@v2
+      uses: aws-actions/configure-aws-credentials@v4
       with:
         aws-access-key-id: ${{ inputs.aws-access-key-id }}
         aws-secret-access-key: ${{ inputs.aws-secret-access-key }}

.github/actions/stop-runner/action.yml

@@ -25,7 +25,7 @@ runs:
   steps:
     - name: Configure AWS credentials
       if: ${{ inputs.label && inputs.ec2-instance-id }}
-      uses: aws-actions/configure-aws-credentials@v2
+      uses: aws-actions/configure-aws-credentials@v4
       with:
         aws-access-key-id: ${{ inputs.aws-access-key-id }}
         aws-secret-access-key: ${{ inputs.aws-secret-access-key }}

.github/workflows/apidoc.yml

@@ -12,9 +12,6 @@ on:
 
 name: Build API docs
 
-env:
-  RUST_TOOLCHAIN: nightly-2024-04-20
-
 jobs:
   apidoc:
     runs-on: ubuntu-20.04
@@ -23,9 +20,7 @@ jobs:
     - uses: arduino/setup-protoc@v3
       with:
         repo-token: ${{ secrets.GITHUB_TOKEN }}
-    - uses: dtolnay/rust-toolchain@master
-      with:
-        toolchain: ${{ env.RUST_TOOLCHAIN }}
+    - uses: actions-rust-lang/setup-rust-toolchain@v1
     - run: cargo doc --workspace --no-deps --document-private-items
     - run: |
         cat <<EOF > target/doc/index.html

.github/workflows/dev-build.yml

@@ -177,6 +177,8 @@ jobs:
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
           dev-mode: true # Only build the standard greptime binary.
           working-dir: ${{ env.CHECKOUT_GREPTIMEDB_PATH }}
+          image-registry: ${{ vars.ECR_IMAGE_REGISTRY }}
+          image-namespace: ${{ vars.ECR_IMAGE_NAMESPACE }}
 
   build-linux-arm64-artifacts:
     name: Build linux-arm64 artifacts
@@ -206,6 +208,8 @@ jobs:
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
           dev-mode: true # Only build the standard greptime binary.
           working-dir: ${{ env.CHECKOUT_GREPTIMEDB_PATH }}
+          image-registry: ${{ vars.ECR_IMAGE_REGISTRY }}
+          image-namespace: ${{ vars.ECR_IMAGE_NAMESPACE }}
 
   release-images-to-dockerhub:
     name: Build and push images to DockerHub

.github/workflows/develop.yml

@@ -29,9 +29,6 @@ concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
   cancel-in-progress: true
 
-env:
-  RUST_TOOLCHAIN: nightly-2024-04-20
-
 jobs:
   check-typos-and-docs:
     name: Check typos and docs
@@ -64,9 +61,7 @@ jobs:
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
@@ -82,9 +77,7 @@ jobs:
     timeout-minutes: 60
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: stable
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
@@ -107,9 +100,7 @@ jobs:
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - uses: Swatinem/rust-cache@v2
         with:
           # Shares across multiple jobs
@@ -145,13 +136,23 @@ jobs:
       matrix:
         target: [ "fuzz_create_table", "fuzz_alter_table", "fuzz_create_database", "fuzz_create_logical_table", "fuzz_alter_logical_table", "fuzz_insert", "fuzz_insert_logical_table" ]
     steps:
+      - name: Remove unused software
+        run: |
+          echo "Disk space before:"
+          df -h
+          [[ -d /usr/share/dotnet ]] && sudo rm -rf /usr/share/dotnet
+          [[ -d /usr/local/lib/android ]] && sudo rm -rf /usr/local/lib/android
+          [[ -d /opt/ghc ]] && sudo rm -rf /opt/ghc
+          [[ -d /opt/hostedtoolcache/CodeQL ]] && sudo rm -rf /opt/hostedtoolcache/CodeQL
+          sudo docker image prune --all --force
+          sudo docker builder prune -a
+          echo "Disk space after:"
+          df -h
       - uses: actions/checkout@v4
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
@@ -169,7 +170,7 @@ jobs:
           name: bins
           path: .
       - name: Unzip binaries
-        run: | 
+        run: |
           tar -xvf ./bins.tar.gz
           rm ./bins.tar.gz
       - name: Run GreptimeDB
@@ -193,13 +194,23 @@ jobs:
       matrix:
         target: [ "unstable_fuzz_create_table_standalone" ]
     steps:
+      - name: Remove unused software
+        run: |
+          echo "Disk space before:"
+          df -h
+          [[ -d /usr/share/dotnet ]] && sudo rm -rf /usr/share/dotnet
+          [[ -d /usr/local/lib/android ]] && sudo rm -rf /usr/local/lib/android
+          [[ -d /opt/ghc ]] && sudo rm -rf /opt/ghc
+          [[ -d /opt/hostedtoolcache/CodeQL ]] && sudo rm -rf /opt/hostedtoolcache/CodeQL
+          sudo docker image prune --all --force
+          sudo docker builder prune -a
+          echo "Disk space after:"
+          df -h
       - uses: actions/checkout@v4
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
@@ -250,9 +261,7 @@ jobs:
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - uses: Swatinem/rust-cache@v2
         with:
           # Shares across multiple jobs
@@ -263,7 +272,7 @@ jobs:
       - name: Build greptime bianry
         shell: bash
         # `cargo gc` will invoke `cargo build` with specified args
-        run: cargo gc --profile ci -- --bin greptime 
+        run: cargo gc --profile ci -- --bin greptime
       - name: Pack greptime binary
         shell: bash
         run: |
@@ -277,7 +286,7 @@ jobs:
           artifacts-dir: bin
           version: current
 
-  distributed-fuzztest: 
+  distributed-fuzztest:
     name: Fuzz Test (Distributed, ${{ matrix.mode.name }}, ${{ matrix.target }})
     runs-on: ubuntu-latest
     needs:  build-greptime-ci
@@ -285,24 +294,24 @@ jobs:
     strategy:
       matrix:
         target: [ "fuzz_create_table", "fuzz_alter_table", "fuzz_create_database", "fuzz_create_logical_table", "fuzz_alter_logical_table", "fuzz_insert", "fuzz_insert_logical_table" ]
-        mode: 
-          - name: "Disk"
-            minio: false
-            kafka: false
-            values: "with-disk.yaml"
-          - name: "Minio"
-            minio: true
-            kafka: false
-            values: "with-minio.yaml"
-          - name: "Minio with Cache"
-            minio: true
-            kafka: false
-            values: "with-minio-and-cache.yaml"
+        mode:
           - name: "Remote WAL"
             minio: true
             kafka: true
             values: "with-remote-wal.yaml"
     steps:
+      - name: Remove unused software
+        run: |
+          echo "Disk space before:"
+          df -h
+          [[ -d /usr/share/dotnet ]] && sudo rm -rf /usr/share/dotnet
+          [[ -d /usr/local/lib/android ]] && sudo rm -rf /usr/local/lib/android
+          [[ -d /opt/ghc ]] && sudo rm -rf /opt/ghc
+          [[ -d /opt/hostedtoolcache/CodeQL ]] && sudo rm -rf /opt/hostedtoolcache/CodeQL
+          sudo docker image prune --all --force
+          sudo docker builder prune -a
+          echo "Disk space after:"
+          df -h
       - uses: actions/checkout@v4
       - name: Setup Kind
         uses: ./.github/actions/setup-kind
@@ -314,13 +323,13 @@ jobs:
         uses: ./.github/actions/setup-kafka-cluster
       - name: Setup Etcd cluser
         uses: ./.github/actions/setup-etcd-cluster
+      - name: Setup Postgres cluser
+        uses: ./.github/actions/setup-postgres-cluster
       # Prepares for fuzz tests
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
@@ -390,12 +399,12 @@ jobs:
       - name: Describe Nodes
         if: failure()
         shell: bash
-        run: | 
-          kubectl describe nodes      
+        run: |
+          kubectl describe nodes
       - name: Export kind logs
         if: failure()
         shell: bash
-        run: | 
+        run: |
           kind export logs /tmp/kind
       - name: Upload logs
         if: failure()
@@ -407,13 +416,13 @@ jobs:
       - name: Delete cluster
         if: success()
         shell: bash
-        run: | 
+        run: |
           kind delete cluster
           docker stop $(docker ps -a -q)
           docker rm $(docker ps -a -q)
           docker system prune -f
 
-  distributed-fuzztest-with-chaos: 
+  distributed-fuzztest-with-chaos:
     name: Fuzz Test with Chaos (Distributed, ${{ matrix.mode.name }}, ${{ matrix.target }})
     runs-on: ubuntu-latest
     needs:  build-greptime-ci
@@ -421,12 +430,24 @@ jobs:
     strategy:
       matrix:
         target: ["fuzz_migrate_mito_regions", "fuzz_failover_mito_regions", "fuzz_failover_metric_regions"]
-        mode: 
+        mode:
           - name: "Remote WAL"
             minio: true
             kafka: true
             values: "with-remote-wal.yaml"
     steps:
+      - name: Remove unused software
+        run: |
+          echo "Disk space before:"
+          df -h
+          [[ -d /usr/share/dotnet ]] && sudo rm -rf /usr/share/dotnet
+          [[ -d /usr/local/lib/android ]] && sudo rm -rf /usr/local/lib/android
+          [[ -d /opt/ghc ]] && sudo rm -rf /opt/ghc
+          [[ -d /opt/hostedtoolcache/CodeQL ]] && sudo rm -rf /opt/hostedtoolcache/CodeQL
+          sudo docker image prune --all --force
+          sudo docker builder prune -a
+          echo "Disk space after:"
+          df -h
       - uses: actions/checkout@v4
       - name: Setup Kind
         uses: ./.github/actions/setup-kind
@@ -440,13 +461,13 @@ jobs:
         uses: ./.github/actions/setup-kafka-cluster
       - name: Setup Etcd cluser
         uses: ./.github/actions/setup-etcd-cluster
+      - name: Setup Postgres cluser
+        uses: ./.github/actions/setup-postgres-cluster
       # Prepares for fuzz tests
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
@@ -517,12 +538,12 @@ jobs:
       - name: Describe Nodes
         if: failure()
         shell: bash
-        run: | 
-          kubectl describe nodes      
+        run: |
+          kubectl describe nodes
       - name: Export kind logs
         if: failure()
         shell: bash
-        run: | 
+        run: |
           kind export logs /tmp/kind
       - name: Upload logs
         if: failure()
@@ -534,7 +555,7 @@ jobs:
       - name: Delete cluster
         if: success()
         shell: bash
-        run: | 
+        run: |
           kind delete cluster
           docker stop $(docker ps -a -q)
           docker rm $(docker ps -a -q)
@@ -557,6 +578,10 @@ jobs:
     timeout-minutes: 60
     steps:
       - uses: actions/checkout@v4
+      - if: matrix.mode.kafka
+        name: Setup kafka server
+        working-directory: tests-integration/fixtures/kafka
+        run: docker compose -f docker-compose-standalone.yml up -d --wait
       - name: Download pre-built binaries
         uses: actions/download-artifact@v4
         with:
@@ -564,10 +589,6 @@ jobs:
           path: .
       - name: Unzip binaries
         run: tar -xvf ./bins.tar.gz
-      - if: matrix.mode.kafka
-        name: Setup kafka server
-        working-directory: tests-integration/fixtures/kafka
-        run: docker compose -f docker-compose-standalone.yml up -d --wait
       - name: Run sqlness
         run: RUST_BACKTRACE=1 ./bins/sqlness-runner ${{ matrix.mode.opts }} -c ./tests/cases --bins-dir ./bins --preserve-state
       - name: Upload sqlness logs
@@ -587,17 +608,16 @@ jobs:
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
         with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
           components: rustfmt
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
         with:
           # Shares across multiple jobs
           shared-key: "check-rust-fmt"
-      - name: Run cargo fmt
-        run: cargo fmt --all -- --check
+      - name: Check format
+        run: make fmt-check
 
   clippy:
     name: Clippy
@@ -608,9 +628,8 @@ jobs:
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
         with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
           components: clippy
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
@@ -634,9 +653,8 @@ jobs:
         with:
           version: "14.0"
       - name: Install toolchain
-        uses: dtolnay/rust-toolchain@master
+        uses: actions-rust-lang/setup-rust-toolchain@v1
         with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
           components: llvm-tools-preview
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
@@ -666,6 +684,9 @@ jobs:
       - name: Setup minio
         working-directory: tests-integration/fixtures/minio
         run: docker compose -f docker-compose-standalone.yml up -d --wait
+      - name: Setup postgres server
+        working-directory: tests-integration/fixtures/postgres
+        run: docker compose -f docker-compose-standalone.yml up -d --wait
       - name: Run nextest cases
         run: cargo llvm-cov nextest --workspace --lcov --output-path lcov.info -F pyo3_backend -F dashboard
         env:
@@ -682,7 +703,9 @@ jobs:
           GT_MINIO_REGION: us-west-2
           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_KAFKA_ENDPOINTS: 127.0.0.1:9092
+          GT_KAFKA_SASL_ENDPOINTS: 127.0.0.1:9093
           UNITTEST_LOG_DIR: "__unittest_logs"
       - name: Codecov upload
         uses: codecov/codecov-action@v4

.github/workflows/nightly-build.yml

@@ -154,6 +154,8 @@ jobs:
           cargo-profile: ${{ env.CARGO_PROFILE }}
           version: ${{ needs.allocate-runners.outputs.version }}
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
+          image-registry: ${{ vars.ECR_IMAGE_REGISTRY }}
+          image-namespace: ${{ vars.ECR_IMAGE_NAMESPACE }}
 
   build-linux-arm64-artifacts:
     name: Build linux-arm64 artifacts
@@ -173,6 +175,8 @@ jobs:
           cargo-profile: ${{ env.CARGO_PROFILE }}
           version: ${{ needs.allocate-runners.outputs.version }}
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
+          image-registry: ${{ vars.ECR_IMAGE_REGISTRY }}
+          image-namespace: ${{ vars.ECR_IMAGE_NAMESPACE }}
 
   release-images-to-dockerhub:
     name: Build and push images to DockerHub

.github/workflows/nightly-ci.yml

@@ -9,9 +9,6 @@ concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
   cancel-in-progress: true
 
-env:
-  RUST_TOOLCHAIN: nightly-2024-04-20
-
 permissions:
   issues: write
 
@@ -33,6 +30,13 @@ jobs:
           aws-region: ${{ vars.AWS_CI_TEST_BUCKET_REGION }}
           aws-access-key-id: ${{ secrets.AWS_CI_TEST_ACCESS_KEY_ID }}
           aws-secret-access-key: ${{ secrets.AWS_CI_TEST_SECRET_ACCESS_KEY }}
+      - name: Upload sqlness logs
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: sqlness-logs-kind
+          path: /tmp/kind/
+          retention-days: 3
 
   sqlness-windows:
     name: Sqlness tests on Windows
@@ -45,9 +49,7 @@ jobs:
       - uses: arduino/setup-protoc@v3
         with:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: dtolnay/rust-toolchain@master
-        with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2
       - name: Run sqlness
@@ -55,11 +57,11 @@ jobs:
         env:
           SQLNESS_OPTS: "--preserve-state"
       - name: Upload sqlness logs
-        if: always()
+        if: failure()
         uses: actions/upload-artifact@v4
         with:
           name: sqlness-logs
-          path: C:\tmp\greptime-*.log
+          path: C:\Users\RUNNER~1\AppData\Local\Temp\sqlness*
           retention-days: 3
 
   test-on-windows:
@@ -78,9 +80,8 @@ jobs:
         with:
           version: "14.0"
       - name: Install Rust toolchain
-        uses: dtolnay/rust-toolchain@master
+        uses: actions-rust-lang/setup-rust-toolchain@v1
         with:
-          toolchain: ${{ env.RUST_TOOLCHAIN }}
           components: llvm-tools-preview
       - name: Rust Cache
         uses: Swatinem/rust-cache@v2

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

@@ -1,12 +1,14 @@
 name: Release dev-builder images
 
 on:
+  push:
+    branches:
+      - main
+    paths:
+      - rust-toolchain.toml
+      - 'docker/dev-builder/**'
   workflow_dispatch: # Allows you to run this workflow manually.
     inputs:
-      version:
-        description: Version of the dev-builder
-        required: false
-        default: latest
       release_dev_builder_ubuntu_image:
         type: boolean
         description: Release dev-builder-ubuntu image
@@ -28,22 +30,103 @@ jobs:
     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-20.04-16-cores
+    outputs:
+      version: ${{ steps.set-version.outputs.version }}
     steps:
       - name: Checkout
         uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
+      - name: Configure build image version
+        id: set-version
+        shell: bash
+        run: |
+          commitShortSHA=`echo ${{ github.sha }} | cut -c1-8`
+          buildTime=`date +%Y%m%d%H%M%S`
+          BUILD_VERSION="$commitShortSHA-$buildTime"
+          RUST_TOOLCHAIN_VERSION=$(cat rust-toolchain.toml | grep -Eo '[0-9]{4}-[0-9]{2}-[0-9]{2}')
+          IMAGE_VERSION="${RUST_TOOLCHAIN_VERSION}-${BUILD_VERSION}"
+          echo "VERSION=${IMAGE_VERSION}" >> $GITHUB_ENV
+          echo "version=$IMAGE_VERSION" >> $GITHUB_OUTPUT
+
       - name: Build and push dev builder images
         uses: ./.github/actions/build-dev-builder-images
         with:
-          version: ${{ inputs.version }}
+          version: ${{ env.VERSION }}
           dockerhub-image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
           dockerhub-image-registry-token: ${{ secrets.DOCKERHUB_TOKEN }}
           build-dev-builder-ubuntu: ${{ inputs.release_dev_builder_ubuntu_image }}
           build-dev-builder-centos: ${{ inputs.release_dev_builder_centos_image }}
           build-dev-builder-android: ${{ inputs.release_dev_builder_android_image }}
 
+  release-dev-builder-images-ecr:
+    name: Release dev builder images to AWS ECR
+    runs-on: ubuntu-20.04
+    needs: [
+      release-dev-builder-images
+    ]
+    steps:
+      - name: Configure AWS credentials
+        uses: aws-actions/configure-aws-credentials@v4
+        with:
+          aws-access-key-id: ${{ secrets.AWS_ECR_ACCESS_KEY_ID }}
+          aws-secret-access-key: ${{ secrets.AWS_ECR_SECRET_ACCESS_KEY }}
+          aws-region: ${{ vars.ECR_REGION }}
+
+      - name: Login to Amazon ECR
+        id: login-ecr-public
+        uses: aws-actions/amazon-ecr-login@v2
+        env:
+          AWS_REGION: ${{ vars.ECR_REGION }}
+        with:
+          registry-type: public
+
+      - name: Push dev-builder-ubuntu image
+        shell: bash
+        if: ${{ inputs.release_dev_builder_ubuntu_image }}
+        run: |
+          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
+            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
+            quay.io/skopeo/stable:latest \
+            copy -a docker://docker.io/${{ 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/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:latest \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-ubuntu:latest
+      - name: Push dev-builder-centos image
+        shell: bash
+        if: ${{ inputs.release_dev_builder_centos_image }}
+        run: |
+          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
+            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
+            quay.io/skopeo/stable:latest \
+            copy -a docker://docker.io/${{ 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/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:latest \
+            docker://${{ vars.ECR_IMAGE_REGISTRY }}/${{ vars.ECR_IMAGE_NAMESPACE }}/dev-builder-centos:latest
+      - name: Push dev-builder-android image
+        shell: bash
+        if: ${{ inputs.release_dev_builder_android_image }}
+        run: |
+          docker run -v "${DOCKER_CONFIG:-$HOME/.docker}:/root/.docker:ro" \
+            -e "REGISTRY_AUTH_FILE=/root/.docker/config.json" \
+            quay.io/skopeo/stable:latest \
+            copy -a docker://docker.io/${{ 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/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:latest \
+            docker://${{ 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-20.04
@@ -51,35 +134,39 @@ jobs:
       release-dev-builder-images
     ]
     steps:
+      - name: Login to AliCloud Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ vars.ACR_IMAGE_REGISTRY }}
+          username: ${{ secrets.ALICLOUD_USERNAME }}
+          password: ${{ secrets.ALICLOUD_PASSWORD }}
+
       - name: Push dev-builder-ubuntu image
         shell: bash
         if: ${{ inputs.release_dev_builder_ubuntu_image }}
-        env:
-          DST_REGISTRY_USERNAME: ${{ secrets.ALICLOUD_USERNAME }}
-          DST_REGISTRY_PASSWORD: ${{ secrets.ALICLOUD_PASSWORD }}
         run: |
-          docker run quay.io/skopeo/stable:latest copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ inputs.version }} \
-            --dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
-            docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ inputs.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/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ 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:
-          DST_REGISTRY_USERNAME: ${{ secrets.ALICLOUD_USERNAME }}
-          DST_REGISTRY_PASSWORD: ${{ secrets.ALICLOUD_PASSWORD }}
         run: |
-          docker run quay.io/skopeo/stable:latest copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ inputs.version }} \
-            --dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
-            docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ inputs.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/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ 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:
-          DST_REGISTRY_USERNAME: ${{ secrets.ALICLOUD_USERNAME }}
-          DST_REGISTRY_PASSWORD: ${{ secrets.ALICLOUD_PASSWORD }}
         run: |
-          docker run quay.io/skopeo/stable:latest copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ inputs.version }} \
-            --dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
-            docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ inputs.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/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ 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 }}

.github/workflows/release.yml

@@ -33,6 +33,7 @@ on:
         description: The runner uses to build linux-arm64 artifacts
         default: ec2-c6g.4xlarge-arm64
         options:
+          - ubuntu-2204-32-cores-arm
           - ec2-c6g.xlarge-arm64 # 4C8G
           - ec2-c6g.2xlarge-arm64 # 8C16G
           - ec2-c6g.4xlarge-arm64 # 16C32G
@@ -82,7 +83,6 @@ on:
 # Use env variables to control all the release process.
 env:
   # The arguments of building greptime.
-  RUST_TOOLCHAIN: nightly-2024-04-20
   CARGO_PROFILE: nightly
 
   # Controls whether to run tests, include unit-test, integration-test and sqlness.
@@ -123,6 +123,11 @@ jobs:
         with:
           fetch-depth: 0
 
+      - name: Check Rust toolchain version
+        shell: bash
+        run: |
+          ./scripts/check-builder-rust-version.sh
+
       # The create-version will create a global variable named 'version' in the global workflows.
       # - If it's a tag push release, the version is the tag name(${{ github.ref_name }});
       # - If it's a scheduled release, the version is '${{ env.NEXT_RELEASE_VERSION }}-nightly-$buildTime', like v0.2.0-nigthly-20230313;
@@ -183,6 +188,8 @@ jobs:
           cargo-profile: ${{ env.CARGO_PROFILE }}
           version: ${{ needs.allocate-runners.outputs.version }}
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
+          image-registry: ${{ vars.ECR_IMAGE_REGISTRY }}
+          image-namespace: ${{ vars.ECR_IMAGE_NAMESPACE }}
 
   build-linux-arm64-artifacts:
     name: Build linux-arm64 artifacts
@@ -202,6 +209,8 @@ jobs:
           cargo-profile: ${{ env.CARGO_PROFILE }}
           version: ${{ needs.allocate-runners.outputs.version }}
           disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
+          image-registry: ${{ vars.ECR_IMAGE_REGISTRY }}
+          image-namespace: ${{ vars.ECR_IMAGE_NAMESPACE }}
 
   build-macos-artifacts:
     name: Build macOS artifacts
@@ -240,11 +249,11 @@ jobs:
       - uses: ./.github/actions/build-macos-artifacts
         with:
           arch: ${{ matrix.arch }}
-          rust-toolchain: ${{ env.RUST_TOOLCHAIN }}
           cargo-profile: ${{ env.CARGO_PROFILE }}
           features: ${{ matrix.features }}
           version: ${{ needs.allocate-runners.outputs.version }}
-          disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
+          # We decide to disable the integration tests on macOS because it's unnecessary and time-consuming.
+          disable-run-tests: true
           artifacts-dir: ${{ matrix.artifacts-dir-prefix }}-${{ needs.allocate-runners.outputs.version }}
 
       - name: Set build macos result
@@ -283,7 +292,6 @@ jobs:
       - uses: ./.github/actions/build-windows-artifacts
         with:
           arch: ${{ matrix.arch }}
-          rust-toolchain: ${{ env.RUST_TOOLCHAIN }}
           cargo-profile: ${{ env.CARGO_PROFILE }}
           features: ${{ matrix.features }}
           version: ${{ needs.allocate-runners.outputs.version }}

Cargo.toml

@@ -64,7 +64,7 @@ members = [
 resolver = "2"
 
 [workspace.package]
-version = "0.9.1"
+version = "0.9.3"
 edition = "2021"
 license = "Apache-2.0"
 
@@ -77,6 +77,7 @@ clippy.readonly_write_lock = "allow"
 rust.unknown_lints = "deny"
 # Remove this after https://github.com/PyO3/pyo3/issues/4094
 rust.non_local_definitions = "allow"
+rust.unexpected_cfgs = { level = "warn", check-cfg = ['cfg(tokio_unstable)'] }
 
 [workspace.dependencies]
 # We turn off default-features for some dependencies here so the workspaces which inherit them can
@@ -104,15 +105,15 @@ clap = { version = "4.4", features = ["derive"] }
 config = "0.13.0"
 crossbeam-utils = "0.8"
 dashmap = "5.4"
-datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-functions = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-physical-plan = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
-datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "d7bda5c9b762426e81f144296deadc87e5f4a0b8" }
+datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-functions = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-physical-plan = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
+datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "7823ef2f63663907edab46af0d51359900f608d6" }
 derive_builder = "0.12"
 dotenv = "0.15"
 etcd-client = { version = "0.13" }
@@ -124,7 +125,7 @@ humantime = "2.1"
 humantime-serde = "1.1"
 itertools = "0.10"
 lazy_static = "1.4"
-meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "049171eb16cb4249d8099751a0c46750d1fe88e7" }
+meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "80eb97c24c88af4dd9a86f8bbaf50e741d4eb8cd" }
 mockall = "0.11.4"
 moka = "0.12"
 notify = "6.1"
@@ -151,14 +152,18 @@ reqwest = { version = "0.12", default-features = false, features = [
     "stream",
     "multipart",
 ] }
-rskafka = "0.5"
+rskafka = { git = "https://github.com/influxdata/rskafka.git", rev = "75535b5ad9bae4a5dbb582c82e44dfd81ec10105", features = [
+    "transport-tls",
+] }
 rstest = "0.21"
 rstest_reuse = "0.7"
 rust_decimal = "1.33"
+rustc-hash = "2.0"
 schemars = "0.8"
 serde = { version = "1.0", features = ["derive"] }
 serde_json = { version = "1.0", features = ["float_roundtrip"] }
 serde_with = "3"
+shadow-rs = "0.31"
 smallvec = { version = "1", features = ["serde"] }
 snafu = "0.8"
 sysinfo = "0.30"
@@ -169,6 +174,7 @@ sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "5
 strum = { version = "0.25", features = ["derive"] }
 tempfile = "3"
 tokio = { version = "1.36", features = ["full"] }
+tokio-postgres = "0.7"
 tokio-stream = { version = "0.1" }
 tokio-util = { version = "0.7", features = ["io-util", "compat"] }
 toml = "0.8.8"
@@ -238,14 +244,14 @@ table = { path = "src/table" }
 
 [workspace.dependencies.meter-macros]
 git = "https://github.com/GreptimeTeam/greptime-meter.git"
-rev = "049171eb16cb4249d8099751a0c46750d1fe88e7"
+rev = "80eb97c24c88af4dd9a86f8bbaf50e741d4eb8cd"
 
 [profile.release]
 debug = 1
 
 [profile.nightly]
 inherits = "release"
-strip = true
+strip = "debuginfo"
 lto = "thin"
 debug = false
 incremental = false

Makefile

@@ -8,6 +8,7 @@ CARGO_BUILD_OPTS := --locked
 IMAGE_REGISTRY ?= docker.io
 IMAGE_NAMESPACE ?= greptime
 IMAGE_TAG ?= latest
+DEV_BUILDER_IMAGE_TAG ?= 2024-06-06-b4b105ad-20240827021230
 BUILDX_MULTI_PLATFORM_BUILD ?= false
 BUILDX_BUILDER_NAME ?= gtbuilder
 BASE_IMAGE ?= ubuntu
@@ -77,7 +78,7 @@ build: ## Build debug version greptime.
 build-by-dev-builder: ## Build greptime by dev-builder.
 	docker run --network=host \
 	-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry \
-	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:latest \
+	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:${DEV_BUILDER_IMAGE_TAG} \
 	make build \
 	CARGO_EXTENSION="${CARGO_EXTENSION}" \
 	CARGO_PROFILE=${CARGO_PROFILE} \
@@ -91,7 +92,7 @@ build-by-dev-builder: ## Build greptime by dev-builder.
 build-android-bin: ## Build greptime binary for android.
 	docker run --network=host \
 	-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry \
-	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-android:latest \
+	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-android:${DEV_BUILDER_IMAGE_TAG} \
 	make build \
 	CARGO_EXTENSION="ndk --platform 23 -t aarch64-linux-android" \
 	CARGO_PROFILE=release \
@@ -105,8 +106,8 @@ build-android-bin: ## Build greptime binary for android.
 strip-android-bin: build-android-bin ## Strip greptime binary for android.
 	docker run --network=host \
 	-v ${PWD}:/greptimedb \
-	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-android:latest \
-	bash -c '$${NDK_ROOT}/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-strip /greptimedb/target/aarch64-linux-android/release/greptime'
+	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-android:${DEV_BUILDER_IMAGE_TAG} \
+	bash -c '$${NDK_ROOT}/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-strip --strip-debug /greptimedb/target/aarch64-linux-android/release/greptime'
 
 .PHONY: clean
 clean: ## Clean the project.
@@ -145,7 +146,7 @@ dev-builder: multi-platform-buildx ## Build dev-builder image.
 	docker buildx build --builder ${BUILDX_BUILDER_NAME} \
 	--build-arg="RUST_TOOLCHAIN=${RUST_TOOLCHAIN}" \
 	-f docker/dev-builder/${BASE_IMAGE}/Dockerfile \
-	-t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:${IMAGE_TAG} ${BUILDX_MULTI_PLATFORM_BUILD_OPTS} .
+	-t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:${DEV_BUILDER_IMAGE_TAG} ${BUILDX_MULTI_PLATFORM_BUILD_OPTS} .
 
 .PHONY: multi-platform-buildx
 multi-platform-buildx: ## Create buildx multi-platform builder.
@@ -190,6 +191,7 @@ fix-clippy: ## Fix clippy violations.
 .PHONY: fmt-check
 fmt-check: ## Check code format.
 	cargo fmt --all -- --check
+	python3 scripts/check-snafu.py
 
 .PHONY: start-etcd
 start-etcd: ## Start single node etcd for testing purpose.
@@ -203,7 +205,7 @@ stop-etcd: ## Stop single node etcd for testing purpose.
 run-it-in-container: start-etcd ## Run integration tests in dev-builder.
 	docker run --network=host \
 	-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry -v /tmp:/tmp \
-	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:latest \
+	-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:${DEV_BUILDER_IMAGE_TAG} \
 	make test sqlness-test BUILD_JOBS=${BUILD_JOBS}
 
 .PHONY: start-cluster

README.md

@@ -150,7 +150,7 @@ Our official Grafana dashboard is available at [grafana](grafana/README.md) dire
 
 ## Project Status
 
-The current version has not yet reached the standards for General Availability. 
+The current version has not yet reached the standards for General Availability.
 According to our Greptime 2024 Roadmap, we aim to achieve a production-level version with the release of v1.0 by the end of 2024. [Join Us](https://github.com/GreptimeTeam/greptimedb/issues/3412)
 
 We welcome you to test and use GreptimeDB. Some users have already adopted it in their production environments. If you're interested in trying it out, please use the latest stable release available.
@@ -172,6 +172,13 @@ In addition, you may:
 - Connect us with [Linkedin](https://www.linkedin.com/company/greptime/)
 - Follow us on [Twitter](https://twitter.com/greptime)
 
+## Commerial Support
+
+If you are running GreptimeDB OSS in your organization, we offer additional
+enterprise addons, installation service, training and consulting. [Contact
+us](https://greptime.com/contactus) and we will reach out to you with more
+detail of our commerial license.
+
 ## License
 
 GreptimeDB uses the [Apache License 2.0](https://apache.org/licenses/LICENSE-2.0.txt) to strike a balance between

config/config.md

@@ -15,6 +15,8 @@
 | `mode` | String | `standalone` | The running mode of the datanode. It can be `standalone` or `distributed`. |
 | `enable_telemetry` | Bool | `true` | Enable telemetry to collect anonymous usage data. |
 | `default_timezone` | String | `None` | The default timezone of the server. |
+| `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. |
 | `runtime` | -- | -- | The runtime options. |
 | `runtime.global_rt_size` | Integer | `8` | The number of threads to execute the runtime for global read operations. |
 | `runtime.compact_rt_size` | Integer | `4` | The number of threads to execute the runtime for global write operations. |
@@ -67,6 +69,12 @@
 | `wal.prefill_log_files` | Bool | `false` | Whether to pre-create log files on start up.<br/>**It's only used when the provider is `raft_engine`**. |
 | `wal.sync_period` | String | `10s` | Duration for fsyncing log files.<br/>**It's only used when the provider is `raft_engine`**. |
 | `wal.broker_endpoints` | Array | -- | The Kafka broker endpoints.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.auto_create_topics` | Bool | `true` | Automatically create topics for WAL.<br/>Set to `true` to automatically create topics for WAL.<br/>Otherwise, use topics named `topic_name_prefix_[0..num_topics)` |
+| `wal.num_topics` | Integer | `64` | Number of topics.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.selector_type` | String | `round_robin` | Topic selector type.<br/>Available selector types:<br/>- `round_robin` (default)<br/>**It's only used when the provider is `kafka`**. |
+| `wal.topic_name_prefix` | String | `greptimedb_wal_topic` | A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.<br/>i.g., greptimedb_wal_topic_0, greptimedb_wal_topic_1.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.replication_factor` | Integer | `1` | Expected number of replicas of each partition.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.create_topic_timeout` | String | `30s` | Above which a topic creation operation will be cancelled.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.max_batch_bytes` | String | `1MB` | The max size of a single producer batch.<br/>Warning: Kafka has a default limit of 1MB per message in a topic.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.consumer_wait_timeout` | String | `100ms` | The consumer wait timeout.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.backoff_init` | String | `500ms` | The initial backoff delay.<br/>**It's only used when the provider is `kafka`**. |
@@ -93,6 +101,7 @@
 | `storage.account_key` | String | `None` | The account key of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
 | `storage.scope` | String | `None` | The scope of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
 | `storage.credential_path` | String | `None` | The credential path of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
+| `storage.credential` | String | `None` | The credential of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
 | `storage.container` | String | `None` | The container of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
 | `storage.sas_token` | String | `None` | The sas token of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
 | `storage.endpoint` | String | `None` | The endpoint of the S3 service.<br/>**It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**. |
@@ -110,15 +119,17 @@
 | `region_engine.mito.global_write_buffer_reject_size` | String | `2GB` | Global write buffer size threshold to reject write requests. If not set, it's default to 2 times of `global_write_buffer_size` |
 | `region_engine.mito.sst_meta_cache_size` | String | `128MB` | Cache size for SST metadata. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/32 of OS memory with a max limitation of 128MB. |
 | `region_engine.mito.vector_cache_size` | String | `512MB` | Cache size for vectors and arrow arrays. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
-| `region_engine.mito.page_cache_size` | String | `512MB` | Cache size for pages of SST row groups. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
+| `region_engine.mito.page_cache_size` | String | `512MB` | Cache size for pages of SST row groups. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/8 of OS memory. |
+| `region_engine.mito.selector_result_cache_size` | String | `512MB` | Cache size for time series selector (e.g. `last_value()`). Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
 | `region_engine.mito.enable_experimental_write_cache` | Bool | `false` | Whether to enable the experimental write cache. |
 | `region_engine.mito.experimental_write_cache_path` | String | `""` | File system path for write cache, defaults to `{data_home}/write_cache`. |
 | `region_engine.mito.experimental_write_cache_size` | String | `512MB` | Capacity for write cache. |
-| `region_engine.mito.experimental_write_cache_ttl` | String | `1h` | TTL for write cache. |
+| `region_engine.mito.experimental_write_cache_ttl` | String | `None` | TTL for write cache. |
 | `region_engine.mito.sst_write_buffer_size` | String | `8MB` | Buffer size for SST writing. |
 | `region_engine.mito.scan_parallelism` | Integer | `0` | Parallelism to scan a region (default: 1/4 of cpu cores).<br/>- `0`: using the default value (1/4 of cpu cores).<br/>- `1`: scan in current thread.<br/>- `n`: scan in parallelism n. |
 | `region_engine.mito.parallel_scan_channel_size` | Integer | `32` | Capacity of the channel to send data from parallel scan tasks to the main task. |
 | `region_engine.mito.allow_stale_entries` | Bool | `false` | Whether to allow stale WAL entries read during replay. |
+| `region_engine.mito.min_compaction_interval` | String | `0m` | Minimum time interval between two compactions.<br/>To align with the old behavior, the default value is 0 (no restrictions). |
 | `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. |
@@ -142,20 +153,21 @@
 | `region_engine.mito.memtable.fork_dictionary_bytes` | String | `1GiB` | Max dictionary bytes.<br/>Only available for `partition_tree` memtable. |
 | `region_engine.file` | -- | -- | Enable the file engine. |
 | `logging` | -- | -- | The logging options. |
-| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
 | `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
 | `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
-| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
 | `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
 | `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
 | `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
 | `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
 | `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
 | `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
-| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself<br/>You must create the database before enabling it. |
 | `export_metrics.self_import.db` | String | `None` | -- |
 | `export_metrics.remote_write` | -- | -- | -- |
-| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`. |
 | `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
 | `tracing` | -- | -- | The tracing options. Only effect when compiled with `tokio-console` feature. |
 | `tracing.tokio_console_addr` | String | `None` | The tokio console address. |
@@ -227,20 +239,21 @@
 | `datanode.client.connect_timeout` | String | `10s` | -- |
 | `datanode.client.tcp_nodelay` | Bool | `true` | -- |
 | `logging` | -- | -- | The logging options. |
-| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
 | `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
 | `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
-| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
 | `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
 | `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
 | `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
 | `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
 | `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
 | `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
-| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself<br/>You must create the database before enabling it. |
 | `export_metrics.self_import.db` | String | `None` | -- |
 | `export_metrics.remote_write` | -- | -- | -- |
-| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`. |
 | `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
 | `tracing` | -- | -- | The tracing options. Only effect when compiled with `tokio-console` feature. |
 | `tracing.tokio_console_addr` | String | `None` | The tokio console address. |
@@ -253,12 +266,13 @@
 | `data_home` | String | `/tmp/metasrv/` | The working home directory. |
 | `bind_addr` | String | `127.0.0.1:3002` | The bind address of metasrv. |
 | `server_addr` | String | `127.0.0.1:3002` | The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost. |
-| `store_addr` | String | `127.0.0.1:2379` | Etcd server address. |
-| `selector` | String | `lease_based` | Datanode selector type.<br/>- `lease_based` (default value).<br/>- `load_based`<br/>For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector". |
+| `store_addr` | String | `127.0.0.1:2379` | Store server address default to etcd store. |
+| `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_telemetry` | Bool | `true` | Whether to enable greptimedb telemetry. |
 | `store_key_prefix` | String | `""` | If it's not empty, the metasrv will store all data with this key prefix. |
 | `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). |
+| `backend` | String | `EtcdStore` | The datastore for meta server. |
 | `runtime` | -- | -- | The runtime options. |
 | `runtime.global_rt_size` | Integer | `8` | The number of threads to execute the runtime for global read operations. |
 | `runtime.compact_rt_size` | Integer | `4` | The number of threads to execute the runtime for global write operations. |
@@ -279,9 +293,10 @@
 | `wal` | -- | -- | -- |
 | `wal.provider` | String | `raft_engine` | -- |
 | `wal.broker_endpoints` | Array | -- | The broker endpoints of the Kafka cluster. |
-| `wal.num_topics` | Integer | `64` | Number of topics to be created upon start. |
+| `wal.auto_create_topics` | Bool | `true` | Automatically create topics for WAL.<br/>Set to `true` to automatically create topics for WAL.<br/>Otherwise, use topics named `topic_name_prefix_[0..num_topics)` |
+| `wal.num_topics` | Integer | `64` | Number of topics. |
 | `wal.selector_type` | String | `round_robin` | Topic selector type.<br/>Available selector types:<br/>- `round_robin` (default) |
-| `wal.topic_name_prefix` | String | `greptimedb_wal_topic` | A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`. |
+| `wal.topic_name_prefix` | String | `greptimedb_wal_topic` | A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.<br/>i.g., greptimedb_wal_topic_0, greptimedb_wal_topic_1. |
 | `wal.replication_factor` | Integer | `1` | Expected number of replicas of each partition. |
 | `wal.create_topic_timeout` | String | `30s` | Above which a topic creation operation will be cancelled. |
 | `wal.backoff_init` | String | `500ms` | The initial backoff for kafka clients. |
@@ -289,20 +304,21 @@
 | `wal.backoff_base` | Integer | `2` | Exponential backoff rate, i.e. next backoff = base * current backoff. |
 | `wal.backoff_deadline` | String | `5mins` | Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate. |
 | `logging` | -- | -- | The logging options. |
-| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
 | `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
 | `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
-| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
 | `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
 | `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
 | `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
 | `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
 | `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
 | `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
-| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself<br/>You must create the database before enabling it. |
 | `export_metrics.self_import.db` | String | `None` | -- |
 | `export_metrics.remote_write` | -- | -- | -- |
-| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`. |
 | `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
 | `tracing` | -- | -- | The tracing options. Only effect when compiled with `tokio-console` feature. |
 | `tracing.tokio_console_addr` | String | `None` | The tokio console address. |
@@ -323,6 +339,10 @@
 | `rpc_runtime_size` | Integer | `None` | Deprecated, use `grpc.runtime_size` instead. |
 | `rpc_max_recv_message_size` | String | `None` | Deprecated, use `grpc.rpc_max_recv_message_size` instead. |
 | `rpc_max_send_message_size` | String | `None` | Deprecated, use `grpc.rpc_max_send_message_size` instead. |
+| `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.addr` | String | `127.0.0.1:3001` | 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 |
@@ -368,6 +388,8 @@
 | `wal.backoff_max` | String | `10s` | The maximum backoff delay.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.backoff_base` | Integer | `2` | The exponential backoff rate, i.e. next backoff = base * current backoff.<br/>**It's only used when the provider is `kafka`**. |
 | `wal.backoff_deadline` | String | `5mins` | The deadline of retries.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.create_index` | Bool | `true` | Whether to enable WAL index creation.<br/>**It's only used when the provider is `kafka`**. |
+| `wal.dump_index_interval` | String | `60s` | The interval for dumping WAL indexes.<br/>**It's only used when the provider is `kafka`**. |
 | `storage` | -- | -- | The data storage options. |
 | `storage.data_home` | String | `/tmp/greptimedb/` | The working home directory. |
 | `storage.type` | String | `File` | The storage type used to store the data.<br/>- `File`: the data is stored in the local file system.<br/>- `S3`: the data is stored in the S3 object storage.<br/>- `Gcs`: the data is stored in the Google Cloud Storage.<br/>- `Azblob`: the data is stored in the Azure Blob Storage.<br/>- `Oss`: the data is stored in the Aliyun OSS. |
@@ -382,6 +404,7 @@
 | `storage.account_key` | String | `None` | The account key of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
 | `storage.scope` | String | `None` | The scope of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
 | `storage.credential_path` | String | `None` | The credential path of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
+| `storage.credential` | String | `None` | The credential of the google cloud storage.<br/>**It's only used when the storage type is `Gcs`**. |
 | `storage.container` | String | `None` | The container of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
 | `storage.sas_token` | String | `None` | The sas token of the azure account.<br/>**It's only used when the storage type is `Azblob`**. |
 | `storage.endpoint` | String | `None` | The endpoint of the S3 service.<br/>**It's only used when the storage type is `S3`, `Oss`, `Gcs` and `Azblob`**. |
@@ -399,15 +422,17 @@
 | `region_engine.mito.global_write_buffer_reject_size` | String | `2GB` | Global write buffer size threshold to reject write requests. If not set, it's default to 2 times of `global_write_buffer_size` |
 | `region_engine.mito.sst_meta_cache_size` | String | `128MB` | Cache size for SST metadata. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/32 of OS memory with a max limitation of 128MB. |
 | `region_engine.mito.vector_cache_size` | String | `512MB` | Cache size for vectors and arrow arrays. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
-| `region_engine.mito.page_cache_size` | String | `512MB` | Cache size for pages of SST row groups. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
+| `region_engine.mito.page_cache_size` | String | `512MB` | Cache size for pages of SST row groups. Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/8 of OS memory. |
+| `region_engine.mito.selector_result_cache_size` | String | `512MB` | Cache size for time series selector (e.g. `last_value()`). Setting it to 0 to disable the cache.<br/>If not set, it's default to 1/16 of OS memory with a max limitation of 512MB. |
 | `region_engine.mito.enable_experimental_write_cache` | Bool | `false` | Whether to enable the experimental write cache. |
 | `region_engine.mito.experimental_write_cache_path` | String | `""` | File system path for write cache, defaults to `{data_home}/write_cache`. |
 | `region_engine.mito.experimental_write_cache_size` | String | `512MB` | Capacity for write cache. |
-| `region_engine.mito.experimental_write_cache_ttl` | String | `1h` | TTL for write cache. |
+| `region_engine.mito.experimental_write_cache_ttl` | String | `None` | TTL for write cache. |
 | `region_engine.mito.sst_write_buffer_size` | String | `8MB` | Buffer size for SST writing. |
 | `region_engine.mito.scan_parallelism` | Integer | `0` | Parallelism to scan a region (default: 1/4 of cpu cores).<br/>- `0`: using the default value (1/4 of cpu cores).<br/>- `1`: scan in current thread.<br/>- `n`: scan in parallelism n. |
 | `region_engine.mito.parallel_scan_channel_size` | Integer | `32` | Capacity of the channel to send data from parallel scan tasks to the main task. |
 | `region_engine.mito.allow_stale_entries` | Bool | `false` | Whether to allow stale WAL entries read during replay. |
+| `region_engine.mito.min_compaction_interval` | String | `0m` | Minimum time interval between two compactions.<br/>To align with the old behavior, the default value is 0 (no restrictions). |
 | `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. |
@@ -429,20 +454,21 @@
 | `region_engine.mito.memtable.fork_dictionary_bytes` | String | `1GiB` | Max dictionary bytes.<br/>Only available for `partition_tree` memtable. |
 | `region_engine.file` | -- | -- | Enable the file engine. |
 | `logging` | -- | -- | The logging options. |
-| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
 | `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
 | `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
-| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
 | `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
 | `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
 | `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
 | `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. send to `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
 | `export_metrics.enable` | Bool | `false` | whether enable export metrics. |
 | `export_metrics.write_interval` | String | `30s` | The interval of export metrics. |
-| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself |
+| `export_metrics.self_import` | -- | -- | For `standalone` mode, `self_import` is recommend to collect metrics generated by itself<br/>You must create the database before enabling it. |
 | `export_metrics.self_import.db` | String | `None` | -- |
 | `export_metrics.remote_write` | -- | -- | -- |
-| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`. |
+| `export_metrics.remote_write.url` | String | `""` | The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`. |
 | `export_metrics.remote_write.headers` | InlineTable | -- | HTTP headers of Prometheus remote-write carry. |
 | `tracing` | -- | -- | The tracing options. Only effect when compiled with `tokio-console` feature. |
 | `tracing.tokio_console_addr` | String | `None` | The tokio console address. |
@@ -474,11 +500,12 @@
 | `heartbeat.interval` | String | `3s` | Interval for sending heartbeat messages to the metasrv. |
 | `heartbeat.retry_interval` | String | `3s` | Interval for retrying to send heartbeat messages to the metasrv. |
 | `logging` | -- | -- | The logging options. |
-| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. |
+| `logging.dir` | String | `/tmp/greptimedb/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
 | `logging.level` | String | `None` | The log level. Can be `info`/`debug`/`warn`/`error`. |
 | `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
-| `logging.otlp_endpoint` | String | `None` | The OTLP tracing endpoint. |
+| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
 | `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
+| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
 | `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
 | `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
 | `tracing` | -- | -- | The tracing options. Only effect when compiled with `tokio-console` feature. |

config/datanode.example.toml

@@ -39,6 +39,18 @@ rpc_max_recv_message_size = "512MB"
 ## +toml2docs:none-default
 rpc_max_send_message_size = "512MB"
 
+
+## The HTTP server options.
+[http]
+## The address to bind the HTTP server.
+addr = "127.0.0.1:4000"
+## HTTP request timeout. Set to 0 to disable timeout.
+timeout = "30s"
+## HTTP request body limit.
+## The following units are supported: `B`, `KB`, `KiB`, `MB`, `MiB`, `GB`, `GiB`, `TB`, `TiB`, `PB`, `PiB`.
+## Set to 0 to disable limit.
+body_limit = "64MB"
+
 ## The gRPC server options.
 [grpc]
 ## The address to bind the gRPC server.
@@ -187,6 +199,32 @@ backoff_base = 2
 ## **It's only used when the provider is `kafka`**.
 backoff_deadline = "5mins"
 
+## Whether to enable WAL index creation.
+## **It's only used when the provider is `kafka`**.
+create_index = true
+
+## The interval for dumping WAL indexes.
+## **It's only used when the provider is `kafka`**.
+dump_index_interval = "60s"
+
+# The Kafka SASL configuration.
+# **It's only used when the provider is `kafka`**.
+# Available SASL mechanisms:
+# - `PLAIN`
+# - `SCRAM-SHA-256`
+# - `SCRAM-SHA-512`
+# [wal.sasl]
+# type = "SCRAM-SHA-512"
+# username = "user_kafka"
+# password = "secret"
+
+# The Kafka TLS configuration.
+# **It's only used when the provider is `kafka`**.
+# [wal.tls]
+# server_ca_cert_path = "/path/to/server_cert"
+# client_cert_path = "/path/to/client_cert"
+# client_key_path = "/path/to/key"
+
 # Example of using S3 as the storage.
 # [storage]
 # type = "S3"
@@ -223,6 +261,7 @@ backoff_deadline = "5mins"
 # root = "data"
 # scope = "test"
 # credential_path = "123456"
+# credential = "base64-credential"
 # endpoint = "https://storage.googleapis.com"
 
 ## The data storage options.
@@ -294,6 +333,11 @@ scope = "test"
 ## +toml2docs:none-default
 credential_path = "test"
 
+## The credential of the google cloud storage.
+## **It's only used when the storage type is `Gcs`**.
+## +toml2docs:none-default
+credential = "base64-credential"
+
 ## The container of the azure account.
 ## **It's only used when the storage type is `Azblob`**.
 ## +toml2docs:none-default
@@ -316,9 +360,23 @@ region = "us-west-2"
 
 # Custom storage options
 # [[storage.providers]]
+# name = "S3"
 # type = "S3"
+# bucket = "greptimedb"
+# root = "data"
+# access_key_id = "test"
+# secret_access_key = "123456"
+# endpoint = "https://s3.amazonaws.com"
+# region = "us-west-2"
 # [[storage.providers]]
+# name = "Gcs"
 # type = "Gcs"
+# bucket = "greptimedb"
+# root = "data"
+# scope = "test"
+# credential_path = "123456"
+# credential = "base64-credential"
+# endpoint = "https://storage.googleapis.com"
 
 ## The region engine options. You can configure multiple region engines.
 [[region_engine]]
@@ -362,9 +420,13 @@ sst_meta_cache_size = "128MB"
 vector_cache_size = "512MB"
 
 ## Cache size for pages of SST row groups. Setting it to 0 to disable the cache.
-## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
+## If not set, it's default to 1/8 of OS memory.
 page_cache_size = "512MB"
 
+## Cache size for time series selector (e.g. `last_value()`). Setting it to 0 to disable the cache.
+## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
+selector_result_cache_size = "512MB"
+
 ## Whether to enable the experimental write cache.
 enable_experimental_write_cache = false
 
@@ -375,7 +437,8 @@ experimental_write_cache_path = ""
 experimental_write_cache_size = "512MB"
 
 ## TTL for write cache.
-experimental_write_cache_ttl = "1h"
+## +toml2docs:none-default
+experimental_write_cache_ttl = "8h"
 
 ## Buffer size for SST writing.
 sst_write_buffer_size = "8MB"
@@ -392,6 +455,10 @@ parallel_scan_channel_size = 32
 ## Whether to allow stale WAL entries read during replay.
 allow_stale_entries = false
 
+## Minimum time interval between two compactions.
+## To align with the old behavior, the default value is 0 (no restrictions).
+min_compaction_interval = "0m"
+
 ## The options for index in Mito engine.
 [region_engine.mito.index]
 
@@ -482,7 +549,7 @@ fork_dictionary_bytes = "1GiB"
 
 ## The logging options.
 [logging]
-## The directory to store the log files.
+## The directory to store the log files. If set to empty, logs will not be written to files.
 dir = "/tmp/greptimedb/logs"
 
 ## The log level. Can be `info`/`debug`/`warn`/`error`.
@@ -493,12 +560,14 @@ level = "info"
 enable_otlp_tracing = false
 
 ## The OTLP tracing endpoint.
-## +toml2docs:none-default
-otlp_endpoint = ""
+otlp_endpoint = "http://localhost:4317"
 
 ## Whether to append logs to stdout.
 append_stdout = true
 
+## The log format. Can be `text`/`json`.
+log_format = "text"
+
 ## The percentage of tracing will be sampled and exported.
 ## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
 ## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
@@ -516,12 +585,13 @@ enable = false
 write_interval = "30s"
 
 ## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+## You must create the database before enabling it.
 [export_metrics.self_import]
 ## +toml2docs:none-default
-db = "information_schema"
+db = "greptime_metrics"
 
 [export_metrics.remote_write]
-## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`.
 url = ""
 
 ## HTTP headers of Prometheus remote-write carry.

config/flownode.example.toml

@@ -59,7 +59,7 @@ retry_interval = "3s"
 
 ## The logging options.
 [logging]
-## The directory to store the log files.
+## The directory to store the log files. If set to empty, logs will not be written to files.
 dir = "/tmp/greptimedb/logs"
 
 ## The log level. Can be `info`/`debug`/`warn`/`error`.
@@ -70,12 +70,14 @@ level = "info"
 enable_otlp_tracing = false
 
 ## The OTLP tracing endpoint.
-## +toml2docs:none-default
-otlp_endpoint = ""
+otlp_endpoint = "http://localhost:4317"
 
 ## Whether to append logs to stdout.
 append_stdout = true
 
+## The log format. Can be `text`/`json`.
+log_format = "text"
+
 ## The percentage of tracing will be sampled and exported.
 ## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
 ## ratio > 1 are treated as 1. Fractions < 0 are treated as 0

config/frontend.example.toml

@@ -166,7 +166,7 @@ tcp_nodelay = true
 
 ## The logging options.
 [logging]
-## The directory to store the log files.
+## The directory to store the log files. If set to empty, logs will not be written to files.
 dir = "/tmp/greptimedb/logs"
 
 ## The log level. Can be `info`/`debug`/`warn`/`error`.
@@ -177,12 +177,14 @@ level = "info"
 enable_otlp_tracing = false
 
 ## The OTLP tracing endpoint.
-## +toml2docs:none-default
-otlp_endpoint = ""
+otlp_endpoint = "http://localhost:4317"
 
 ## Whether to append logs to stdout.
 append_stdout = true
 
+## The log format. Can be `text`/`json`.
+log_format = "text"
+
 ## The percentage of tracing will be sampled and exported.
 ## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
 ## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
@@ -200,12 +202,13 @@ enable = false
 write_interval = "30s"
 
 ## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+## You must create the database before enabling it.
 [export_metrics.self_import]
 ## +toml2docs:none-default
-db = "information_schema"
+db = "greptime_metrics"
 
 [export_metrics.remote_write]
-## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`.
 url = ""
 
 ## HTTP headers of Prometheus remote-write carry.

config/metasrv.example.toml

@@ -7,14 +7,15 @@ bind_addr = "127.0.0.1:3002"
 ## The communication server address for frontend and datanode to connect to metasrv,  "127.0.0.1:3002" by default for localhost.
 server_addr = "127.0.0.1:3002"
 
-## Etcd server address.
+## Store server address default to etcd store.
 store_addr = "127.0.0.1:2379"
 
 ## Datanode selector type.
-## - `lease_based` (default value).
+## - `round_robin` (default value)
+## - `lease_based`
 ## - `load_based`
 ## For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector".
-selector = "lease_based"
+selector = "round_robin"
 
 ## Store data in memory.
 use_memory_store = false
@@ -31,6 +32,9 @@ store_key_prefix = ""
 ## - Using shared storage (e.g., s3).
 enable_region_failover = false
 
+## The datastore for meta server.
+backend = "EtcdStore"
+
 ## The runtime options.
 [runtime]
 ## The number of threads to execute the runtime for global read operations.
@@ -95,7 +99,12 @@ provider = "raft_engine"
 ## The broker endpoints of the Kafka cluster.
 broker_endpoints = ["127.0.0.1:9092"]
 
-## Number of topics to be created upon start.
+## Automatically create topics for WAL.
+## Set to `true` to automatically create topics for WAL.
+## Otherwise, use topics named `topic_name_prefix_[0..num_topics)`
+auto_create_topics = true
+
+## Number of topics.
 num_topics = 64
 
 ## Topic selector type.
@@ -104,6 +113,7 @@ num_topics = 64
 selector_type = "round_robin"
 
 ## A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.
+## i.g., greptimedb_wal_topic_0, greptimedb_wal_topic_1.
 topic_name_prefix = "greptimedb_wal_topic"
 
 ## Expected number of replicas of each partition.
@@ -123,9 +133,27 @@ backoff_base = 2
 ## Stop reconnecting if the total wait time reaches the deadline. If this config is missing, the reconnecting won't terminate.
 backoff_deadline = "5mins"
 
+# The Kafka SASL configuration.
+# **It's only used when the provider is `kafka`**.
+# Available SASL mechanisms:
+# - `PLAIN`
+# - `SCRAM-SHA-256`
+# - `SCRAM-SHA-512`
+# [wal.sasl]
+# type = "SCRAM-SHA-512"
+# username = "user_kafka"
+# password = "secret"
+
+# The Kafka TLS configuration.
+# **It's only used when the provider is `kafka`**.
+# [wal.tls]
+# server_ca_cert_path = "/path/to/server_cert"
+# client_cert_path = "/path/to/client_cert"
+# client_key_path = "/path/to/key"
+
 ## The logging options.
 [logging]
-## The directory to store the log files.
+## The directory to store the log files. If set to empty, logs will not be written to files.
 dir = "/tmp/greptimedb/logs"
 
 ## The log level. Can be `info`/`debug`/`warn`/`error`.
@@ -136,12 +164,14 @@ level = "info"
 enable_otlp_tracing = false
 
 ## The OTLP tracing endpoint.
-## +toml2docs:none-default
-otlp_endpoint = ""
+otlp_endpoint = "http://localhost:4317"
 
 ## Whether to append logs to stdout.
 append_stdout = true
 
+## The log format. Can be `text`/`json`.
+log_format = "text"
+
 ## The percentage of tracing will be sampled and exported.
 ## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
 ## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
@@ -159,12 +189,13 @@ enable = false
 write_interval = "30s"
 
 ## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+## You must create the database before enabling it.
 [export_metrics.self_import]
 ## +toml2docs:none-default
-db = "information_schema"
+db = "greptime_metrics"
 
 [export_metrics.remote_write]
-## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`.
 url = ""
 
 ## HTTP headers of Prometheus remote-write carry.

config/standalone.example.toml

@@ -8,6 +8,13 @@ enable_telemetry = true
 ## +toml2docs:none-default
 default_timezone = "UTC"
 
+## Initialize all regions in the background during the startup.
+## By default, it provides services after all regions have been initialized.
+init_regions_in_background = false
+
+## Parallelism of initializing regions.
+init_regions_parallelism = 16
+
 ## The runtime options.
 [runtime]
 ## The number of threads to execute the runtime for global read operations.
@@ -171,6 +178,34 @@ sync_period = "10s"
 ## **It's only used when the provider is `kafka`**.
 broker_endpoints = ["127.0.0.1:9092"]
 
+## Automatically create topics for WAL.
+## Set to `true` to automatically create topics for WAL.
+## Otherwise, use topics named `topic_name_prefix_[0..num_topics)`
+auto_create_topics = true
+
+## Number of topics.
+## **It's only used when the provider is `kafka`**.
+num_topics = 64
+
+## Topic selector type.
+## Available selector types:
+## - `round_robin` (default)
+## **It's only used when the provider is `kafka`**.
+selector_type = "round_robin"
+
+## A Kafka topic is constructed by concatenating `topic_name_prefix` and `topic_id`.
+## i.g., greptimedb_wal_topic_0, greptimedb_wal_topic_1.
+## **It's only used when the provider is `kafka`**.
+topic_name_prefix = "greptimedb_wal_topic"
+
+## Expected number of replicas of each partition.
+## **It's only used when the provider is `kafka`**.
+replication_factor = 1
+
+## Above which a topic creation operation will be cancelled.
+## **It's only used when the provider is `kafka`**.
+create_topic_timeout = "30s"
+
 ## The max size of a single producer batch.
 ## Warning: Kafka has a default limit of 1MB per message in a topic.
 ## **It's only used when the provider is `kafka`**.
@@ -196,6 +231,24 @@ backoff_base = 2
 ## **It's only used when the provider is `kafka`**.
 backoff_deadline = "5mins"
 
+# The Kafka SASL configuration.
+# **It's only used when the provider is `kafka`**.
+# Available SASL mechanisms:
+# - `PLAIN`
+# - `SCRAM-SHA-256`
+# - `SCRAM-SHA-512`
+# [wal.sasl]
+# type = "SCRAM-SHA-512"
+# username = "user_kafka"
+# password = "secret"
+
+# The Kafka TLS configuration.
+# **It's only used when the provider is `kafka`**.
+# [wal.tls]
+# server_ca_cert_path = "/path/to/server_cert"
+# client_cert_path = "/path/to/client_cert"
+# client_key_path = "/path/to/key"
+
 ## Metadata storage options.
 [metadata_store]
 ## Kv file size in bytes.
@@ -246,6 +299,7 @@ retry_delay = "500ms"
 # root = "data"
 # scope = "test"
 # credential_path = "123456"
+# credential = "base64-credential"
 # endpoint = "https://storage.googleapis.com"
 
 ## The data storage options.
@@ -317,6 +371,11 @@ scope = "test"
 ## +toml2docs:none-default
 credential_path = "test"
 
+## The credential of the google cloud storage.
+## **It's only used when the storage type is `Gcs`**.
+## +toml2docs:none-default
+credential = "base64-credential"
+
 ## The container of the azure account.
 ## **It's only used when the storage type is `Azblob`**.
 ## +toml2docs:none-default
@@ -339,9 +398,23 @@ region = "us-west-2"
 
 # Custom storage options
 # [[storage.providers]]
+# name = "S3"
 # type = "S3"
+# bucket = "greptimedb"
+# root = "data"
+# access_key_id = "test"
+# secret_access_key = "123456"
+# endpoint = "https://s3.amazonaws.com"
+# region = "us-west-2"
 # [[storage.providers]]
+# name = "Gcs"
 # type = "Gcs"
+# bucket = "greptimedb"
+# root = "data"
+# scope = "test"
+# credential_path = "123456"
+# credential = "base64-credential"
+# endpoint = "https://storage.googleapis.com"
 
 ## The region engine options. You can configure multiple region engines.
 [[region_engine]]
@@ -385,9 +458,13 @@ sst_meta_cache_size = "128MB"
 vector_cache_size = "512MB"
 
 ## Cache size for pages of SST row groups. Setting it to 0 to disable the cache.
-## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
+## If not set, it's default to 1/8 of OS memory.
 page_cache_size = "512MB"
 
+## Cache size for time series selector (e.g. `last_value()`). Setting it to 0 to disable the cache.
+## If not set, it's default to 1/16 of OS memory with a max limitation of 512MB.
+selector_result_cache_size = "512MB"
+
 ## Whether to enable the experimental write cache.
 enable_experimental_write_cache = false
 
@@ -398,7 +475,8 @@ experimental_write_cache_path = ""
 experimental_write_cache_size = "512MB"
 
 ## TTL for write cache.
-experimental_write_cache_ttl = "1h"
+## +toml2docs:none-default
+experimental_write_cache_ttl = "8h"
 
 ## Buffer size for SST writing.
 sst_write_buffer_size = "8MB"
@@ -415,6 +493,10 @@ parallel_scan_channel_size = 32
 ## Whether to allow stale WAL entries read during replay.
 allow_stale_entries = false
 
+## Minimum time interval between two compactions.
+## To align with the old behavior, the default value is 0 (no restrictions).
+min_compaction_interval = "0m"
+
 ## The options for index in Mito engine.
 [region_engine.mito.index]
 
@@ -511,7 +593,7 @@ fork_dictionary_bytes = "1GiB"
 
 ## The logging options.
 [logging]
-## The directory to store the log files.
+## The directory to store the log files. If set to empty, logs will not be written to files.
 dir = "/tmp/greptimedb/logs"
 
 ## The log level. Can be `info`/`debug`/`warn`/`error`.
@@ -522,12 +604,14 @@ level = "info"
 enable_otlp_tracing = false
 
 ## The OTLP tracing endpoint.
-## +toml2docs:none-default
-otlp_endpoint = ""
+otlp_endpoint = "http://localhost:4317"
 
 ## Whether to append logs to stdout.
 append_stdout = true
 
+## The log format. Can be `text`/`json`.
+log_format = "text"
+
 ## The percentage of tracing will be sampled and exported.
 ## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
 ## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
@@ -545,12 +629,13 @@ enable = false
 write_interval = "30s"
 
 ## For `standalone` mode, `self_import` is recommend to collect metrics generated by itself
+## You must create the database before enabling it.
 [export_metrics.self_import]
 ## +toml2docs:none-default
-db = "information_schema"
+db = "greptime_metrics"
 
 [export_metrics.remote_write]
-## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=information_schema`.
+## The url the metrics send to. The url example can be: `http://127.0.0.1:4000/v1/prometheus/write?db=greptime_metrics`.
 url = ""
 
 ## HTTP headers of Prometheus remote-write carry.

docs/benchmarks/log/README.md

@@ -0,0 +1,51 @@
+# Log benchmark configuration
+This repo holds the configuration we used to benchmark GreptimeDB, Clickhouse and Elastic Search.
+
+Here are the versions of databases we used in the benchmark
+
+| name          | version    |
+| :------------ | :--------- |
+| GreptimeDB    | v0.9.2     |
+| Clickhouse    | 24.9.1.219 |
+| Elasticsearch | 8.15.0     |
+
+## Structured model vs Unstructured model
+We divide test into two parts, using structured model and unstructured model accordingly. You can also see the difference in create table clause.
+
+__Structured model__
+
+The log data is pre-processed into columns by vector. For example an insert request looks like following
+```SQL
+INSERT INTO test_table (bytes, http_version, ip, method, path, status, user, timestamp) VALUES ()
+```
+The goal is to test string/text support for each database. In real scenarios it means the datasource(or log data producers) have separate fields defined, or have already processed the raw input.
+
+__Unstructured model__
+
+The log data is inserted as a long string, and then we build fulltext index upon these strings. For example an insert request looks like following
+```SQL
+INSERT INTO test_table (message, timestamp) VALUES ()
+```
+The goal is to test fuzzy search performance for each database. In real scenarios it means the log is produced by some kind of middleware and inserted directly into the database.
+
+## Creating tables
+See [here](./create_table.sql) for GreptimeDB and Clickhouse's create table clause.
+The mapping of Elastic search is created automatically.
+
+## Vector Configuration
+We use vector to generate random log data and send inserts to databases.
+Please refer to [structured config](./structured_vector.toml) and [unstructured config](./unstructured_vector.toml) for detailed configuration.
+
+## SQLs and payloads
+Please refer to [SQL query](./query.sql) for GreptimeDB and Clickhouse, and [query payload](./query.md) for Elastic search.
+
+## Steps to reproduce
+0. Decide whether to run structured model test or unstructured mode test.
+1. Build vector binary(see vector's config file for specific branch) and databases binaries accordingly.
+2. Create table in GreptimeDB and Clickhouse in advance.
+3. Run vector to insert data.
+4. When data insertion is finished, run queries against each database. Note: you'll need to update timerange value after data insertion.
+
+## Addition
+- You can tune GreptimeDB's configuration to get better performance.
+- You can setup GreptimeDB to use S3 as storage, see [here](https://docs.greptime.com/user-guide/operations/configuration/#storage-options).

docs/benchmarks/log/create_table.sql

@@ -0,0 +1,56 @@
+-- GreptimeDB create table clause
+-- structured test, use vector to pre-process log data into fields
+CREATE TABLE IF NOT EXISTS `test_table` (
+    `bytes` Int64 NULL,
+    `http_version` STRING NULL,
+    `ip` STRING NULL,
+    `method` STRING NULL,
+    `path` STRING NULL,
+    `status` SMALLINT UNSIGNED NULL,
+    `user` STRING NULL,
+    `timestamp` TIMESTAMP(3) NOT NULL,
+    PRIMARY KEY (`user`, `path`, `status`),
+    TIME INDEX (`timestamp`)
+)
+ENGINE=mito
+WITH(
+    append_mode = 'true'
+);
+
+-- unstructured test, build fulltext index on message column
+CREATE TABLE IF NOT EXISTS `test_table` (
+    `message` STRING NULL FULLTEXT WITH(analyzer = 'English', case_sensitive = 'false'),
+    `timestamp` TIMESTAMP(3) NOT NULL,
+    TIME INDEX (`timestamp`)
+)
+ENGINE=mito
+WITH(
+    append_mode = 'true'
+);
+
+-- Clickhouse create table clause
+-- structured test
+CREATE TABLE IF NOT EXISTS test_table
+(
+    bytes UInt64 NOT NULL,
+    http_version String NOT NULL,
+    ip String NOT NULL,
+    method String NOT NULL,
+    path String NOT NULL,
+    status UInt8 NOT NULL,
+    user String NOT NULL,
+    timestamp String NOT NULL,
+)
+ENGINE = MergeTree()
+ORDER BY (user, path, status);
+
+-- unstructured test
+SET allow_experimental_full_text_index = true;
+CREATE TABLE IF NOT EXISTS test_table
+(
+    message String,
+    timestamp String,
+    INDEX inv_idx(message) TYPE full_text(0) GRANULARITY 1
+)
+ENGINE = MergeTree()
+ORDER BY tuple();

docs/benchmarks/log/query.md

@@ -0,0 +1,199 @@
+# Query URL and payload for Elastic Search
+## Count
+URL: `http://127.0.0.1:9200/_count`
+
+## Query by timerange
+URL: `http://127.0.0.1:9200/_search`
+
+You can use the following payload to get the full timerange first.
+```JSON
+{"size":0,"aggs":{"max_timestamp":{"max":{"field":"timestamp"}},"min_timestamp":{"min":{"field":"timestamp"}}}}
+```
+
+And then use this payload to query by timerange.
+```JSON
+{
+  "from": 0,
+  "size": 1000,
+  "query": {
+    "range": {
+      "timestamp": {
+        "gte": "2024-08-16T04:30:44.000Z",
+        "lte": "2024-08-16T04:51:52.000Z"
+      }
+    }
+  }
+}
+```
+
+## Query by condition
+URL: `http://127.0.0.1:9200/_search`
+### Structured payload
+```JSON
+{
+  "from": 0,
+  "size": 10000,
+  "query": {
+    "bool": {
+      "must": [
+        {
+          "term": {
+            "user.keyword": "CrucifiX"
+          }
+        },
+        {
+          "term": {
+            "method.keyword": "OPTION"
+          }
+        },
+        {
+          "term": {
+            "path.keyword": "/user/booperbot124"
+          }
+        },
+        {
+          "term": {
+            "http_version.keyword": "HTTP/1.1"
+          }
+        },
+        {
+          "term": {
+            "status": "401"
+          }
+        }
+      ]
+    }
+  }
+}
+```
+### Unstructured payload
+```JSON
+{
+  "from": 0,
+  "size": 10000,
+  "query": {
+    "bool": {
+      "must": [
+        {
+          "match_phrase": {
+            "message": "CrucifiX"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "OPTION"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "/user/booperbot124"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "HTTP/1.1"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "401"
+          }
+        }
+      ]
+    }
+  }
+}
+```
+
+## Query by condition and timerange
+URL: `http://127.0.0.1:9200/_search`
+### Structured payload
+```JSON
+{
+  "size": 10000,
+  "query": {
+    "bool": {
+      "must": [
+        {
+          "term": {
+            "user.keyword": "CrucifiX"
+          }
+        },
+        {
+          "term": {
+            "method.keyword": "OPTION"
+          }
+        },
+        {
+          "term": {
+            "path.keyword": "/user/booperbot124"
+          }
+        },
+        {
+          "term": {
+            "http_version.keyword": "HTTP/1.1"
+          }
+        },
+        {
+          "term": {
+            "status": "401"
+          }
+        },
+        {
+          "range": {
+            "timestamp": {
+              "gte": "2024-08-19T07:03:37.383Z",
+              "lte": "2024-08-19T07:24:58.883Z"
+            }
+          }
+        }
+      ]
+    }
+  }
+}
+```
+### Unstructured payload
+```JSON
+{
+  "size": 10000,
+  "query": {
+    "bool": {
+      "must": [
+        {
+          "match_phrase": {
+            "message": "CrucifiX"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "OPTION"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "/user/booperbot124"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "HTTP/1.1"
+          }
+        },
+        {
+          "match_phrase": {
+            "message": "401"
+          }
+        },
+        {
+          "range": {
+            "timestamp": {
+              "gte": "2024-08-19T05:16:17.099Z",
+              "lte": "2024-08-19T05:46:02.722Z"
+            }
+          }
+        }
+      ]
+    }
+  }
+}
+```

docs/benchmarks/log/query.sql

@@ -0,0 +1,50 @@
+-- Structured query for GreptimeDB and Clickhouse
+
+-- query count
+select count(*) from test_table;
+
+-- query by timerange. Note: place the timestamp range in the where clause
+-- GreptimeDB
+-- you can use `select max(timestamp)::bigint from test_table;` and `select min(timestamp)::bigint from test_table;`
+-- to get the full timestamp range
+select * from test_table where timestamp between 1723710843619 and 1723711367588;
+-- Clickhouse
+-- you can use `select max(timestamp) from test_table;` and `select min(timestamp) from test_table;`
+-- to get the full timestamp range
+select * from test_table where timestamp between '2024-08-16T03:58:46Z' and '2024-08-16T04:03:50Z';
+
+-- query by condition
+SELECT * FROM test_table WHERE user = 'CrucifiX' and method = 'OPTION' and path = '/user/booperbot124' and http_version = 'HTTP/1.1' and status = 401;
+
+-- query by condition and timerange
+-- GreptimeDB
+SELECT * FROM test_table WHERE user = "CrucifiX" and method = "OPTION" and path = "/user/booperbot124" and http_version = "HTTP/1.1" and status = 401 
+and timestamp between 1723774396760 and 1723774788760;
+-- Clickhouse
+SELECT * FROM test_table WHERE user = 'CrucifiX' and method = 'OPTION' and path = '/user/booperbot124' and http_version = 'HTTP/1.1' and status = 401 
+and timestamp between '2024-08-16T03:58:46Z' and '2024-08-16T04:03:50Z';
+
+-- Unstructured query for GreptimeDB and Clickhouse
+
+
+-- query by condition
+-- GreptimeDB
+SELECT * FROM test_table WHERE MATCHES(message, "+CrucifiX +OPTION +/user/booperbot124 +HTTP/1.1 +401");
+-- Clickhouse
+SELECT * FROM test_table WHERE (message LIKE '%CrucifiX%') 
+AND (message LIKE '%OPTION%') 
+AND (message LIKE '%/user/booperbot124%') 
+AND (message LIKE '%HTTP/1.1%') 
+AND (message LIKE '%401%');
+
+-- query by condition and timerange
+-- GreptimeDB
+SELECT * FROM test_table WHERE MATCHES(message, "+CrucifiX +OPTION +/user/booperbot124 +HTTP/1.1 +401") 
+and timestamp between 1723710843619 and 1723711367588;
+-- Clickhouse
+SELECT * FROM test_table WHERE (message LIKE '%CrucifiX%') 
+AND (message LIKE '%OPTION%') 
+AND (message LIKE '%/user/booperbot124%') 
+AND (message LIKE '%HTTP/1.1%') 
+AND (message LIKE '%401%') 
+AND timestamp between '2024-08-15T10:25:26.524000000Z' AND '2024-08-15T10:31:31.746000000Z';

docs/benchmarks/log/structured_vector.toml

@@ -0,0 +1,57 @@
+# Please note we use patched branch to build vector
+# https://github.com/shuiyisong/vector/tree/chore/greptime_log_ingester_logitem
+
+[sources.demo_logs]
+type = "demo_logs"
+format = "apache_common"
+# interval value = 1 / rps
+# say you want to insert at 20k/s, that is 1 / 20000 = 0.00005
+# set to 0 to run as fast as possible
+interval = 0
+# total rows to insert
+count = 100000000
+lines = [ "line1" ]
+
+[transforms.parse_logs]
+type = "remap"
+inputs = ["demo_logs"]
+source = '''
+. = parse_regex!(.message, r'^(?P<ip>\S+) - (?P<user>\S+) \[(?P<timestamp>[^\]]+)\] "(?P<method>\S+) (?P<path>\S+) (?P<http_version>\S+)" (?P<status>\d+) (?P<bytes>\d+)$')
+
+# Convert timestamp to a standard format
+.timestamp = parse_timestamp!(.timestamp, format: "%d/%b/%Y:%H:%M:%S %z")
+
+# Convert status and bytes to integers
+.status = to_int!(.status)
+.bytes = to_int!(.bytes)
+'''
+
+[sinks.sink_greptime_logs]
+type = "greptimedb_logs"
+# The table to insert into
+table = "test_table"
+pipeline_name = "demo_pipeline"
+compression = "none"
+inputs = [ "parse_logs" ]
+endpoint = "http://127.0.0.1:4000"
+# Batch size for each insertion
+batch.max_events = 4000
+
+[sinks.clickhouse]
+type = "clickhouse"
+inputs = [ "parse_logs" ]
+database = "default"
+endpoint = "http://127.0.0.1:8123"
+format = "json_each_row"
+# The table to insert into
+table = "test_table"
+
+[sinks.sink_elasticsearch]
+type = "elasticsearch"
+inputs = [ "parse_logs" ]
+api_version = "auto"
+compression = "none"
+doc_type = "_doc"
+endpoints = [ "http://127.0.0.1:9200" ]
+id_key = "id"
+mode = "bulk"

docs/benchmarks/log/unstructured_vector.toml

@@ -0,0 +1,43 @@
+# Please note we use patched branch to build vector
+# https://github.com/shuiyisong/vector/tree/chore/greptime_log_ingester_ft
+
+[sources.demo_logs]
+type = "demo_logs"
+format = "apache_common"
+# interval value = 1 / rps
+# say you want to insert at 20k/s, that is 1 / 20000 = 0.00005
+# set to 0 to run as fast as possible
+interval = 0
+# total rows to insert
+count = 100000000
+lines = [ "line1" ]
+
+[sinks.sink_greptime_logs]
+type = "greptimedb_logs"
+# The table to insert into
+table = "test_table"
+pipeline_name = "demo_pipeline"
+compression = "none"
+inputs = [ "demo_logs" ]
+endpoint = "http://127.0.0.1:4000"
+# Batch size for each insertion
+batch.max_events = 500
+
+[sinks.clickhouse]
+type = "clickhouse"
+inputs = [ "demo_logs" ]
+database = "default"
+endpoint = "http://127.0.0.1:8123"
+format = "json_each_row"
+# The table to insert into
+table = "test_table"
+
+[sinks.sink_elasticsearch]
+type = "elasticsearch"
+inputs = [ "demo_logs" ]
+api_version = "auto"
+compression = "none"
+doc_type = "_doc"
+endpoints = [ "http://127.0.0.1:9200" ]
+id_key = "id"
+mode = "bulk"

docs/benchmarks/tsbs/v0.9.1.md

@@ -0,0 +1,58 @@
+# TSBS benchmark - v0.9.1
+
+## Environment
+
+### Local
+
+|        |                                    |
+| ------ | ---------------------------------- |
+| CPU    | AMD Ryzen 7 7735HS (8 core 3.2GHz) |
+| Memory | 32GB                               |
+| Disk   | SOLIDIGM SSDPFKNU010TZ             |
+| OS     | Ubuntu 22.04.2 LTS                 |
+
+### Amazon EC2
+
+|         |                         |
+| ------- | ----------------------- |
+| Machine | c5d.2xlarge             |
+| CPU     | 8 core                  |
+| Memory  | 16GB                    |
+| Disk    | 100GB (GP3)             |
+| OS      | Ubuntu Server 24.04 LTS |
+
+## Write performance
+
+| Environment     | Ingest rate (rows/s) |
+| --------------- | -------------------- |
+| Local           | 387697.68            |
+| EC2 c5d.2xlarge | 234620.19            |
+
+## Query performance
+
+| Query type            | Local (ms) | EC2 c5d.2xlarge (ms) |
+| --------------------- | ---------- | -------------------- |
+| cpu-max-all-1         | 21.14      | 14.75                |
+| cpu-max-all-8         | 36.79      | 30.69                |
+| double-groupby-1      | 529.02     | 987.85               |
+| double-groupby-5      | 1064.53    | 1455.95              |
+| double-groupby-all    | 1625.33    | 2143.96              |
+| groupby-orderby-limit | 529.19     | 1353.49              |
+| high-cpu-1            | 12.09      | 8.24                 |
+| high-cpu-all          | 3619.47    | 5312.82              |
+| lastpoint             | 224.91     | 576.06               |
+| single-groupby-1-1-1  | 10.82      | 6.01                 |
+| single-groupby-1-1-12 | 11.16      | 7.42                 |
+| single-groupby-1-8-1  | 13.50      | 10.20                |
+| single-groupby-5-1-1  | 11.99      | 6.70                 |
+| single-groupby-5-1-12 | 13.17      | 8.72                 |
+| single-groupby-5-8-1  | 16.01      | 12.07                |
+
+`single-groupby-1-1-1` query throughput
+
+| Environment     | Client concurrency | mean time (ms) | qps (queries/sec) |
+| --------------- | ------------------ | -------------- | ----------------- |
+| Local           | 50                 | 33.04          | 1511.74           |
+| Local           | 100                | 67.70          | 1476.14           |
+| EC2 c5d.2xlarge | 50                 | 61.93          | 806.97            |
+| EC2 c5d.2xlarge | 100                | 126.31         | 791.40            |

grafana/README.md

@@ -25,7 +25,7 @@ Please ensure the following configuration before importing the dashboard into Gr
 
 __1. Prometheus scrape config__
 
-Assign `greptime_pod` label to each host target. We use this label to identify each node instance.
+Configure Prometheus to scrape the cluster.
 
 ```yml
 # example config
@@ -34,27 +34,15 @@ Assign `greptime_pod` label to each host target. We use this label to identify e
 scrape_configs:
   - job_name: metasrv
     static_configs:
-    - targets: ['<ip>:<port>']
-      labels:
-        greptime_pod: metasrv
+    - targets: ['<metasrv-ip>:<port>']
 
   - job_name: datanode
     static_configs:
-    - targets: ['<ip>:<port>']
-      labels:
-        greptime_pod: datanode1
-    - targets: ['<ip>:<port>']
-      labels:
-        greptime_pod: datanode2
-    - targets: ['<ip>:<port>']
-      labels:
-        greptime_pod: datanode3
+    - targets: ['<datanode0-ip>:<port>', '<datanode1-ip>:<port>', '<datanode2-ip>:<port>']
 
   - job_name: frontend
     static_configs:
-    - targets: ['<ip>:<port>']
-      labels:
-        greptime_pod: frontend
+    - targets: ['<frontend-ip>:<port>']
 ```
 
 __2. Grafana config__
@@ -63,4 +51,4 @@ Create a Prometheus data source in Grafana before using this dashboard. We use `
 
 ### Usage
 
-Use `datasource` or `greptime_pod` on the upper-left corner to filter data from certain node.
+Use `datasource` or `instance` on the upper-left corner to filter data from certain node.

rust-toolchain.toml

@@ -1,2 +1,3 @@
 [toolchain]
-channel = "nightly-2024-04-20"
+channel = "nightly-2024-06-06"
+

scripts/check-builder-rust-version.sh

@@ -0,0 +1,42 @@
+#!/usr/bin/env bash
+
+set -e
+
+RUST_TOOLCHAIN_VERSION_FILE="rust-toolchain.toml"
+DEV_BUILDER_UBUNTU_REGISTRY="docker.io"
+DEV_BUILDER_UBUNTU_NAMESPACE="greptime"
+DEV_BUILDER_UBUNTU_NAME="dev-builder-ubuntu"
+
+function check_rust_toolchain_version() {
+  DEV_BUILDER_IMAGE_TAG=$(grep "DEV_BUILDER_IMAGE_TAG ?= " Makefile | cut -d= -f2 | sed 's/^[ \t]*//')
+  if [ -z "$DEV_BUILDER_IMAGE_TAG" ]; then
+    echo "Error: No DEV_BUILDER_IMAGE_TAG found in Makefile"
+    exit 1
+  fi
+
+  DEV_BUILDER_UBUNTU_IMAGE="$DEV_BUILDER_UBUNTU_REGISTRY/$DEV_BUILDER_UBUNTU_NAMESPACE/$DEV_BUILDER_UBUNTU_NAME:$DEV_BUILDER_IMAGE_TAG"
+
+  CURRENT_VERSION=$(grep -Eo '[0-9]{4}-[0-9]{2}-[0-9]{2}' "$RUST_TOOLCHAIN_VERSION_FILE")
+  if [ -z "$CURRENT_VERSION" ]; then
+    echo "Error: No rust toolchain version found in $RUST_TOOLCHAIN_VERSION_FILE"
+    exit 1
+  fi
+
+  RUST_TOOLCHAIN_VERSION_IN_BUILDER=$(docker run "$DEV_BUILDER_UBUNTU_IMAGE" rustc --version | grep  -Eo '[0-9]{4}-[0-9]{2}-[0-9]{2}')
+  if [ -z "$RUST_TOOLCHAIN_VERSION_IN_BUILDER" ]; then
+    echo "Error: No rustc version found in $DEV_BUILDER_UBUNTU_IMAGE"
+    exit 1
+  fi
+
+  # Compare the version and the difference should be less than 1 day.
+  current_rust_toolchain_seconds=$(date -d "$CURRENT_VERSION" +%s)
+  rust_toolchain_in_dev_builder_ubuntu_seconds=$(date -d "$RUST_TOOLCHAIN_VERSION_IN_BUILDER" +%s)
+  date_diff=$(( (current_rust_toolchain_seconds - rust_toolchain_in_dev_builder_ubuntu_seconds) / 86400 ))
+
+  if [ $date_diff -gt 1 ]; then
+    echo "Error: The rust toolchain '$RUST_TOOLCHAIN_VERSION_IN_BUILDER' in builder '$DEV_BUILDER_UBUNTU_IMAGE' maybe outdated, please update it to '$CURRENT_VERSION'"
+    exit 1
+  fi
+}
+
+check_rust_toolchain_version

scripts/check-snafu.py

@@ -0,0 +1,69 @@
+# 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.
+
+import os
+import re
+
+
+def find_rust_files(directory):
+    error_files = []
+    other_rust_files = []
+    for root, _, files in os.walk(directory):
+        for file in files:
+            if file == "error.rs":
+                error_files.append(os.path.join(root, file))
+            elif file.endswith(".rs"):
+                other_rust_files.append(os.path.join(root, file))
+    return error_files, other_rust_files
+
+
+def extract_branch_names(file_content):
+    pattern = re.compile(r"#\[snafu\(display\([^\)]*\)\)\]\s*(\w+)\s*\{")
+    return pattern.findall(file_content)
+
+
+def check_snafu_in_files(branch_name, rust_files):
+    branch_name_snafu = f"{branch_name}Snafu"
+    for rust_file in rust_files:
+        with open(rust_file, "r") as file:
+            content = file.read()
+            if branch_name_snafu in content:
+                return True
+    return False
+
+
+def main():
+    error_files, other_rust_files = find_rust_files(".")
+    branch_names = []
+
+    for error_file in error_files:
+        with open(error_file, "r") as file:
+            content = file.read()
+            branch_names.extend(extract_branch_names(content))
+
+    unused_snafu = [
+        branch_name
+        for branch_name in branch_names
+        if not check_snafu_in_files(branch_name, other_rust_files)
+    ]
+
+    for name in unused_snafu:
+        print(name)
+
+    if unused_snafu:
+        raise SystemExit(1)
+
+
+if __name__ == "__main__":
+    main()

scripts/install.sh

@@ -1,62 +1,72 @@
-#!/bin/sh
+#!/usr/bin/env bash
 
 set -ue
 
 OS_TYPE=
 ARCH_TYPE=
+
+# Set the GitHub token to avoid GitHub API rate limit.
+# You can run with `GITHUB_TOKEN`:
+#  GITHUB_TOKEN=<your_token> ./scripts/install.sh
+GITHUB_TOKEN=${GITHUB_TOKEN:-}
+
 VERSION=${1:-latest}
 GITHUB_ORG=GreptimeTeam
 GITHUB_REPO=greptimedb
 BIN=greptime
 
-get_os_type() {
-    os_type="$(uname -s)"
+function get_os_type() {
+  os_type="$(uname -s)"
 
-    case "$os_type" in
+  case "$os_type" in
     Darwin)
-        OS_TYPE=darwin
-        ;;
+      OS_TYPE=darwin
+      ;;
     Linux)
-        OS_TYPE=linux
-        ;;
+      OS_TYPE=linux
+      ;;
     *)
-        echo "Error: Unknown OS type: $os_type"
-        exit 1
-    esac
+      echo "Error: Unknown OS type: $os_type"
+      exit 1
+  esac
 }
 
-get_arch_type() {
-    arch_type="$(uname -m)"
+function get_arch_type() {
+  arch_type="$(uname -m)"
 
-    case "$arch_type" in
+  case "$arch_type" in
     arm64)
-        ARCH_TYPE=arm64
-        ;;
+      ARCH_TYPE=arm64
+      ;;
     aarch64)
-        ARCH_TYPE=arm64
-        ;;
+      ARCH_TYPE=arm64
+      ;;
     x86_64)
-        ARCH_TYPE=amd64
-        ;;
+      ARCH_TYPE=amd64
+      ;;
     amd64)
-        ARCH_TYPE=amd64
-        ;;
+      ARCH_TYPE=amd64
+      ;;
     *)
-        echo "Error: Unknown CPU type: $arch_type"
-        exit 1
-    esac
+      echo "Error: Unknown CPU type: $arch_type"
+      exit 1
+  esac
 }
 
-get_os_type
-get_arch_type
-
-if [ -n "${OS_TYPE}" ] && [ -n "${ARCH_TYPE}" ]; then
-    # Use the latest nightly version.
+function download_artifact() {
+  if [ -n "${OS_TYPE}" ] && [ -n "${ARCH_TYPE}" ]; then
+    # Use the latest stable released version.
+    # GitHub API reference: https://docs.github.com/en/rest/releases/releases?apiVersion=2022-11-28#get-the-latest-release.
     if [ "${VERSION}" = "latest" ]; then
-        VERSION=$(curl -s -XGET "https://api.github.com/repos/${GITHUB_ORG}/${GITHUB_REPO}/releases" | grep tag_name | grep nightly | cut -d: -f 2 | sed 's/.*"\(.*\)".*/\1/' | uniq | sort -r | head -n 1)
-        if [ -z "${VERSION}" ]; then
-            echo "Failed to get the latest version."
-            exit 1
+      # To avoid other tools dependency, we choose to use `curl` to get the version metadata and parsed by `sed`.
+      VERSION=$(curl -sL \
+        -H "Accept: application/vnd.github+json" \
+        -H "X-GitHub-Api-Version: 2022-11-28" \
+        ${GITHUB_TOKEN:+-H "Authorization: Bearer $GITHUB_TOKEN"} \
+        "https://api.github.com/repos/${GITHUB_ORG}/${GITHUB_REPO}/releases/latest" | sed -n 's/.*"tag_name": "\([^"]*\)".*/\1/p')
+      if [ -z "${VERSION}" ]; then
+        echo "Failed to get the latest stable released version."
+        exit 1
         fi
     fi
 
@@ -73,4 +83,9 @@ if [ -n "${OS_TYPE}" ] && [ -n "${ARCH_TYPE}" ]; then
       rm -r "${PACKAGE_NAME%.tar.gz}" && \
       echo "Run './${BIN} --help' to get started"
     fi
-fi
+  fi
+}
+
+get_os_type
+get_arch_type
+download_artifact

src/auth/src/tests.rs

@@ -13,9 +13,11 @@
 // limitations under the License.
 
 use common_base::secrets::ExposeSecret;
+use common_error::ext::BoxedError;
+use snafu::{OptionExt, ResultExt};
 
 use crate::error::{
-    AccessDeniedSnafu, Result, UnsupportedPasswordTypeSnafu, UserNotFoundSnafu,
+    AccessDeniedSnafu, AuthBackendSnafu, Result, UnsupportedPasswordTypeSnafu, UserNotFoundSnafu,
     UserPasswordMismatchSnafu,
 };
 use crate::user_info::DefaultUserInfo;
@@ -49,6 +51,19 @@ impl MockUserProvider {
         info.schema.clone_into(&mut self.schema);
         info.username.clone_into(&mut self.username);
     }
+
+    // this is a deliberate function to ref AuthBackendSnafu
+    // so that it won't get deleted in the future
+    pub fn ref_auth_backend_snafu(&self) -> Result<()> {
+        let none_option = None;
+
+        none_option
+            .context(UserNotFoundSnafu {
+                username: "no_user".to_string(),
+            })
+            .map_err(BoxedError::new)
+            .context(AuthBackendSnafu)
+    }
 }
 
 #[async_trait::async_trait]

src/auth/tests/mod.rs

@@ -18,6 +18,7 @@ use std::sync::Arc;
 
 use api::v1::greptime_request::Request;
 use auth::error::Error::InternalState;
+use auth::error::InternalStateSnafu;
 use auth::{PermissionChecker, PermissionCheckerRef, PermissionReq, PermissionResp, UserInfoRef};
 use sql::statements::show::{ShowDatabases, ShowKind};
 use sql::statements::statement::Statement;
@@ -33,9 +34,10 @@ impl PermissionChecker for DummyPermissionChecker {
         match req {
             PermissionReq::GrpcRequest(_) => Ok(PermissionResp::Allow),
             PermissionReq::SqlStatement(_) => Ok(PermissionResp::Reject),
-            _ => Err(InternalState {
+            _ => InternalStateSnafu {
                 msg: "testing".to_string(),
-            }),
+            }
+            .fail(),
         }
     }
 }

src/catalog/Cargo.toml

@@ -40,6 +40,7 @@ moka = { workspace = true, features = ["future", "sync"] }
 partition.workspace = true
 paste = "1.0"
 prometheus.workspace = true
+rustc-hash.workspace = true
 serde_json.workspace = true
 session.workspace = true
 snafu.workspace = true

src/catalog/src/error.rs

@@ -97,13 +97,6 @@ pub enum Error {
         source: table::error::Error,
     },
 
-    #[snafu(display("System catalog is not valid: {}", msg))]
-    SystemCatalog {
-        msg: String,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
     #[snafu(display("Cannot find catalog by name: {}", catalog_name))]
     CatalogNotFound {
         catalog_name: String,
@@ -186,13 +179,6 @@ pub enum Error {
         source: common_query::error::Error,
     },
 
-    #[snafu(display("Failed to perform metasrv operation"))]
-    Metasrv {
-        #[snafu(implicit)]
-        location: Location,
-        source: meta_client::error::Error,
-    },
-
     #[snafu(display("Invalid table info in catalog"))]
     InvalidTableInfoInCatalog {
         #[snafu(implicit)]
@@ -288,8 +274,6 @@ impl ErrorExt for Error {
 
             Error::FlowInfoNotFound { .. } => StatusCode::FlowNotFound,
 
-            Error::SystemCatalog { .. } => StatusCode::StorageUnavailable,
-
             Error::UpgradeWeakCatalogManagerRef { .. } => StatusCode::Internal,
 
             Error::CreateRecordBatch { source, .. } => source.status_code(),
@@ -303,7 +287,6 @@ impl ErrorExt for Error {
 
             Error::CreateTable { source, .. } => source.status_code(),
 
-            Error::Metasrv { source, .. } => source.status_code(),
             Error::DecodePlan { source, .. } => source.status_code(),
             Error::InvalidTableInfoInCatalog { source, .. } => source.status_code(),
 
@@ -338,27 +321,6 @@ mod tests {
 
     use super::*;
 
-    #[test]
-    pub fn test_error_status_code() {
-        assert_eq!(
-            StatusCode::TableAlreadyExists,
-            Error::TableExists {
-                table: "some_table".to_string(),
-                location: Location::generate(),
-            }
-            .status_code()
-        );
-
-        assert_eq!(
-            StatusCode::StorageUnavailable,
-            Error::SystemCatalog {
-                msg: String::default(),
-                location: Location::generate(),
-            }
-            .status_code()
-        );
-    }
-
     #[test]
     pub fn test_errors_to_datafusion_error() {
         let e: DataFusionError = Error::TableExists {

src/catalog/src/kvbackend/client.rs

@@ -20,8 +20,8 @@ use std::time::Duration;
 
 use common_error::ext::BoxedError;
 use common_meta::cache_invalidator::KvCacheInvalidator;
-use common_meta::error::Error::{CacheNotGet, GetKvCache};
-use common_meta::error::{CacheNotGetSnafu, Error, ExternalSnafu, Result};
+use common_meta::error::Error::CacheNotGet;
+use common_meta::error::{CacheNotGetSnafu, Error, ExternalSnafu, GetKvCacheSnafu, Result};
 use common_meta::kv_backend::{KvBackend, KvBackendRef, TxnService};
 use common_meta::rpc::store::{
     BatchDeleteRequest, BatchDeleteResponse, BatchGetRequest, BatchGetResponse, BatchPutRequest,
@@ -282,8 +282,11 @@ impl KvBackend for CachedMetaKvBackend {
                 _ => Err(e),
             },
         }
-        .map_err(|e| GetKvCache {
-            err_msg: e.to_string(),
+        .map_err(|e| {
+            GetKvCacheSnafu {
+                err_msg: e.to_string(),
+            }
+            .build()
         });
 
         // "cache.invalidate_key" and "cache.try_get_with_by_ref" are not mutually exclusive. So we need

src/catalog/src/kvbackend/manager.rs

@@ -313,7 +313,7 @@ struct SystemCatalog {
     catalog_cache: Cache<String, Arc<InformationSchemaProvider>>,
     pg_catalog_cache: Cache<String, Arc<PGCatalogProvider>>,
 
-    // system_schema_provier for default catalog
+    // system_schema_provider for default catalog
     information_schema_provider: Arc<InformationSchemaProvider>,
     pg_catalog_provider: Arc<PGCatalogProvider>,
     backend: KvBackendRef,

src/catalog/src/system_schema/pg_catalog.rs

@@ -30,6 +30,7 @@ use pg_namespace::PGNamespace;
 use table::TableRef;
 pub use table_names::*;
 
+use self::pg_namespace::oid_map::{PGNamespaceOidMap, PGNamespaceOidMapRef};
 use super::memory_table::MemoryTable;
 use super::utils::tables::u32_column;
 use super::{SystemSchemaProvider, SystemSchemaProviderInner, SystemTableRef};
@@ -52,6 +53,9 @@ pub struct PGCatalogProvider {
     catalog_name: String,
     catalog_manager: Weak<dyn CatalogManager>,
     tables: HashMap<String, TableRef>,
+
+    // Workaround to store mapping of schema_name to a numeric id
+    namespace_oid_map: PGNamespaceOidMapRef,
 }
 
 impl SystemSchemaProvider for PGCatalogProvider {
@@ -85,6 +89,7 @@ impl PGCatalogProvider {
             catalog_name,
             catalog_manager,
             tables: HashMap::new(),
+            namespace_oid_map: Arc::new(PGNamespaceOidMap::new()),
         };
         provider.build_tables();
         provider
@@ -122,10 +127,12 @@ impl SystemSchemaProviderInner for PGCatalogProvider {
             table_names::PG_NAMESPACE => Some(Arc::new(PGNamespace::new(
                 self.catalog_name.clone(),
                 self.catalog_manager.clone(),
+                self.namespace_oid_map.clone(),
             ))),
             table_names::PG_CLASS => Some(Arc::new(PGClass::new(
                 self.catalog_name.clone(),
                 self.catalog_manager.clone(),
+                self.namespace_oid_map.clone(),
             ))),
             _ => None,
         }

src/catalog/src/system_schema/pg_catalog/pg_class.rs

@@ -31,6 +31,7 @@ use snafu::{OptionExt, ResultExt};
 use store_api::storage::ScanRequest;
 use table::metadata::TableType;
 
+use super::pg_namespace::oid_map::PGNamespaceOidMapRef;
 use super::{OID_COLUMN_NAME, PG_CLASS};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
@@ -60,14 +61,22 @@ pub(super) struct PGClass {
     schema: SchemaRef,
     catalog_name: String,
     catalog_manager: Weak<dyn CatalogManager>,
+
+    // Workaround to convert schema_name to a numeric id
+    namespace_oid_map: PGNamespaceOidMapRef,
 }
 
 impl PGClass {
-    pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
+    pub(super) fn new(
+        catalog_name: String,
+        catalog_manager: Weak<dyn CatalogManager>,
+        namespace_oid_map: PGNamespaceOidMapRef,
+    ) -> Self {
         Self {
             schema: Self::schema(),
             catalog_name,
             catalog_manager,
+            namespace_oid_map,
         }
     }
 
@@ -75,7 +84,7 @@ impl PGClass {
         Arc::new(Schema::new(vec![
             u32_column(OID_COLUMN_NAME),
             string_column(RELNAME),
-            string_column(RELNAMESPACE),
+            u32_column(RELNAMESPACE),
             string_column(RELKIND),
             u32_column(RELOWNER),
         ]))
@@ -86,6 +95,7 @@ impl PGClass {
             self.schema.clone(),
             self.catalog_name.clone(),
             self.catalog_manager.clone(),
+            self.namespace_oid_map.clone(),
         )
     }
 }
@@ -155,10 +165,11 @@ struct PGClassBuilder {
     schema: SchemaRef,
     catalog_name: String,
     catalog_manager: Weak<dyn CatalogManager>,
+    namespace_oid_map: PGNamespaceOidMapRef,
 
     oid: UInt32VectorBuilder,
     relname: StringVectorBuilder,
-    relnamespace: StringVectorBuilder,
+    relnamespace: UInt32VectorBuilder,
     relkind: StringVectorBuilder,
     relowner: UInt32VectorBuilder,
 }
@@ -168,15 +179,17 @@ impl PGClassBuilder {
         schema: SchemaRef,
         catalog_name: String,
         catalog_manager: Weak<dyn CatalogManager>,
+        namespace_oid_map: PGNamespaceOidMapRef,
     ) -> Self {
         Self {
             schema,
             catalog_name,
             catalog_manager,
+            namespace_oid_map,
 
             oid: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
             relname: StringVectorBuilder::with_capacity(INIT_CAPACITY),
-            relnamespace: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            relnamespace: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
             relkind: StringVectorBuilder::with_capacity(INIT_CAPACITY),
             relowner: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
         }
@@ -217,6 +230,7 @@ impl PGClassBuilder {
         table: &str,
         kind: &str,
     ) {
+        let namespace_oid = self.namespace_oid_map.get_oid(schema);
         let row = [
             (OID_COLUMN_NAME, &Value::from(oid)),
             (RELNAMESPACE, &Value::from(schema)),
@@ -230,7 +244,7 @@ impl PGClassBuilder {
         }
 
         self.oid.push(Some(oid));
-        self.relnamespace.push(Some(schema));
+        self.relnamespace.push(Some(namespace_oid));
         self.relname.push(Some(table));
         self.relkind.push(Some(kind));
         self.relowner.push(Some(DUMMY_OWNER_ID));

src/catalog/src/system_schema/pg_catalog/pg_namespace.rs

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+pub(super) mod oid_map;
+
 use std::sync::{Arc, Weak};
 
 use arrow_schema::SchemaRef as ArrowSchemaRef;
@@ -25,16 +27,16 @@ use datafusion::physical_plan::streaming::PartitionStream as DfPartitionStream;
 use datatypes::scalars::ScalarVectorBuilder;
 use datatypes::schema::{Schema, SchemaRef};
 use datatypes::value::Value;
-use datatypes::vectors::{StringVectorBuilder, VectorRef};
+use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder, VectorRef};
 use snafu::{OptionExt, ResultExt};
 use store_api::storage::ScanRequest;
 
-use super::{OID_COLUMN_NAME, PG_NAMESPACE};
+use super::{PGNamespaceOidMapRef, OID_COLUMN_NAME, PG_NAMESPACE};
 use crate::error::{
     CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
 };
 use crate::information_schema::Predicates;
-use crate::system_schema::utils::tables::string_column;
+use crate::system_schema::utils::tables::{string_column, u32_column};
 use crate::system_schema::SystemTable;
 use crate::CatalogManager;
 
@@ -48,21 +50,29 @@ pub(super) struct PGNamespace {
     schema: SchemaRef,
     catalog_name: String,
     catalog_manager: Weak<dyn CatalogManager>,
+
+    // Workaround to convert schema_name to a numeric id
+    oid_map: PGNamespaceOidMapRef,
 }
 
 impl PGNamespace {
-    pub(super) fn new(catalog_name: String, catalog_manager: Weak<dyn CatalogManager>) -> Self {
+    pub(super) fn new(
+        catalog_name: String,
+        catalog_manager: Weak<dyn CatalogManager>,
+        oid_map: PGNamespaceOidMapRef,
+    ) -> Self {
         Self {
             schema: Self::schema(),
             catalog_name,
             catalog_manager,
+            oid_map,
         }
     }
 
     fn schema() -> SchemaRef {
         Arc::new(Schema::new(vec![
             // TODO(J0HN50N133): we do not have a numeric schema id, use schema name as a workaround. Use a proper schema id once we have it.
-            string_column(OID_COLUMN_NAME),
+            u32_column(OID_COLUMN_NAME),
             string_column(NSPNAME),
         ]))
     }
@@ -72,6 +82,7 @@ impl PGNamespace {
             self.schema.clone(),
             self.catalog_name.clone(),
             self.catalog_manager.clone(),
+            self.oid_map.clone(),
         )
     }
 }
@@ -138,8 +149,9 @@ struct PGNamespaceBuilder {
     schema: SchemaRef,
     catalog_name: String,
     catalog_manager: Weak<dyn CatalogManager>,
+    namespace_oid_map: PGNamespaceOidMapRef,
 
-    oid: StringVectorBuilder,
+    oid: UInt32VectorBuilder,
     nspname: StringVectorBuilder,
 }
 
@@ -148,12 +160,14 @@ impl PGNamespaceBuilder {
         schema: SchemaRef,
         catalog_name: String,
         catalog_manager: Weak<dyn CatalogManager>,
+        namespace_oid_map: PGNamespaceOidMapRef,
     ) -> Self {
         Self {
             schema,
             catalog_name,
             catalog_manager,
-            oid: StringVectorBuilder::with_capacity(INIT_CAPACITY),
+            namespace_oid_map,
+            oid: UInt32VectorBuilder::with_capacity(INIT_CAPACITY),
             nspname: StringVectorBuilder::with_capacity(INIT_CAPACITY),
         }
     }
@@ -178,14 +192,15 @@ impl PGNamespaceBuilder {
     }
 
     fn add_namespace(&mut self, predicates: &Predicates, schema_name: &str) {
+        let oid = self.namespace_oid_map.get_oid(schema_name);
         let row = [
-            (OID_COLUMN_NAME, &Value::from(schema_name)),
+            (OID_COLUMN_NAME, &Value::from(oid)),
             (NSPNAME, &Value::from(schema_name)),
         ];
         if !predicates.eval(&row) {
             return;
         }
-        self.oid.push(Some(schema_name));
+        self.oid.push(Some(oid));
         self.nspname.push(Some(schema_name));
     }
 }

src/catalog/src/system_schema/pg_catalog/pg_namespace/oid_map.rs

@@ -0,0 +1,100 @@
+// 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::hash::BuildHasher;
+use std::sync::Arc;
+
+use dashmap::DashMap;
+use rustc_hash::FxSeededState;
+
+pub type PGNamespaceOidMapRef = Arc<PGNamespaceOidMap>;
+// Workaround to convert schema_name to a numeric id,
+// remove this when we have numeric schema id in greptime
+pub struct PGNamespaceOidMap {
+    oid_map: DashMap<String, u32>,
+
+    // Rust use SipHasher by default, which provides resistance against DOS attacks.
+    // This will produce different hash value between each greptime instance. This will
+    // cause the sqlness test fail. We need a deterministic hash here to provide
+    // same oid for the same schema name with best effort and DOS attacks aren't concern here.
+    hasher: FxSeededState,
+}
+
+impl PGNamespaceOidMap {
+    pub fn new() -> Self {
+        Self {
+            oid_map: DashMap::new(),
+            hasher: FxSeededState::with_seed(0), // PLEASE DO NOT MODIFY THIS SEED VALUE!!!
+        }
+    }
+
+    fn oid_is_used(&self, oid: u32) -> bool {
+        self.oid_map.iter().any(|e| *e.value() == oid)
+    }
+
+    pub fn get_oid(&self, schema_name: &str) -> u32 {
+        if let Some(oid) = self.oid_map.get(schema_name) {
+            *oid
+        } else {
+            let mut oid = self.hasher.hash_one(schema_name) as u32;
+            while self.oid_is_used(oid) {
+                oid = self.hasher.hash_one(oid) as u32;
+            }
+            self.oid_map.insert(schema_name.to_string(), oid);
+            oid
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+
+    use super::*;
+
+    #[test]
+    fn oid_is_stable() {
+        let oid_map_1 = PGNamespaceOidMap::new();
+        let oid_map_2 = PGNamespaceOidMap::new();
+
+        let schema = "schema";
+        let oid = oid_map_1.get_oid(schema);
+
+        // oid keep stable in the same instance
+        assert_eq!(oid, oid_map_1.get_oid(schema));
+
+        // oid keep stable between different instances
+        assert_eq!(oid, oid_map_2.get_oid(schema));
+    }
+
+    #[test]
+    fn oid_collision() {
+        let oid_map = PGNamespaceOidMap::new();
+
+        let key1 = "3178510";
+        let key2 = "4215648";
+
+        // have collision
+        assert_eq!(
+            oid_map.hasher.hash_one(key1) as u32,
+            oid_map.hasher.hash_one(key2) as u32
+        );
+
+        // insert them into oid_map
+        let oid1 = oid_map.get_oid(key1);
+        let oid2 = oid_map.get_oid(key2);
+
+        // they should have different id
+        assert_ne!(oid1, oid2);
+    }
+}

src/client/src/database.rs

@@ -37,7 +37,8 @@ use tonic::metadata::AsciiMetadataKey;
 use tonic::transport::Channel;
 
 use crate::error::{
-    ConvertFlightDataSnafu, Error, IllegalFlightMessagesSnafu, InvalidAsciiSnafu, ServerSnafu,
+    ConvertFlightDataSnafu, Error, FlightGetSnafu, IllegalFlightMessagesSnafu, InvalidAsciiSnafu,
+    ServerSnafu,
 };
 use crate::{from_grpc_response, Client, Result};
 
@@ -91,7 +92,7 @@ impl Database {
     ///
     /// - the name of database when using GreptimeDB standalone or cluster
     /// - the name provided by GreptimeCloud or other multi-tenant GreptimeDB
-    /// environment
+    ///   environment
     pub fn new_with_dbname(dbname: impl Into<String>, client: Client) -> Self {
         Self {
             catalog: String::default(),
@@ -225,16 +226,18 @@ impl Database {
 
         let mut client = self.client.make_flight_client()?;
 
-        let response = client.mut_inner().do_get(request).await.map_err(|e| {
+        let response = client.mut_inner().do_get(request).await.or_else(|e| {
             let tonic_code = e.code();
             let e: Error = e.into();
             let code = e.status_code();
             let msg = e.to_string();
-            let error = Error::FlightGet {
-                tonic_code,
-                addr: client.addr().to_string(),
-                source: BoxedError::new(ServerSnafu { code, msg }.build()),
-            };
+            let error =
+                Err(BoxedError::new(ServerSnafu { code, msg }.build())).with_context(|_| {
+                    FlightGetSnafu {
+                        addr: client.addr().to_string(),
+                        tonic_code,
+                    }
+                });
             error!(
                 "Failed to do Flight get, addr: {}, code: {}, source: {:?}",
                 client.addr(),

src/client/src/error.rs

@@ -39,13 +39,6 @@ pub enum Error {
         source: BoxedError,
     },
 
-    #[snafu(display("Failure occurs during handling request"))]
-    HandleRequest {
-        #[snafu(implicit)]
-        location: Location,
-        source: BoxedError,
-    },
-
     #[snafu(display("Failed to convert FlightData"))]
     ConvertFlightData {
         #[snafu(implicit)]
@@ -116,13 +109,6 @@ pub enum Error {
         location: Location,
     },
 
-    #[snafu(display("Failed to send request with streaming: {}", err_msg))]
-    ClientStreaming {
-        err_msg: String,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
     #[snafu(display("Failed to parse ascii string: {}", value))]
     InvalidAscii {
         value: String,
@@ -138,12 +124,10 @@ impl ErrorExt for Error {
         match self {
             Error::IllegalFlightMessages { .. }
             | Error::MissingField { .. }
-            | Error::IllegalDatabaseResponse { .. }
-            | Error::ClientStreaming { .. } => StatusCode::Internal,
+            | Error::IllegalDatabaseResponse { .. } => StatusCode::Internal,
 
             Error::Server { code, .. } => *code,
             Error::FlightGet { source, .. }
-            | Error::HandleRequest { source, .. }
             | Error::RegionServer { source, .. }
             | Error::FlowServer { source, .. } => source.status_code(),
             Error::CreateChannel { source, .. }

src/client/src/flow.rs

@@ -16,9 +16,9 @@ use api::v1::flow::{FlowRequest, FlowResponse};
 use api::v1::region::InsertRequests;
 use common_error::ext::BoxedError;
 use common_meta::node_manager::Flownode;
-use snafu::{location, Location, ResultExt};
+use snafu::ResultExt;
 
-use crate::error::Result;
+use crate::error::{FlowServerSnafu, Result};
 use crate::Client;
 
 #[derive(Debug)]
@@ -57,15 +57,10 @@ impl FlowRequester {
         let response = client
             .handle_create_remove(request)
             .await
-            .map_err(|e| {
+            .or_else(|e| {
                 let code = e.code();
                 let err: crate::error::Error = e.into();
-                crate::error::Error::FlowServer {
-                    addr,
-                    code,
-                    source: BoxedError::new(err),
-                    location: location!(),
-                }
+                Err(BoxedError::new(err)).context(FlowServerSnafu { addr, code })
             })?
             .into_inner();
         Ok(response)
@@ -88,15 +83,10 @@ impl FlowRequester {
         let response = client
             .handle_mirror_request(requests)
             .await
-            .map_err(|e| {
+            .or_else(|e| {
                 let code = e.code();
                 let err: crate::error::Error = e.into();
-                crate::error::Error::FlowServer {
-                    addr,
-                    code,
-                    source: BoxedError::new(err),
-                    location: location!(),
-                }
+                Err(BoxedError::new(err)).context(FlowServerSnafu { addr, code })
             })?
             .into_inner();
         Ok(response)

src/client/src/region.rs

@@ -33,13 +33,13 @@ use common_telemetry::error;
 use common_telemetry::tracing_context::TracingContext;
 use prost::Message;
 use query::query_engine::DefaultSerializer;
-use snafu::{location, Location, OptionExt, ResultExt};
+use snafu::{location, OptionExt, ResultExt};
 use substrait::{DFLogicalSubstraitConvertor, SubstraitPlan};
 use tokio_stream::StreamExt;
 
 use crate::error::{
-    self, ConvertFlightDataSnafu, IllegalDatabaseResponseSnafu, IllegalFlightMessagesSnafu,
-    MissingFieldSnafu, Result, ServerSnafu,
+    self, ConvertFlightDataSnafu, FlightGetSnafu, IllegalDatabaseResponseSnafu,
+    IllegalFlightMessagesSnafu, MissingFieldSnafu, Result, ServerSnafu,
 };
 use crate::{metrics, Client, Error};
 
@@ -103,11 +103,14 @@ impl RegionRequester {
                 let e: error::Error = e.into();
                 let code = e.status_code();
                 let msg = e.to_string();
-                let error = Error::FlightGet {
-                    tonic_code,
-                    addr: flight_client.addr().to_string(),
-                    source: BoxedError::new(ServerSnafu { code, msg }.build()),
-                };
+                let error = ServerSnafu { code, msg }
+                    .fail::<()>()
+                    .map_err(BoxedError::new)
+                    .with_context(|_| FlightGetSnafu {
+                        tonic_code,
+                        addr: flight_client.addr().to_string(),
+                    })
+                    .unwrap_err();
                 error!(
                     e; "Failed to do Flight get, addr: {}, code: {}",
                     flight_client.addr(),

src/cmd/Cargo.toml

@@ -51,7 +51,7 @@ file-engine.workspace = true
 flow.workspace = true
 frontend = { workspace = true, default-features = false }
 futures.workspace = true
-human-panic = "1.2.2"
+human-panic = "2.0"
 lazy_static.workspace = true
 meta-client.workspace = true
 meta-srv.workspace = true
@@ -80,7 +80,7 @@ tonic.workspace = true
 tracing-appender = "0.2"
 
 [target.'cfg(not(windows))'.dependencies]
-tikv-jemallocator = "0.5"
+tikv-jemallocator = "0.6"
 
 [dev-dependencies]
 client = { workspace = true, features = ["testing"] }

src/cmd/src/bin/greptime.rs

@@ -62,8 +62,37 @@ enum SubCommand {
 #[global_allocator]
 static ALLOC: tikv_jemallocator::Jemalloc = tikv_jemallocator::Jemalloc;
 
+#[cfg(debug_assertions)]
+fn main() -> Result<()> {
+    use snafu::ResultExt;
+    // Set the stack size to 8MB for the thread so it wouldn't overflow on large stack usage in debug mode
+    // see https://github.com/GreptimeTeam/greptimedb/pull/4317
+    // and https://github.com/rust-lang/rust/issues/34283
+    std::thread::Builder::new()
+        .name("main_spawn".to_string())
+        .stack_size(8 * 1024 * 1024)
+        .spawn(|| {
+            {
+                tokio::runtime::Builder::new_multi_thread()
+                    .thread_stack_size(8 * 1024 * 1024)
+                    .enable_all()
+                    .build()
+                    .expect("Failed building the Runtime")
+                    .block_on(main_body())
+            }
+        })
+        .context(cmd::error::SpawnThreadSnafu)?
+        .join()
+        .expect("Couldn't join on the associated thread")
+}
+
+#[cfg(not(debug_assertions))]
 #[tokio::main]
 async fn main() -> Result<()> {
+    main_body().await
+}
+
+async fn main_body() -> Result<()> {
     setup_human_panic();
     start(Command::parse()).await
 }
@@ -110,13 +139,10 @@ async fn start(cli: Command) -> Result<()> {
 }
 
 fn setup_human_panic() {
-    let metadata = human_panic::Metadata {
-        version: env!("CARGO_PKG_VERSION").into(),
-        name: "GreptimeDB".into(),
-        authors: Default::default(),
-        homepage: "https://github.com/GreptimeTeam/greptimedb/discussions".into(),
-    };
-    human_panic::setup_panic!(metadata);
+    human_panic::setup_panic!(
+        human_panic::Metadata::new("GreptimeDB", env!("CARGO_PKG_VERSION"))
+            .homepage("https://github.com/GreptimeTeam/greptimedb/discussions")
+    );
 
     common_telemetry::set_panic_hook();
 }

src/cmd/src/cli.rs

@@ -21,6 +21,8 @@ mod export;
 mod helper;
 
 // Wait for https://github.com/GreptimeTeam/greptimedb/issues/2373
+mod database;
+mod import;
 #[allow(unused)]
 mod repl;
 
@@ -32,6 +34,7 @@ pub use repl::Repl;
 use tracing_appender::non_blocking::WorkerGuard;
 
 use self::export::ExportCommand;
+use crate::cli::import::ImportCommand;
 use crate::error::Result;
 use crate::options::GlobalOptions;
 use crate::App;
@@ -114,6 +117,7 @@ enum SubCommand {
     // Attach(AttachCommand),
     Bench(BenchTableMetadataCommand),
     Export(ExportCommand),
+    Import(ImportCommand),
 }
 
 impl SubCommand {
@@ -122,6 +126,7 @@ impl SubCommand {
             // SubCommand::Attach(cmd) => cmd.build().await,
             SubCommand::Bench(cmd) => cmd.build(guard).await,
             SubCommand::Export(cmd) => cmd.build(guard).await,
+            SubCommand::Import(cmd) => cmd.build(guard).await,
         }
     }
 }

src/cmd/src/cli/database.rs

@@ -0,0 +1,119 @@
+// 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 base64::engine::general_purpose;
+use base64::Engine;
+use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
+use serde_json::Value;
+use servers::http::greptime_result_v1::GreptimedbV1Response;
+use servers::http::GreptimeQueryOutput;
+use snafu::ResultExt;
+
+use crate::error::{HttpQuerySqlSnafu, Result, SerdeJsonSnafu};
+
+pub(crate) struct DatabaseClient {
+    addr: String,
+    catalog: String,
+    auth_header: Option<String>,
+}
+
+impl DatabaseClient {
+    pub fn new(addr: String, catalog: String, auth_basic: Option<String>) -> Self {
+        let auth_header = if let Some(basic) = auth_basic {
+            let encoded = general_purpose::STANDARD.encode(basic);
+            Some(format!("basic {}", encoded))
+        } else {
+            None
+        };
+
+        Self {
+            addr,
+            catalog,
+            auth_header,
+        }
+    }
+
+    pub async fn sql_in_public(&self, sql: &str) -> Result<Option<Vec<Vec<Value>>>> {
+        self.sql(sql, DEFAULT_SCHEMA_NAME).await
+    }
+
+    /// Execute sql query.
+    pub async fn sql(&self, sql: &str, schema: &str) -> Result<Option<Vec<Vec<Value>>>> {
+        let url = format!("http://{}/v1/sql", self.addr);
+        let params = [
+            ("db", format!("{}-{}", self.catalog, schema)),
+            ("sql", sql.to_string()),
+        ];
+        let mut request = reqwest::Client::new()
+            .post(&url)
+            .form(&params)
+            .header("Content-Type", "application/x-www-form-urlencoded");
+        if let Some(ref auth) = self.auth_header {
+            request = request.header("Authorization", auth);
+        }
+
+        let response = request.send().await.with_context(|_| HttpQuerySqlSnafu {
+            reason: format!("bad url: {}", url),
+        })?;
+        let response = response
+            .error_for_status()
+            .with_context(|_| HttpQuerySqlSnafu {
+                reason: format!("query failed: {}", sql),
+            })?;
+
+        let text = response.text().await.with_context(|_| HttpQuerySqlSnafu {
+            reason: "cannot get response text".to_string(),
+        })?;
+
+        let body = serde_json::from_str::<GreptimedbV1Response>(&text).context(SerdeJsonSnafu)?;
+        Ok(body.output().first().and_then(|output| match output {
+            GreptimeQueryOutput::Records(records) => Some(records.rows().clone()),
+            GreptimeQueryOutput::AffectedRows(_) => None,
+        }))
+    }
+}
+
+/// Split at `-`.
+pub(crate) fn split_database(database: &str) -> Result<(String, Option<String>)> {
+    let (catalog, schema) = match database.split_once('-') {
+        Some((catalog, schema)) => (catalog, schema),
+        None => (DEFAULT_CATALOG_NAME, database),
+    };
+
+    if schema == "*" {
+        Ok((catalog.to_string(), None))
+    } else {
+        Ok((catalog.to_string(), Some(schema.to_string())))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_split_database() {
+        let result = split_database("catalog-schema").unwrap();
+        assert_eq!(result, ("catalog".to_string(), Some("schema".to_string())));
+
+        let result = split_database("schema").unwrap();
+        assert_eq!(result, ("greptime".to_string(), Some("schema".to_string())));
+
+        let result = split_database("catalog-*").unwrap();
+        assert_eq!(result, ("catalog".to_string(), None));
+
+        let result = split_database("*").unwrap();
+        assert_eq!(result, ("greptime".to_string(), None));
+    }
+}

src/cmd/src/cli/export.rs

@@ -13,30 +13,23 @@
 // limitations under the License.
 
 use std::collections::HashSet;
-use std::path::Path;
+use std::path::PathBuf;
 use std::sync::Arc;
 
 use async_trait::async_trait;
-use base64::engine::general_purpose;
-use base64::Engine;
 use clap::{Parser, ValueEnum};
-use client::DEFAULT_SCHEMA_NAME;
-use common_catalog::consts::DEFAULT_CATALOG_NAME;
 use common_telemetry::{debug, error, info};
 use serde_json::Value;
-use servers::http::greptime_result_v1::GreptimedbV1Response;
-use servers::http::GreptimeQueryOutput;
-use snafu::ResultExt;
+use snafu::{OptionExt, ResultExt};
 use tokio::fs::File;
 use tokio::io::{AsyncWriteExt, BufWriter};
 use tokio::sync::Semaphore;
 use tokio::time::Instant;
 use tracing_appender::non_blocking::WorkerGuard;
 
-use crate::cli::{Instance, Tool};
-use crate::error::{
-    EmptyResultSnafu, Error, FileIoSnafu, HttpQuerySqlSnafu, Result, SerdeJsonSnafu,
-};
+use crate::cli::database::DatabaseClient;
+use crate::cli::{database, Instance, Tool};
+use crate::error::{EmptyResultSnafu, Error, FileIoSnafu, Result, SchemaNotFoundSnafu};
 
 type TableReference = (String, String, String);
 
@@ -94,26 +87,21 @@ pub struct ExportCommand {
 
 impl ExportCommand {
     pub async fn build(&self, guard: Vec<WorkerGuard>) -> Result<Instance> {
-        let (catalog, schema) = split_database(&self.database)?;
+        let (catalog, schema) = database::split_database(&self.database)?;
 
-        let auth_header = if let Some(basic) = &self.auth_basic {
-            let encoded = general_purpose::STANDARD.encode(basic);
-            Some(format!("basic {}", encoded))
-        } else {
-            None
-        };
+        let database_client =
+            DatabaseClient::new(self.addr.clone(), catalog.clone(), self.auth_basic.clone());
 
         Ok(Instance::new(
             Box::new(Export {
-                addr: self.addr.clone(),
                 catalog,
                 schema,
+                database_client,
                 output_dir: self.output_dir.clone(),
                 parallelism: self.export_jobs,
                 target: self.target.clone(),
                 start_time: self.start_time.clone(),
                 end_time: self.end_time.clone(),
-                auth_header,
             }),
             guard,
         ))
@@ -121,78 +109,59 @@ impl ExportCommand {
 }
 
 pub struct Export {
-    addr: String,
     catalog: String,
     schema: Option<String>,
+    database_client: DatabaseClient,
     output_dir: String,
     parallelism: usize,
     target: ExportTarget,
     start_time: Option<String>,
     end_time: Option<String>,
-    auth_header: Option<String>,
 }
 
 impl Export {
-    /// Execute one single sql query.
-    async fn sql(&self, sql: &str) -> Result<Option<Vec<Vec<Value>>>> {
-        let url = format!(
-            "http://{}/v1/sql?db={}-{}&sql={}",
-            self.addr,
-            self.catalog,
-            self.schema.as_deref().unwrap_or(DEFAULT_SCHEMA_NAME),
-            sql
-        );
+    fn catalog_path(&self) -> PathBuf {
+        PathBuf::from(&self.output_dir).join(&self.catalog)
+    }
 
-        let mut request = reqwest::Client::new()
-            .get(&url)
-            .header("Content-Type", "application/x-www-form-urlencoded");
-        if let Some(ref auth) = self.auth_header {
-            request = request.header("Authorization", auth);
-        }
+    async fn get_db_names(&self) -> Result<Vec<String>> {
+        let db_names = self.all_db_names().await?;
+        let Some(schema) = &self.schema else {
+            return Ok(db_names);
+        };
 
-        let response = request.send().await.with_context(|_| HttpQuerySqlSnafu {
-            reason: format!("bad url: {}", url),
-        })?;
-        let response = response
-            .error_for_status()
-            .with_context(|_| HttpQuerySqlSnafu {
-                reason: format!("query failed: {}", sql),
-            })?;
-
-        let text = response.text().await.with_context(|_| HttpQuerySqlSnafu {
-            reason: "cannot get response text".to_string(),
-        })?;
-
-        let body = serde_json::from_str::<GreptimedbV1Response>(&text).context(SerdeJsonSnafu)?;
-        Ok(body.output().first().and_then(|output| match output {
-            GreptimeQueryOutput::Records(records) => Some(records.rows().clone()),
-            GreptimeQueryOutput::AffectedRows(_) => None,
-        }))
+        // Check if the schema exists
+        db_names
+            .into_iter()
+            .find(|db_name| db_name.to_lowercase() == schema.to_lowercase())
+            .map(|name| vec![name])
+            .context(SchemaNotFoundSnafu {
+                catalog: &self.catalog,
+                schema,
+            })
     }
 
     /// Iterate over all db names.
-    ///
-    /// Newbie: `db_name` is catalog + schema.
-    async fn iter_db_names(&self) -> Result<Vec<(String, String)>> {
-        if let Some(schema) = &self.schema {
-            Ok(vec![(self.catalog.clone(), schema.clone())])
-        } else {
-            let result = self.sql("SHOW DATABASES").await?;
-            let Some(records) = result else {
-                EmptyResultSnafu.fail()?
+    async fn all_db_names(&self) -> Result<Vec<String>> {
+        let records = self
+            .database_client
+            .sql_in_public("SHOW DATABASES")
+            .await?
+            .context(EmptyResultSnafu)?;
+        let mut result = Vec::with_capacity(records.len());
+        for value in records {
+            let Value::String(schema) = &value[0] else {
+                unreachable!()
             };
-            let mut result = Vec::with_capacity(records.len());
-            for value in records {
-                let Value::String(schema) = &value[0] else {
-                    unreachable!()
-                };
-                if schema == common_catalog::consts::INFORMATION_SCHEMA_NAME {
-                    continue;
-                }
-                result.push((self.catalog.clone(), schema.clone()));
+            if schema == common_catalog::consts::INFORMATION_SCHEMA_NAME {
+                continue;
             }
-            Ok(result)
+            if schema == common_catalog::consts::PG_CATALOG_NAME {
+                continue;
+            }
+            result.push(schema.clone());
         }
+        Ok(result)
     }
 
     /// Return a list of [`TableReference`] to be exported.
@@ -201,7 +170,11 @@ impl Export {
         &self,
         catalog: &str,
         schema: &str,
-    ) -> Result<(Vec<TableReference>, Vec<TableReference>)> {
+    ) -> Result<(
+        Vec<TableReference>,
+        Vec<TableReference>,
+        Vec<TableReference>,
+    )> {
         // Puts all metric table first
         let sql = format!(
             "SELECT table_catalog, table_schema, table_name \
@@ -210,15 +183,16 @@ impl Export {
                 and table_catalog = \'{catalog}\' \
                 and table_schema = \'{schema}\'"
         );
-        let result = self.sql(&sql).await?;
-        let Some(records) = result else {
-            EmptyResultSnafu.fail()?
-        };
+        let records = self
+            .database_client
+            .sql_in_public(&sql)
+            .await?
+            .context(EmptyResultSnafu)?;
         let mut metric_physical_tables = HashSet::with_capacity(records.len());
         for value in records {
             let mut t = Vec::with_capacity(3);
             for v in &value {
-                let serde_json::Value::String(value) = v else {
+                let Value::String(value) = v else {
                     unreachable!()
                 };
                 t.push(value);
@@ -226,100 +200,142 @@ impl Export {
             metric_physical_tables.insert((t[0].clone(), t[1].clone(), t[2].clone()));
         }
 
-        // TODO: SQL injection hurts
         let sql = format!(
-            "SELECT table_catalog, table_schema, table_name \
+            "SELECT table_catalog, table_schema, table_name, table_type \
             FROM information_schema.tables \
-            WHERE table_type = \'BASE TABLE\' \
+            WHERE (table_type = \'BASE TABLE\' OR table_type = \'VIEW\') \
                 and table_catalog = \'{catalog}\' \
                 and table_schema = \'{schema}\'",
         );
-        let result = self.sql(&sql).await?;
-        let Some(records) = result else {
-            EmptyResultSnafu.fail()?
-        };
+        let records = self
+            .database_client
+            .sql_in_public(&sql)
+            .await?
+            .context(EmptyResultSnafu)?;
 
-        debug!("Fetched table list: {:?}", records);
+        debug!("Fetched table/view list: {:?}", records);
 
         if records.is_empty() {
-            return Ok((vec![], vec![]));
+            return Ok((vec![], vec![], vec![]));
         }
 
         let mut remaining_tables = Vec::with_capacity(records.len());
+        let mut views = Vec::new();
         for value in records {
-            let mut t = Vec::with_capacity(3);
+            let mut t = Vec::with_capacity(4);
             for v in &value {
-                let serde_json::Value::String(value) = v else {
+                let Value::String(value) = v else {
                     unreachable!()
                 };
                 t.push(value);
             }
             let table = (t[0].clone(), t[1].clone(), t[2].clone());
+            let table_type = t[3].as_str();
             // Ignores the physical table
             if !metric_physical_tables.contains(&table) {
-                remaining_tables.push(table);
+                if table_type == "VIEW" {
+                    views.push(table);
+                } else {
+                    remaining_tables.push(table);
+                }
             }
         }
 
         Ok((
             metric_physical_tables.into_iter().collect(),
             remaining_tables,
+            views,
         ))
     }
 
-    async fn show_create_table(&self, catalog: &str, schema: &str, table: &str) -> Result<String> {
-        let sql = format!(
-            r#"SHOW CREATE TABLE "{}"."{}"."{}""#,
-            catalog, schema, table
-        );
-        let result = self.sql(&sql).await?;
-        let Some(records) = result else {
-            EmptyResultSnafu.fail()?
+    async fn show_create(
+        &self,
+        show_type: &str,
+        catalog: &str,
+        schema: &str,
+        table: Option<&str>,
+    ) -> Result<String> {
+        let sql = match table {
+            Some(table) => format!(
+                r#"SHOW CREATE {} "{}"."{}"."{}""#,
+                show_type, catalog, schema, table
+            ),
+            None => format!(r#"SHOW CREATE {} "{}"."{}""#, show_type, catalog, schema),
         };
-        let Value::String(create_table) = &records[0][1] else {
+        let records = self
+            .database_client
+            .sql_in_public(&sql)
+            .await?
+            .context(EmptyResultSnafu)?;
+        let Value::String(create) = &records[0][1] else {
             unreachable!()
         };
 
-        Ok(format!("{};\n", create_table))
+        Ok(format!("{};\n", create))
+    }
+
+    async fn export_create_database(&self) -> Result<()> {
+        let timer = Instant::now();
+        let db_names = self.get_db_names().await?;
+        let db_count = db_names.len();
+        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?;
+            file.write_all(create_database.as_bytes())
+                .await
+                .context(FileIoSnafu)?;
+        }
+
+        let elapsed = timer.elapsed();
+        info!("Success {db_count} jobs, cost: {elapsed:?}");
+
+        Ok(())
     }
 
     async fn export_create_table(&self) -> Result<()> {
         let timer = Instant::now();
         let semaphore = Arc::new(Semaphore::new(self.parallelism));
-        let db_names = self.iter_db_names().await?;
+        let db_names = self.get_db_names().await?;
         let db_count = db_names.len();
         let mut tasks = Vec::with_capacity(db_names.len());
-        for (catalog, schema) in db_names {
+        for schema in db_names {
             let semaphore_moved = semaphore.clone();
             tasks.push(async move {
                 let _permit = semaphore_moved.acquire().await.unwrap();
-                let (metric_physical_tables, remaining_tables) =
-                    self.get_table_list(&catalog, &schema).await?;
-                let table_count = metric_physical_tables.len() + remaining_tables.len();
-                let output_dir = Path::new(&self.output_dir)
-                    .join(&catalog)
-                    .join(format!("{schema}/"));
-                tokio::fs::create_dir_all(&output_dir)
+                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 output_file = Path::new(&output_dir).join("create_tables.sql");
-                let mut file = File::create(output_file).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) {
-                    match self.show_create_table(&c, &s, &t).await {
-                        Err(e) => {
-                            error!(e; r#"Failed to export table "{}"."{}"."{}""#, c, s, t)
-                        }
-                        Ok(create_table) => {
-                            file.write_all(create_table.as_bytes())
-                                .await
-                                .context(FileIoSnafu)?;
-                        }
-                    }
+                    let create_table = self.show_create("TABLE", &c, &s, Some(&t)).await?;
+                    file.write_all(create_table.as_bytes())
+                        .await
+                        .context(FileIoSnafu)?;
+                }
+                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)?;
                 }
 
                 info!(
-                    "Finished exporting {catalog}.{schema} with {table_count} table schemas to path: {}",
-                    output_dir.to_string_lossy()
+                    "Finished exporting {}.{schema} with {table_count} table schemas to path: {}",
+                    self.catalog,
+                    db_dir.to_string_lossy()
                 );
 
                 Ok::<(), Error>(())
@@ -332,14 +348,14 @@ impl Export {
             .filter(|r| match r {
                 Ok(_) => true,
                 Err(e) => {
-                    error!(e; "export job failed");
+                    error!(e; "export schema job failed");
                     false
                 }
             })
             .count();
 
         let elapsed = timer.elapsed();
-        info!("Success {success}/{db_count} jobs, cost: {:?}", elapsed);
+        info!("Success {success}/{db_count} jobs, cost: {elapsed:?}");
 
         Ok(())
     }
@@ -347,17 +363,15 @@ impl Export {
     async fn export_database_data(&self) -> Result<()> {
         let timer = Instant::now();
         let semaphore = Arc::new(Semaphore::new(self.parallelism));
-        let db_names = self.iter_db_names().await?;
+        let db_names = self.get_db_names().await?;
         let db_count = db_names.len();
-        let mut tasks = Vec::with_capacity(db_names.len());
-        for (catalog, schema) in db_names {
+        let mut tasks = Vec::with_capacity(db_count);
+        for schema in db_names {
             let semaphore_moved = semaphore.clone();
             tasks.push(async move {
                 let _permit = semaphore_moved.acquire().await.unwrap();
-                let output_dir = Path::new(&self.output_dir)
-                    .join(&catalog)
-                    .join(format!("{schema}/"));
-                tokio::fs::create_dir_all(&output_dir)
+                let db_dir = self.catalog_path().join(format!("{schema}/"));
+                tokio::fs::create_dir_all(&db_dir)
                     .await
                     .context(FileIoSnafu)?;
 
@@ -379,30 +393,31 @@ impl Export {
 
                 let sql = format!(
                     r#"COPY DATABASE "{}"."{}" TO '{}' {};"#,
-                    catalog,
+                    self.catalog,
                     schema,
-                    output_dir.to_str().unwrap(),
+                    db_dir.to_str().unwrap(),
                     with_options
                 );
 
                 info!("Executing sql: {sql}");
 
-                self.sql(&sql).await?;
+                self.database_client.sql_in_public(&sql).await?;
 
                 info!(
-                    "Finished exporting {catalog}.{schema} data into path: {}",
-                    output_dir.to_string_lossy()
+                    "Finished exporting {}.{schema} data into path: {}",
+                    self.catalog,
+                    db_dir.to_string_lossy()
                 );
 
                 // The export copy from sql
-                let copy_from_file = output_dir.join("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');"#,
-                    catalog,
+                    self.catalog,
                     schema,
-                    output_dir.to_str().unwrap()
+                    db_dir.to_str().unwrap()
                 );
                 writer
                     .write(copy_database_from_sql.as_bytes())
@@ -410,7 +425,7 @@ impl Export {
                     .context(FileIoSnafu)?;
                 writer.flush().await.context(FileIoSnafu)?;
 
-                info!("Finished exporting {catalog}.{schema} copy_from.sql");
+                info!("Finished exporting {}.{schema} copy_from.sql", self.catalog);
 
                 Ok::<(), Error>(())
             })
@@ -429,20 +444,23 @@ impl Export {
             .count();
         let elapsed = timer.elapsed();
 
-        info!("Success {success}/{db_count} jobs, costs: {:?}", elapsed);
+        info!("Success {success}/{db_count} jobs, costs: {elapsed:?}");
 
         Ok(())
     }
 }
 
-#[allow(deprecated)]
 #[async_trait]
 impl Tool for Export {
     async fn do_work(&self) -> Result<()> {
         match self.target {
-            ExportTarget::Schema => self.export_create_table().await,
+            ExportTarget::Schema => {
+                self.export_create_database().await?;
+                self.export_create_table().await
+            }
             ExportTarget::Data => self.export_database_data().await,
             ExportTarget::All => {
+                self.export_create_database().await?;
                 self.export_create_table().await?;
                 self.export_database_data().await
             }
@@ -450,20 +468,6 @@ impl Tool for Export {
     }
 }
 
-/// Split at `-`.
-fn split_database(database: &str) -> Result<(String, Option<String>)> {
-    let (catalog, schema) = match database.split_once('-') {
-        Some((catalog, schema)) => (catalog, schema),
-        None => (DEFAULT_CATALOG_NAME, database),
-    };
-
-    if schema == "*" {
-        Ok((catalog.to_string(), None))
-    } else {
-        Ok((catalog.to_string(), Some(schema.to_string())))
-    }
-}
-
 #[cfg(test)]
 mod tests {
     use clap::Parser;
@@ -471,26 +475,10 @@ mod tests {
     use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
     use common_telemetry::logging::LoggingOptions;
 
-    use crate::cli::export::split_database;
     use crate::error::Result as CmdResult;
     use crate::options::GlobalOptions;
     use crate::{cli, standalone, App};
 
-    #[test]
-    fn test_split_database() {
-        let result = split_database("catalog-schema").unwrap();
-        assert_eq!(result, ("catalog".to_string(), Some("schema".to_string())));
-
-        let result = split_database("schema").unwrap();
-        assert_eq!(result, ("greptime".to_string(), Some("schema".to_string())));
-
-        let result = split_database("catalog-*").unwrap();
-        assert_eq!(result, ("catalog".to_string(), None));
-
-        let result = split_database("*").unwrap();
-        assert_eq!(result, ("greptime".to_string(), None));
-    }
-
     #[tokio::test(flavor = "multi_thread")]
     async fn test_export_create_table_with_quoted_names() -> CmdResult<()> {
         let output_dir = tempfile::tempdir().unwrap();

src/cmd/src/cli/import.rs

@@ -0,0 +1,218 @@
+// 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::path::PathBuf;
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use clap::{Parser, ValueEnum};
+use common_catalog::consts::DEFAULT_SCHEMA_NAME;
+use common_telemetry::{error, info, warn};
+use snafu::{OptionExt, ResultExt};
+use tokio::sync::Semaphore;
+use tokio::time::Instant;
+use tracing_appender::non_blocking::WorkerGuard;
+
+use crate::cli::database::DatabaseClient;
+use crate::cli::{database, Instance, Tool};
+use crate::error::{Error, FileIoSnafu, Result, SchemaNotFoundSnafu};
+
+#[derive(Debug, Default, Clone, ValueEnum)]
+enum ImportTarget {
+    /// Import all table schemas into the database.
+    Schema,
+    /// Import all table data into the database.
+    Data,
+    /// Export all table schemas and data at once.
+    #[default]
+    All,
+}
+
+#[derive(Debug, Default, Parser)]
+pub struct ImportCommand {
+    /// Server address to connect
+    #[clap(long)]
+    addr: String,
+
+    /// Directory of the data. E.g.: /tmp/greptimedb-backup
+    #[clap(long)]
+    input_dir: String,
+
+    /// The name of the catalog to import.
+    #[clap(long, default_value = "greptime-*")]
+    database: String,
+
+    /// Parallelism of the import.
+    #[clap(long, short = 'j', default_value = "1")]
+    import_jobs: usize,
+
+    /// Max retry times for each job.
+    #[clap(long, default_value = "3")]
+    max_retry: usize,
+
+    /// Things to export
+    #[clap(long, short = 't', value_enum, default_value = "all")]
+    target: ImportTarget,
+
+    /// The basic authentication for connecting to the server
+    #[clap(long)]
+    auth_basic: Option<String>,
+}
+
+impl ImportCommand {
+    pub async fn build(&self, guard: Vec<WorkerGuard>) -> Result<Instance> {
+        let (catalog, schema) = database::split_database(&self.database)?;
+        let database_client =
+            DatabaseClient::new(self.addr.clone(), catalog.clone(), self.auth_basic.clone());
+
+        Ok(Instance::new(
+            Box::new(Import {
+                catalog,
+                schema,
+                database_client,
+                input_dir: self.input_dir.clone(),
+                parallelism: self.import_jobs,
+                target: self.target.clone(),
+            }),
+            guard,
+        ))
+    }
+}
+
+pub struct Import {
+    catalog: String,
+    schema: Option<String>,
+    database_client: DatabaseClient,
+    input_dir: String,
+    parallelism: usize,
+    target: ImportTarget,
+}
+
+impl Import {
+    async fn import_create_table(&self) -> Result<()> {
+        // Use default db to creates other dbs
+        self.do_sql_job("create_database.sql", Some(DEFAULT_SCHEMA_NAME))
+            .await?;
+        self.do_sql_job("create_tables.sql", None).await
+    }
+
+    async fn import_database_data(&self) -> Result<()> {
+        self.do_sql_job("copy_from.sql", None).await
+    }
+
+    async fn do_sql_job(&self, filename: &str, exec_db: Option<&str>) -> 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);
+        for schema in db_names {
+            let semaphore_moved = semaphore.clone();
+            tasks.push(async move {
+                let _permit = semaphore_moved.acquire().await.unwrap();
+                let database_input_dir = self.catalog_path().join(&schema);
+                let sql_file = database_input_dir.join(filename);
+                let sql = tokio::fs::read_to_string(sql_file)
+                    .await
+                    .context(FileIoSnafu)?;
+                if sql.is_empty() {
+                    info!("Empty `{filename}` {database_input_dir:?}");
+                } else {
+                    let db = exec_db.unwrap_or(&schema);
+                    self.database_client.sql(&sql, db).await?;
+                    info!("Imported `{filename}` for database {schema}");
+                }
+
+                Ok::<(), Error>(())
+            })
+        }
+
+        let success = futures::future::join_all(tasks)
+            .await
+            .into_iter()
+            .filter(|r| match r {
+                Ok(_) => true,
+                Err(e) => {
+                    error!(e; "import {filename} job failed");
+                    false
+                }
+            })
+            .count();
+        let elapsed = timer.elapsed();
+        info!("Success {success}/{db_count} `{filename}` jobs, cost: {elapsed:?}");
+
+        Ok(())
+    }
+
+    fn catalog_path(&self) -> PathBuf {
+        PathBuf::from(&self.input_dir).join(&self.catalog)
+    }
+
+    async fn get_db_names(&self) -> Result<Vec<String>> {
+        let db_names = self.all_db_names().await?;
+        let Some(schema) = &self.schema else {
+            return Ok(db_names);
+        };
+
+        // Check if the schema exists
+        db_names
+            .into_iter()
+            .find(|db_name| db_name.to_lowercase() == schema.to_lowercase())
+            .map(|name| vec![name])
+            .context(SchemaNotFoundSnafu {
+                catalog: &self.catalog,
+                schema,
+            })
+    }
+
+    // Get all database names in the input directory.
+    // The directory structure should be like:
+    // /tmp/greptimedb-backup
+    // ├── greptime-1
+    // │   ├── db1
+    // │   └── db2
+    async fn all_db_names(&self) -> Result<Vec<String>> {
+        let mut db_names = vec![];
+        let path = self.catalog_path();
+        let mut entries = tokio::fs::read_dir(path).await.context(FileIoSnafu)?;
+        while let Some(entry) = entries.next_entry().await.context(FileIoSnafu)? {
+            let path = entry.path();
+            if path.is_dir() {
+                let db_name = match path.file_name() {
+                    Some(name) => name.to_string_lossy().to_string(),
+                    None => {
+                        warn!("Failed to get the file name of {:?}", path);
+                        continue;
+                    }
+                };
+                db_names.push(db_name);
+            }
+        }
+        Ok(db_names)
+    }
+}
+
+#[async_trait]
+impl Tool for Import {
+    async fn do_work(&self) -> Result<()> {
+        match self.target {
+            ImportTarget::Schema => self.import_create_table().await,
+            ImportTarget::Data => self.import_database_data().await,
+            ImportTarget::All => {
+                self.import_create_table().await?;
+                self.import_database_data().await
+            }
+        }
+    }
+}

src/cmd/src/datanode.rs

@@ -267,7 +267,7 @@ impl StartCommand {
             &opts.component.tracing,
             opts.component.node_id.map(|x| x.to_string()),
         );
-        log_versions(version(), short_version());
+        log_versions(version(), short_version(), APP_NAME);
 
         info!("Datanode start command: {:#?}", self);
         info!("Datanode options: {:#?}", opts);

src/cmd/src/error.rs

@@ -31,13 +31,6 @@ pub enum Error {
         source: common_meta::error::Error,
     },
 
-    #[snafu(display("Failed to iter stream"))]
-    IterStream {
-        #[snafu(implicit)]
-        location: Location,
-        source: common_meta::error::Error,
-    },
-
     #[snafu(display("Failed to init DDL manager"))]
     InitDdlManager {
         #[snafu(implicit)]
@@ -237,13 +230,6 @@ pub enum Error {
         location: Location,
     },
 
-    #[snafu(display("Failed to start catalog manager"))]
-    StartCatalogManager {
-        #[snafu(implicit)]
-        location: Location,
-        source: catalog::error::Error,
-    },
-
     #[snafu(display("Failed to connect to Etcd at {etcd_addr}"))]
     ConnectEtcd {
         etcd_addr: String,
@@ -253,14 +239,6 @@ pub enum Error {
         location: Location,
     },
 
-    #[snafu(display("Failed to connect server at {addr}"))]
-    ConnectServer {
-        addr: String,
-        source: client::error::Error,
-        #[snafu(implicit)]
-        location: Location,
-    },
-
     #[snafu(display("Failed to serde json"))]
     SerdeJson {
         #[snafu(source)]
@@ -278,12 +256,6 @@ pub enum Error {
         location: Location,
     },
 
-    #[snafu(display("Expect data from output, but got another thing"))]
-    NotDataFromOutput {
-        #[snafu(implicit)]
-        location: Location,
-    },
-
     #[snafu(display("Empty result from output"))]
     EmptyResult {
         #[snafu(implicit)]
@@ -305,6 +277,12 @@ pub enum Error {
         error: std::io::Error,
     },
 
+    #[snafu(display("Failed to spawn thread"))]
+    SpawnThread {
+        #[snafu(source)]
+        error: std::io::Error,
+    },
+
     #[snafu(display("Other error"))]
     Other {
         source: BoxedError,
@@ -340,13 +318,12 @@ pub enum Error {
         source: meta_client::error::Error,
     },
 
-    #[snafu(display("Tonic transport error: {error:?} with msg: {msg:?}"))]
-    TonicTransport {
+    #[snafu(display("Cannot find schema {schema} in catalog {catalog}"))]
+    SchemaNotFound {
+        catalog: String,
+        schema: String,
         #[snafu(implicit)]
         location: Location,
-        #[snafu(source)]
-        error: tonic::transport::Error,
-        msg: Option<String>,
     },
 }
 
@@ -364,18 +341,16 @@ impl ErrorExt for Error {
             Error::BuildMetaServer { source, .. } => source.status_code(),
             Error::UnsupportedSelectorType { source, .. } => source.status_code(),
 
-            Error::IterStream { source, .. }
-            | Error::InitMetadata { source, .. }
-            | Error::InitDdlManager { source, .. } => source.status_code(),
+            Error::InitMetadata { source, .. } | Error::InitDdlManager { source, .. } => {
+                source.status_code()
+            }
 
-            Error::ConnectServer { source, .. } => source.status_code(),
             Error::MissingConfig { .. }
             | Error::LoadLayeredConfig { .. }
             | Error::IllegalConfig { .. }
             | Error::InvalidReplCommand { .. }
             | Error::InitTimezone { .. }
             | Error::ConnectEtcd { .. }
-            | Error::NotDataFromOutput { .. }
             | Error::CreateDir { .. }
             | Error::EmptyResult { .. } => StatusCode::InvalidArguments,
 
@@ -393,9 +368,10 @@ impl ErrorExt for Error {
                 source.status_code()
             }
             Error::SubstraitEncodeLogicalPlan { source, .. } => source.status_code(),
-            Error::StartCatalogManager { source, .. } => source.status_code(),
 
-            Error::SerdeJson { .. } | Error::FileIo { .. } => StatusCode::Unexpected,
+            Error::SerdeJson { .. } | Error::FileIo { .. } | Error::SpawnThread { .. } => {
+                StatusCode::Unexpected
+            }
 
             Error::Other { source, .. } => source.status_code(),
 
@@ -406,7 +382,7 @@ impl ErrorExt for Error {
                 source.status_code()
             }
             Error::MetaClientInit { source, .. } => source.status_code(),
-            Error::TonicTransport { .. } => StatusCode::Internal,
+            Error::SchemaNotFound { .. } => StatusCode::DatabaseNotFound,
         }
     }
 

src/cmd/src/flownode.rs

@@ -24,6 +24,7 @@ use common_grpc::channel_manager::ChannelConfig;
 use common_meta::cache::{CacheRegistryBuilder, LayeredCacheRegistryBuilder};
 use common_meta::heartbeat::handler::parse_mailbox_message::ParseMailboxMessageHandler;
 use common_meta::heartbeat::handler::HandlerGroupExecutor;
+use common_meta::key::flow::FlowMetadataManager;
 use common_meta::key::TableMetadataManager;
 use common_telemetry::info;
 use common_telemetry::logging::TracingOptions;
@@ -214,7 +215,7 @@ impl StartCommand {
             &opts.component.tracing,
             opts.component.node_id.map(|x| x.to_string()),
         );
-        log_versions(version(), short_version());
+        log_versions(version(), short_version(), APP_NAME);
 
         info!("Flownode start command: {:#?}", self);
         info!("Flownode options: {:#?}", opts);
@@ -296,11 +297,13 @@ impl StartCommand {
             Arc::new(executor),
         );
 
+        let flow_metadata_manager = Arc::new(FlowMetadataManager::new(cached_meta_backend.clone()));
         let flownode_builder = FlownodeBuilder::new(
             opts,
             Plugins::new(),
             table_metadata_manager,
             catalog_manager.clone(),
+            flow_metadata_manager,
         )
         .with_heartbeat_task(heartbeat_task);
 

src/cmd/src/frontend.rs

@@ -261,7 +261,7 @@ impl StartCommand {
             &opts.component.tracing,
             opts.component.node_id.clone(),
         );
-        log_versions(version(), short_version());
+        log_versions(version(), short_version(), APP_NAME);
 
         info!("Frontend start command: {:#?}", self);
         info!("Frontend options: {:#?}", opts);

src/cmd/src/lib.rs

@@ -30,7 +30,7 @@ pub mod standalone;
 
 lazy_static::lazy_static! {
     static ref APP_VERSION: prometheus::IntGaugeVec =
-        prometheus::register_int_gauge_vec!("greptime_app_version", "app version", &["short_version", "version"]).unwrap();
+        prometheus::register_int_gauge_vec!("greptime_app_version", "app version", &["version", "short_version", "app"]).unwrap();
 }
 
 #[async_trait]
@@ -74,16 +74,16 @@ pub trait App: Send {
 }
 
 /// Log the versions of the application, and the arguments passed to the cli.
-/// `version_string` should be the same as the output of cli "--version";
-/// and the `app_version` is the short version of the codes, often consist of git branch and commit.
-pub fn log_versions(version_string: &str, app_version: &str) {
+/// `version` should be the same as the output of cli "--version";
+/// and the `short_version` is the short version of the codes, often consist of git branch and commit.
+pub fn log_versions(version: &str, short_version: &str, app: &str) {
     // Report app version as gauge.
     APP_VERSION
-        .with_label_values(&[env!("CARGO_PKG_VERSION"), app_version])
+        .with_label_values(&[env!("CARGO_PKG_VERSION"), short_version, app])
         .inc();
 
     // Log version and argument flags.
-    info!("GreptimeDB version: {}", version_string);
+    info!("GreptimeDB version: {}", version);
 
     log_env_flags();
 }

src/cmd/src/metasrv.rs

@@ -22,6 +22,7 @@ use common_telemetry::info;
 use common_telemetry::logging::TracingOptions;
 use common_version::{short_version, version};
 use meta_srv::bootstrap::MetasrvInstance;
+use meta_srv::metasrv::BackendImpl;
 use snafu::ResultExt;
 use tracing_appender::non_blocking::WorkerGuard;
 
@@ -137,6 +138,9 @@ struct StartCommand {
     /// The max operations per txn
     #[clap(long)]
     max_txn_ops: Option<usize>,
+    /// The database backend.
+    #[clap(long, value_enum)]
+    backend: Option<BackendImpl>,
 }
 
 impl StartCommand {
@@ -219,6 +223,12 @@ impl StartCommand {
             opts.max_txn_ops = max_txn_ops;
         }
 
+        if let Some(backend) = &self.backend {
+            opts.backend.clone_from(backend);
+        } else {
+            opts.backend = BackendImpl::default()
+        }
+
         // Disable dashboard in metasrv.
         opts.http.disable_dashboard = true;
 
@@ -234,7 +244,7 @@ impl StartCommand {
             &opts.component.tracing,
             None,
         );
-        log_versions(version(), short_version());
+        log_versions(version(), short_version(), APP_NAME);
 
         info!("Metasrv start command: {:#?}", self);
         info!("Metasrv options: {:#?}", opts);

src/cmd/src/standalone.rs

@@ -141,6 +141,8 @@ pub struct StandaloneOptions {
     pub region_engine: Vec<RegionEngineConfig>,
     pub export_metrics: ExportMetricsOption,
     pub tracing: TracingOptions,
+    pub init_regions_in_background: bool,
+    pub init_regions_parallelism: usize,
 }
 
 impl Default for StandaloneOptions {
@@ -168,6 +170,8 @@ impl Default for StandaloneOptions {
                 RegionEngineConfig::File(FileEngineConfig::default()),
             ],
             tracing: TracingOptions::default(),
+            init_regions_in_background: false,
+            init_regions_parallelism: 16,
         }
     }
 }
@@ -178,6 +182,16 @@ impl Configurable for StandaloneOptions {
     }
 }
 
+/// The [`StandaloneOptions`] is only defined in cmd crate,
+/// we don't want to make `frontend` depends on it, so impl [`Into`]
+/// rather than [`From`].
+#[allow(clippy::from_over_into)]
+impl Into<FrontendOptions> for StandaloneOptions {
+    fn into(self) -> FrontendOptions {
+        self.frontend_options()
+    }
+}
+
 impl StandaloneOptions {
     pub fn frontend_options(&self) -> FrontendOptions {
         let cloned_opts = self.clone();
@@ -208,6 +222,9 @@ impl StandaloneOptions {
             storage: cloned_opts.storage,
             region_engine: cloned_opts.region_engine,
             grpc: cloned_opts.grpc,
+            init_regions_in_background: cloned_opts.init_regions_in_background,
+            init_regions_parallelism: cloned_opts.init_regions_parallelism,
+            mode: Mode::Standalone,
             ..Default::default()
         }
     }
@@ -415,7 +432,7 @@ impl StartCommand {
             &opts.component.tracing,
             None,
         );
-        log_versions(version(), short_version());
+        log_versions(version(), short_version(), APP_NAME);
 
         info!("Standalone start command: {:#?}", self);
         info!("Standalone options: {opts:#?}");
@@ -476,11 +493,13 @@ impl StartCommand {
             .await
             .context(StartDatanodeSnafu)?;
 
+        let flow_metadata_manager = Arc::new(FlowMetadataManager::new(kv_backend.clone()));
         let flow_builder = FlownodeBuilder::new(
             Default::default(),
             plugins.clone(),
             table_metadata_manager.clone(),
             catalog_manager.clone(),
+            flow_metadata_manager.clone(),
         );
         let flownode = Arc::new(
             flow_builder
@@ -508,10 +527,9 @@ impl StartCommand {
                 .build(),
         );
         let wal_options_allocator = Arc::new(WalOptionsAllocator::new(
-            opts.wal.into(),
+            opts.wal.clone().into(),
             kv_backend.clone(),
         ));
-        let flow_metadata_manager = Arc::new(FlowMetadataManager::new(kv_backend.clone()));
         let table_meta_allocator = Arc::new(TableMetadataAllocator::new(
             table_id_sequence,
             wal_options_allocator.clone(),
@@ -532,7 +550,7 @@ impl StartCommand {
         .await?;
 
         let mut frontend = FrontendBuilder::new(
-            fe_opts.clone(),
+            fe_opts,
             kv_backend.clone(),
             layered_cache_registry.clone(),
             catalog_manager.clone(),
@@ -560,7 +578,7 @@ impl StartCommand {
 
         let (tx, _rx) = broadcast::channel(1);
 
-        let servers = Services::new(fe_opts, Arc::new(frontend.clone()), plugins)
+        let servers = Services::new(opts, Arc::new(frontend.clone()), plugins)
             .build()
             .await
             .context(StartFrontendSnafu)?;

src/cmd/tests/load_config_test.rs

@@ -22,7 +22,7 @@ use common_grpc::channel_manager::{
     DEFAULT_MAX_GRPC_RECV_MESSAGE_SIZE, DEFAULT_MAX_GRPC_SEND_MESSAGE_SIZE,
 };
 use common_runtime::global::RuntimeOptions;
-use common_telemetry::logging::LoggingOptions;
+use common_telemetry::logging::{LoggingOptions, DEFAULT_OTLP_ENDPOINT};
 use common_wal::config::raft_engine::RaftEngineConfig;
 use common_wal::config::DatanodeWalConfig;
 use datanode::config::{DatanodeOptions, RegionEngineConfig, StorageConfig};
@@ -81,14 +81,16 @@ fn test_load_datanode_example_config() {
                     sst_meta_cache_size: ReadableSize::mb(128),
                     vector_cache_size: ReadableSize::mb(512),
                     page_cache_size: ReadableSize::mb(512),
+                    selector_result_cache_size: ReadableSize::mb(512),
                     max_background_jobs: 4,
+                    experimental_write_cache_ttl: Some(Duration::from_secs(60 * 60 * 8)),
                     ..Default::default()
                 }),
                 RegionEngineConfig::File(EngineConfig {}),
             ],
             logging: LoggingOptions {
                 level: Some("info".to_string()),
-                otlp_endpoint: Some("".to_string()),
+                otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
                 tracing_sample_ratio: Some(Default::default()),
                 ..Default::default()
             },
@@ -136,7 +138,7 @@ fn test_load_frontend_example_config() {
             }),
             logging: LoggingOptions {
                 level: Some("info".to_string()),
-                otlp_endpoint: Some("".to_string()),
+                otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
                 tracing_sample_ratio: Some(Default::default()),
                 ..Default::default()
             },
@@ -169,12 +171,12 @@ fn test_load_metasrv_example_config() {
             compact_rt_size: 4,
         },
         component: MetasrvOptions {
-            selector: SelectorType::LeaseBased,
+            selector: SelectorType::default(),
             data_home: "/tmp/metasrv/".to_string(),
             logging: LoggingOptions {
                 dir: "/tmp/greptimedb/logs".to_string(),
                 level: Some("info".to_string()),
-                otlp_endpoint: Some("".to_string()),
+                otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
                 tracing_sample_ratio: Some(Default::default()),
                 ..Default::default()
             },
@@ -217,7 +219,9 @@ fn test_load_standalone_example_config() {
                     sst_meta_cache_size: ReadableSize::mb(128),
                     vector_cache_size: ReadableSize::mb(512),
                     page_cache_size: ReadableSize::mb(512),
+                    selector_result_cache_size: ReadableSize::mb(512),
                     max_background_jobs: 4,
+                    experimental_write_cache_ttl: Some(Duration::from_secs(60 * 60 * 8)),
                     ..Default::default()
                 }),
                 RegionEngineConfig::File(EngineConfig {}),
@@ -228,7 +232,7 @@ fn test_load_standalone_example_config() {
             },
             logging: LoggingOptions {
                 level: Some("info".to_string()),
-                otlp_endpoint: Some("".to_string()),
+                otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
                 tracing_sample_ratio: Some(Default::default()),
                 ..Default::default()
             },

src/common/catalog/src/error.rs

@@ -1,46 +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::any::Any;
-
-use common_error::ext::ErrorExt;
-use common_error::status_code::StatusCode;
-use common_macro::stack_trace_debug;
-use snafu::{Location, Snafu};
-
-#[derive(Snafu)]
-#[snafu(visibility(pub))]
-#[stack_trace_debug]
-pub enum Error {
-    #[snafu(display("Invalid full table name: {}", table_name))]
-    InvalidFullTableName {
-        table_name: String,
-        #[snafu(implicit)]
-        location: Location,
-    },
-}
-
-impl ErrorExt for Error {
-    fn status_code(&self) -> StatusCode {
-        match self {
-            Error::InvalidFullTableName { .. } => StatusCode::Unexpected,
-        }
-    }
-
-    fn as_any(&self) -> &dyn Any {
-        self
-    }
-}
-
-pub type Result<T> = std::result::Result<T, Error>;

src/common/catalog/src/lib.rs

@@ -15,7 +15,6 @@
 use consts::DEFAULT_CATALOG_NAME;
 
 pub mod consts;
-pub mod error;
 
 #[inline]
 pub fn format_schema_name(catalog: &str, schema: &str) -> String {
@@ -48,19 +47,19 @@ pub fn build_db_string(catalog: &str, schema: &str) -> String {
 /// The database name may come from different sources:
 ///
 /// - MySQL `schema` name in MySQL protocol login request: it's optional and user
-/// and switch database using `USE` command
+///   and switch database using `USE` command
 /// - Postgres `database` parameter in Postgres wire protocol, required
 /// - HTTP RESTful API: the database parameter, optional
 /// - gRPC: the dbname field in header, optional but has a higher priority than
-/// original catalog/schema
+///   original catalog/schema
 ///
 /// When database name is provided, we attempt to parse catalog and schema from
 /// it. We assume the format `[<catalog>-]<schema>`:
 ///
 /// - If `[<catalog>-]` part is not provided, we use whole database name as
-/// schema name
+///   schema name
 /// - if `[<catalog>-]` is provided, we split database name with `-` and use
-/// `<catalog>` and `<schema>`.
+///   `<catalog>` and `<schema>`.
 pub fn parse_catalog_and_schema_from_db_string(db: &str) -> (String, String) {
     match parse_optional_catalog_and_schema_from_db_string(db) {
         (Some(catalog), schema) => (catalog, schema),

src/common/datasource/src/object_store/fs.rs

@@ -13,23 +13,19 @@
 // limitations under the License.
 
 use object_store::services::Fs;
+use object_store::util::DefaultLoggingInterceptor;
 use object_store::ObjectStore;
 use snafu::ResultExt;
 
 use crate::error::{BuildBackendSnafu, Result};
 
 pub fn build_fs_backend(root: &str) -> Result<ObjectStore> {
-    let mut builder = Fs::default();
-    let _ = builder.root(root);
-    let object_store = ObjectStore::new(builder)
+    let builder = Fs::default();
+    let object_store = ObjectStore::new(builder.root(root))
         .context(BuildBackendSnafu)?
-        .layer(
-            object_store::layers::LoggingLayer::default()
-                // Print the expected error only in DEBUG level.
-                // See https://docs.rs/opendal/latest/opendal/layers/struct.LoggingLayer.html#method.with_error_level
-                .with_error_level(Some("debug"))
-                .expect("input error level must be valid"),
-        )
+        .layer(object_store::layers::LoggingLayer::new(
+            DefaultLoggingInterceptor,
+        ))
         .layer(object_store::layers::TracingLayer)
         .layer(object_store::layers::PrometheusMetricsLayer::new(true))
         .finish();

src/common/datasource/src/object_store/s3.rs

@@ -15,6 +15,7 @@
 use std::collections::HashMap;
 
 use object_store::services::S3;
+use object_store::util::DefaultLoggingInterceptor;
 use object_store::ObjectStore;
 use snafu::ResultExt;
 
@@ -44,28 +45,26 @@ pub fn build_s3_backend(
     path: &str,
     connection: &HashMap<String, String>,
 ) -> Result<ObjectStore> {
-    let mut builder = S3::default();
-
-    let _ = builder.root(path).bucket(host);
+    let mut builder = S3::default().root(path).bucket(host);
 
     if let Some(endpoint) = connection.get(ENDPOINT) {
-        let _ = builder.endpoint(endpoint);
+        builder = builder.endpoint(endpoint);
     }
 
     if let Some(region) = connection.get(REGION) {
-        let _ = builder.region(region);
+        builder = builder.region(region);
     }
 
     if let Some(key_id) = connection.get(ACCESS_KEY_ID) {
-        let _ = builder.access_key_id(key_id);
+        builder = builder.access_key_id(key_id);
     }
 
     if let Some(key) = connection.get(SECRET_ACCESS_KEY) {
-        let _ = builder.secret_access_key(key);
+        builder = builder.secret_access_key(key);
     }
 
     if let Some(session_token) = connection.get(SESSION_TOKEN) {
-        let _ = builder.security_token(session_token);
+        builder = builder.session_token(session_token);
     }
 
     if let Some(enable_str) = connection.get(ENABLE_VIRTUAL_HOST_STYLE) {
@@ -79,20 +78,16 @@ pub fn build_s3_backend(
             .build()
         })?;
         if enable {
-            let _ = builder.enable_virtual_host_style();
+            builder = builder.enable_virtual_host_style();
         }
     }
 
     // TODO(weny): Consider finding a better way to eliminate duplicate code.
     Ok(ObjectStore::new(builder)
         .context(error::BuildBackendSnafu)?
-        .layer(
-            object_store::layers::LoggingLayer::default()
-                // Print the expected error only in DEBUG level.
-                // See https://docs.rs/opendal/latest/opendal/layers/struct.LoggingLayer.html#method.with_error_level
-                .with_error_level(Some("debug"))
-                .expect("input error level must be valid"),
-        )
+        .layer(object_store::layers::LoggingLayer::new(
+            DefaultLoggingInterceptor,
+        ))
         .layer(object_store::layers::TracingLayer)
         .layer(object_store::layers::PrometheusMetricsLayer::new(true))
         .finish())

src/common/datasource/src/test_util.rs

@@ -47,19 +47,15 @@ pub fn format_schema(schema: Schema) -> Vec<String> {
 }
 
 pub fn test_store(root: &str) -> ObjectStore {
-    let mut builder = Fs::default();
-    let _ = builder.root(root);
-
-    ObjectStore::new(builder).unwrap().finish()
+    let builder = Fs::default();
+    ObjectStore::new(builder.root(root)).unwrap().finish()
 }
 
 pub fn test_tmp_store(root: &str) -> (ObjectStore, TempDir) {
     let dir = create_temp_dir(root);
 
-    let mut builder = Fs::default();
-    let _ = builder.root("/");
-
-    (ObjectStore::new(builder).unwrap().finish(), dir)
+    let builder = Fs::default();
+    (ObjectStore::new(builder.root("/")).unwrap().finish(), dir)
 }
 
 pub fn test_basic_schema() -> SchemaRef {

src/common/function/Cargo.toml

@@ -7,6 +7,10 @@ license.workspace = true
 [lints]
 workspace = true
 
+[features]
+default = ["geo"]
+geo = ["geohash", "h3o"]
+
 [dependencies]
 api.workspace = true
 arc-swap = "1.0"
@@ -23,6 +27,8 @@ common-time.workspace = true
 common-version.workspace = true
 datafusion.workspace = true
 datatypes.workspace = true
+geohash = { version = "0.13", optional = true }
+h3o = { version = "0.6", optional = true }
 num = "0.4"
 num-traits = "0.2"
 once_cell.workspace = true
@@ -39,3 +45,4 @@ table.workspace = true
 [dev-dependencies]
 ron = "0.7"
 serde = { version = "1.0", features = ["derive"] }
+tokio.workspace = true

src/common/function/src/flush_flow.rs

@@ -110,7 +110,7 @@ mod test {
     use session::context::QueryContext;
 
     use super::*;
-    use crate::function::{Function, FunctionContext};
+    use crate::function::{AsyncFunction, FunctionContext};
 
     #[test]
     fn test_flush_flow_metadata() {
@@ -130,8 +130,8 @@ mod test {
         );
     }
 
-    #[test]
-    fn test_missing_flow_service() {
+    #[tokio::test]
+    async fn test_missing_flow_service() {
         let f = FlushFlowFunction;
 
         let args = vec!["flow_name"];
@@ -140,7 +140,7 @@ mod test {
             .map(|arg| Arc::new(StringVector::from_slice(&[arg])) as _)
             .collect::<Vec<_>>();
 
-        let result = f.eval(FunctionContext::default(), &args).unwrap_err();
+        let result = f.eval(FunctionContext::default(), &args).await.unwrap_err();
         assert_eq!(
             "Missing FlowServiceHandler, not expected",
             result.to_string()

src/common/function/src/function.rs

@@ -32,7 +32,7 @@ pub struct FunctionContext {
 
 impl FunctionContext {
     /// Create a mock [`FunctionContext`] for test.
-    #[cfg(any(test, feature = "testing"))]
+    #[cfg(test)]
     pub fn mock() -> Self {
         Self {
             query_ctx: QueryContextBuilder::default().build().into(),
@@ -56,8 +56,10 @@ pub trait Function: fmt::Display + Sync + Send {
     /// Returns the name of the function, should be unique.
     fn name(&self) -> &str;
 
+    /// The returned data type of function execution.
     fn return_type(&self, input_types: &[ConcreteDataType]) -> Result<ConcreteDataType>;
 
+    /// The signature of function.
     fn signature(&self) -> Signature;
 
     /// Evaluate the function, e.g. run/execute the function.
@@ -65,3 +67,22 @@ pub trait Function: fmt::Display + Sync + Send {
 }
 
 pub type FunctionRef = Arc<dyn Function>;
+
+/// Async Scalar function trait
+#[async_trait::async_trait]
+pub trait AsyncFunction: fmt::Display + Sync + Send {
+    /// Returns the name of the function, should be unique.
+    fn name(&self) -> &str;
+
+    /// The returned data type of function execution.
+    fn return_type(&self, input_types: &[ConcreteDataType]) -> Result<ConcreteDataType>;
+
+    /// The signature of function.
+    fn signature(&self) -> Signature;
+
+    /// Evaluate the function, e.g. run/execute the function.
+    /// TODO(dennis): simplify the signature and refactor all the admin functions.
+    async fn eval(&self, _func_ctx: FunctionContext, _columns: &[VectorRef]) -> Result<VectorRef>;
+}
+
+pub type AsyncFunctionRef = Arc<dyn AsyncFunction>;

src/common/function/src/function_registry.rs

@@ -18,7 +18,7 @@ use std::sync::{Arc, RwLock};
 
 use once_cell::sync::Lazy;
 
-use crate::function::FunctionRef;
+use crate::function::{AsyncFunctionRef, FunctionRef};
 use crate::scalars::aggregate::{AggregateFunctionMetaRef, AggregateFunctions};
 use crate::scalars::date::DateFunction;
 use crate::scalars::expression::ExpressionFunction;
@@ -32,6 +32,7 @@ use crate::table::TableFunction;
 #[derive(Default)]
 pub struct FunctionRegistry {
     functions: RwLock<HashMap<String, FunctionRef>>,
+    async_functions: RwLock<HashMap<String, AsyncFunctionRef>>,
     aggregate_functions: RwLock<HashMap<String, AggregateFunctionMetaRef>>,
 }
 
@@ -44,6 +45,27 @@ impl FunctionRegistry {
             .insert(func.name().to_string(), func);
     }
 
+    pub fn register_async(&self, func: AsyncFunctionRef) {
+        let _ = self
+            .async_functions
+            .write()
+            .unwrap()
+            .insert(func.name().to_string(), func);
+    }
+
+    pub fn get_async_function(&self, name: &str) -> Option<AsyncFunctionRef> {
+        self.async_functions.read().unwrap().get(name).cloned()
+    }
+
+    pub fn async_functions(&self) -> Vec<AsyncFunctionRef> {
+        self.async_functions
+            .read()
+            .unwrap()
+            .values()
+            .cloned()
+            .collect()
+    }
+
     pub fn register_aggregate_function(&self, func: AggregateFunctionMetaRef) {
         let _ = self
             .aggregate_functions
@@ -94,6 +116,10 @@ pub static FUNCTION_REGISTRY: Lazy<Arc<FunctionRegistry>> = Lazy::new(|| {
     SystemFunction::register(&function_registry);
     TableFunction::register(&function_registry);
 
+    // Geo functions
+    #[cfg(feature = "geo")]
+    crate::scalars::geo::GeoFunctions::register(&function_registry);
+
     Arc::new(function_registry)
 });
 

src/common/function/src/scalars.rs

@@ -15,6 +15,8 @@
 pub mod aggregate;
 pub(crate) mod date;
 pub mod expression;
+#[cfg(feature = "geo")]
+pub mod geo;
 pub mod matches;
 pub mod math;
 pub mod numpy;

src/common/function/src/scalars/aggregate/percentile.rs

@@ -75,7 +75,7 @@ where
         // to keep the not_greater length == floor+1
         // so to ensure the peek of the not_greater is array[floor]
         // and the peek of the greater is array[floor+1]
-        let p = if let Some(p) = self.p { p } else { 0.0_f64 };
+        let p = self.p.unwrap_or(0.0_f64);
         let floor = (((self.n - 1) as f64) * p / (100_f64)).floor();
         if value <= *self.not_greater.peek().unwrap() {
             self.not_greater.push(value);

src/common/function/src/scalars/geo.rs

@@ -0,0 +1,31 @@
+// 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;
+mod geohash;
+mod h3;
+
+use geohash::GeohashFunction;
+use h3::H3Function;
+
+use crate::function_registry::FunctionRegistry;
+
+pub(crate) struct GeoFunctions;
+
+impl GeoFunctions {
+    pub fn register(registry: &FunctionRegistry) {
+        registry.register(Arc::new(GeohashFunction));
+        registry.register(Arc::new(H3Function));
+    }
+}

src/common/function/src/scalars/geo/geohash.rs

@@ -0,0 +1,135 @@
+// 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::fmt;
+
+use common_error::ext::{BoxedError, PlainError};
+use common_error::status_code::StatusCode;
+use common_query::error::{self, InvalidFuncArgsSnafu, Result};
+use common_query::prelude::{Signature, TypeSignature};
+use datafusion::logical_expr::Volatility;
+use datatypes::prelude::ConcreteDataType;
+use datatypes::scalars::ScalarVectorBuilder;
+use datatypes::value::Value;
+use datatypes::vectors::{MutableVector, StringVectorBuilder, VectorRef};
+use geohash::Coord;
+use snafu::{ensure, ResultExt};
+
+use crate::function::{Function, FunctionContext};
+
+/// Function that return geohash string for a given geospatial coordinate.
+#[derive(Clone, Debug, Default)]
+pub struct GeohashFunction;
+
+const NAME: &str = "geohash";
+
+impl Function for GeohashFunction {
+    fn name(&self) -> &str {
+        NAME
+    }
+
+    fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::string_datatype())
+    }
+
+    fn signature(&self) -> Signature {
+        let mut signatures = Vec::new();
+        for coord_type in &[
+            ConcreteDataType::float32_datatype(),
+            ConcreteDataType::float64_datatype(),
+        ] {
+            for resolution_type in &[
+                ConcreteDataType::int8_datatype(),
+                ConcreteDataType::int16_datatype(),
+                ConcreteDataType::int32_datatype(),
+                ConcreteDataType::int64_datatype(),
+                ConcreteDataType::uint8_datatype(),
+                ConcreteDataType::uint16_datatype(),
+                ConcreteDataType::uint32_datatype(),
+                ConcreteDataType::uint64_datatype(),
+            ] {
+                signatures.push(TypeSignature::Exact(vec![
+                    // latitude
+                    coord_type.clone(),
+                    // longitude
+                    coord_type.clone(),
+                    // resolution
+                    resolution_type.clone(),
+                ]));
+            }
+        }
+        Signature::one_of(signatures, Volatility::Stable)
+    }
+
+    fn eval(&self, _func_ctx: FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
+        ensure!(
+            columns.len() == 3,
+            InvalidFuncArgsSnafu {
+                err_msg: format!(
+                    "The length of the args is not correct, expect 3, provided : {}",
+                    columns.len()
+                ),
+            }
+        );
+
+        let lat_vec = &columns[0];
+        let lon_vec = &columns[1];
+        let resolution_vec = &columns[2];
+
+        let size = lat_vec.len();
+        let mut results = StringVectorBuilder::with_capacity(size);
+
+        for i in 0..size {
+            let lat = lat_vec.get(i).as_f64_lossy();
+            let lon = lon_vec.get(i).as_f64_lossy();
+            let r = match resolution_vec.get(i) {
+                Value::Int8(v) => v as usize,
+                Value::Int16(v) => v as usize,
+                Value::Int32(v) => v as usize,
+                Value::Int64(v) => v as usize,
+                Value::UInt8(v) => v as usize,
+                Value::UInt16(v) => v as usize,
+                Value::UInt32(v) => v as usize,
+                Value::UInt64(v) => v as usize,
+                _ => unreachable!(),
+            };
+
+            let result = match (lat, lon) {
+                (Some(lat), Some(lon)) => {
+                    let coord = Coord { x: lon, y: lat };
+                    let encoded = geohash::encode(coord, r)
+                        .map_err(|e| {
+                            BoxedError::new(PlainError::new(
+                                format!("Geohash error: {}", e),
+                                StatusCode::EngineExecuteQuery,
+                            ))
+                        })
+                        .context(error::ExecuteSnafu)?;
+                    Some(encoded)
+                }
+                _ => None,
+            };
+
+            results.push(result.as_deref());
+        }
+
+        Ok(results.to_vector())
+    }
+}
+
+impl fmt::Display for GeohashFunction {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "{}", NAME)
+    }
+}

src/common/function/src/scalars/geo/h3.rs

@@ -0,0 +1,145 @@
+// 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::fmt;
+
+use common_error::ext::{BoxedError, PlainError};
+use common_error::status_code::StatusCode;
+use common_query::error::{self, InvalidFuncArgsSnafu, Result};
+use common_query::prelude::{Signature, TypeSignature};
+use datafusion::logical_expr::Volatility;
+use datatypes::prelude::ConcreteDataType;
+use datatypes::scalars::ScalarVectorBuilder;
+use datatypes::value::Value;
+use datatypes::vectors::{MutableVector, StringVectorBuilder, VectorRef};
+use h3o::{LatLng, Resolution};
+use snafu::{ensure, ResultExt};
+
+use crate::function::{Function, FunctionContext};
+
+/// Function that returns [h3] encoding string for a given geospatial coordinate.
+///
+/// [h3]: https://h3geo.org/
+#[derive(Clone, Debug, Default)]
+pub struct H3Function;
+
+const NAME: &str = "h3";
+
+impl Function for H3Function {
+    fn name(&self) -> &str {
+        NAME
+    }
+
+    fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::string_datatype())
+    }
+
+    fn signature(&self) -> Signature {
+        let mut signatures = Vec::new();
+        for coord_type in &[
+            ConcreteDataType::float32_datatype(),
+            ConcreteDataType::float64_datatype(),
+        ] {
+            for resolution_type in &[
+                ConcreteDataType::int8_datatype(),
+                ConcreteDataType::int16_datatype(),
+                ConcreteDataType::int32_datatype(),
+                ConcreteDataType::int64_datatype(),
+                ConcreteDataType::uint8_datatype(),
+                ConcreteDataType::uint16_datatype(),
+                ConcreteDataType::uint32_datatype(),
+                ConcreteDataType::uint64_datatype(),
+            ] {
+                signatures.push(TypeSignature::Exact(vec![
+                    // latitude
+                    coord_type.clone(),
+                    // longitude
+                    coord_type.clone(),
+                    // resolution
+                    resolution_type.clone(),
+                ]));
+            }
+        }
+        Signature::one_of(signatures, Volatility::Stable)
+    }
+
+    fn eval(&self, _func_ctx: FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
+        ensure!(
+            columns.len() == 3,
+            InvalidFuncArgsSnafu {
+                err_msg: format!(
+                    "The length of the args is not correct, expect 3, provided : {}",
+                    columns.len()
+                ),
+            }
+        );
+
+        let lat_vec = &columns[0];
+        let lon_vec = &columns[1];
+        let resolution_vec = &columns[2];
+
+        let size = lat_vec.len();
+        let mut results = StringVectorBuilder::with_capacity(size);
+
+        for i in 0..size {
+            let lat = lat_vec.get(i).as_f64_lossy();
+            let lon = lon_vec.get(i).as_f64_lossy();
+            let r = match resolution_vec.get(i) {
+                Value::Int8(v) => v as u8,
+                Value::Int16(v) => v as u8,
+                Value::Int32(v) => v as u8,
+                Value::Int64(v) => v as u8,
+                Value::UInt8(v) => v,
+                Value::UInt16(v) => v as u8,
+                Value::UInt32(v) => v as u8,
+                Value::UInt64(v) => v as u8,
+                _ => unreachable!(),
+            };
+
+            let result = match (lat, lon) {
+                (Some(lat), Some(lon)) => {
+                    let coord = LatLng::new(lat, lon)
+                        .map_err(|e| {
+                            BoxedError::new(PlainError::new(
+                                format!("H3 error: {}", e),
+                                StatusCode::EngineExecuteQuery,
+                            ))
+                        })
+                        .context(error::ExecuteSnafu)?;
+                    let r = Resolution::try_from(r)
+                        .map_err(|e| {
+                            BoxedError::new(PlainError::new(
+                                format!("H3 error: {}", e),
+                                StatusCode::EngineExecuteQuery,
+                            ))
+                        })
+                        .context(error::ExecuteSnafu)?;
+                    let encoded = coord.to_cell(r).to_string();
+                    Some(encoded)
+                }
+                _ => None,
+            };
+
+            results.push(result.as_deref());
+        }
+
+        Ok(results.to_vector())
+    }
+}
+
+impl fmt::Display for H3Function {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "{}", NAME)
+    }
+}

src/common/function/src/state.rs

@@ -28,7 +28,7 @@ pub struct FunctionState {
 
 impl FunctionState {
     /// Create a mock [`FunctionState`] for test.
-    #[cfg(any(test, feature = "testing"))]
+    #[cfg(test)]
     pub fn mock() -> Self {
         use std::sync::Arc;
 

src/common/function/src/system.rs

@@ -22,7 +22,7 @@ mod version;
 use std::sync::Arc;
 
 use build::BuildFunction;
-use database::DatabaseFunction;
+use database::{CurrentSchemaFunction, DatabaseFunction};
 use pg_catalog::PGCatalogFunction;
 use procedure_state::ProcedureStateFunction;
 use timezone::TimezoneFunction;
@@ -37,8 +37,9 @@ impl SystemFunction {
         registry.register(Arc::new(BuildFunction));
         registry.register(Arc::new(VersionFunction));
         registry.register(Arc::new(DatabaseFunction));
+        registry.register(Arc::new(CurrentSchemaFunction));
         registry.register(Arc::new(TimezoneFunction));
-        registry.register(Arc::new(ProcedureStateFunction));
+        registry.register_async(Arc::new(ProcedureStateFunction));
         PGCatalogFunction::register(registry);
     }
 }

src/common/function/src/system/database.rs

@@ -26,11 +26,35 @@ use crate::function::{Function, FunctionContext};
 #[derive(Clone, Debug, Default)]
 pub struct DatabaseFunction;
 
-const NAME: &str = "database";
+#[derive(Clone, Debug, Default)]
+pub struct CurrentSchemaFunction;
+
+const DATABASE_FUNCTION_NAME: &str = "database";
+const CURRENT_SCHEMA_FUNCTION_NAME: &str = "current_schema";
 
 impl Function for DatabaseFunction {
     fn name(&self) -> &str {
-        NAME
+        DATABASE_FUNCTION_NAME
+    }
+
+    fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::string_datatype())
+    }
+
+    fn signature(&self) -> Signature {
+        Signature::uniform(0, vec![], Volatility::Immutable)
+    }
+
+    fn eval(&self, func_ctx: FunctionContext, _columns: &[VectorRef]) -> Result<VectorRef> {
+        let db = func_ctx.query_ctx.current_schema();
+
+        Ok(Arc::new(StringVector::from_slice(&[&db])) as _)
+    }
+}
+
+impl Function for CurrentSchemaFunction {
+    fn name(&self) -> &str {
+        CURRENT_SCHEMA_FUNCTION_NAME
     }
 
     fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
@@ -54,6 +78,12 @@ impl fmt::Display for DatabaseFunction {
     }
 }
 
+impl fmt::Display for CurrentSchemaFunction {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "CURRENT_SCHEMA")
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use std::sync::Arc;

src/common/function/src/system/pg_catalog.rs

@@ -14,11 +14,13 @@
 
 mod pg_get_userbyid;
 mod table_is_visible;
+mod version;
 
 use std::sync::Arc;
 
 use pg_get_userbyid::PGGetUserByIdFunction;
 use table_is_visible::PGTableIsVisibleFunction;
+use version::PGVersionFunction;
 
 use crate::function_registry::FunctionRegistry;
 
@@ -35,5 +37,6 @@ impl PGCatalogFunction {
     pub fn register(registry: &FunctionRegistry) {
         registry.register(Arc::new(PGTableIsVisibleFunction));
         registry.register(Arc::new(PGGetUserByIdFunction));
+        registry.register(Arc::new(PGVersionFunction));
     }
 }

src/common/function/src/system/pg_catalog/version.rs

@@ -0,0 +1,54 @@
+// 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 std::{env, fmt};
+
+use common_query::error::Result;
+use common_query::prelude::{Signature, Volatility};
+use datatypes::data_type::ConcreteDataType;
+use datatypes::vectors::{StringVector, VectorRef};
+
+use crate::function::{Function, FunctionContext};
+
+#[derive(Clone, Debug, Default)]
+pub(crate) struct PGVersionFunction;
+
+impl fmt::Display for PGVersionFunction {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, crate::pg_catalog_func_fullname!("VERSION"))
+    }
+}
+
+impl Function for PGVersionFunction {
+    fn name(&self) -> &str {
+        crate::pg_catalog_func_fullname!("version")
+    }
+
+    fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
+        Ok(ConcreteDataType::string_datatype())
+    }
+
+    fn signature(&self) -> Signature {
+        Signature::exact(vec![], Volatility::Immutable)
+    }
+
+    fn eval(&self, _func_ctx: FunctionContext, _columns: &[VectorRef]) -> Result<VectorRef> {
+        let result = StringVector::from(vec![format!(
+            "PostgreSQL 16.3 GreptimeDB {}",
+            env!("CARGO_PKG_VERSION")
+        )]);
+        Ok(Arc::new(result))
+    }
+}

src/common/function/src/system/procedure_state.rs

@@ -96,7 +96,7 @@ mod tests {
     use datatypes::vectors::StringVector;
 
     use super::*;
-    use crate::function::{Function, FunctionContext};
+    use crate::function::{AsyncFunction, FunctionContext};
 
     #[test]
     fn test_procedure_state_misc() {
@@ -114,8 +114,8 @@ mod tests {
         ));
     }
 
-    #[test]
-    fn test_missing_procedure_service() {
+    #[tokio::test]
+    async fn test_missing_procedure_service() {
         let f = ProcedureStateFunction;
 
         let args = vec!["pid"];
@@ -125,15 +125,15 @@ mod tests {
             .map(|arg| Arc::new(StringVector::from_slice(&[arg])) as _)
             .collect::<Vec<_>>();
 
-        let result = f.eval(FunctionContext::default(), &args).unwrap_err();
+        let result = f.eval(FunctionContext::default(), &args).await.unwrap_err();
         assert_eq!(
             "Missing ProcedureServiceHandler, not expected",
             result.to_string()
         );
     }
 
-    #[test]
-    fn test_procedure_state() {
+    #[tokio::test]
+    async fn test_procedure_state() {
         let f = ProcedureStateFunction;
 
         let args = vec!["pid"];
@@ -143,7 +143,7 @@ mod tests {
             .map(|arg| Arc::new(StringVector::from_slice(&[arg])) as _)
             .collect::<Vec<_>>();
 
-        let result = f.eval(FunctionContext::mock(), &args).unwrap();
+        let result = f.eval(FunctionContext::mock(), &args).await.unwrap();
 
         let expect: VectorRef = Arc::new(StringVector::from(vec![
             "{\"status\":\"Done\",\"error\":\"OK\"}",

src/common/function/src/table.rs

@@ -31,11 +31,11 @@ pub(crate) struct TableFunction;
 impl TableFunction {
     /// Register all table functions to [`FunctionRegistry`].
     pub fn register(registry: &FunctionRegistry) {
-        registry.register(Arc::new(MigrateRegionFunction));
-        registry.register(Arc::new(FlushRegionFunction));
-        registry.register(Arc::new(CompactRegionFunction));
-        registry.register(Arc::new(FlushTableFunction));
-        registry.register(Arc::new(CompactTableFunction));
-        registry.register(Arc::new(FlushFlowFunction));
+        registry.register_async(Arc::new(MigrateRegionFunction));
+        registry.register_async(Arc::new(FlushRegionFunction));
+        registry.register_async(Arc::new(CompactRegionFunction));
+        registry.register_async(Arc::new(FlushTableFunction));
+        registry.register_async(Arc::new(CompactTableFunction));
+        registry.register_async(Arc::new(FlushFlowFunction));
     }
 }

src/common/function/src/table/flush_compact_region.rs

@@ -77,7 +77,7 @@ mod tests {
     use datatypes::vectors::UInt64Vector;
 
     use super::*;
-    use crate::function::{Function, FunctionContext};
+    use crate::function::{AsyncFunction, FunctionContext};
 
     macro_rules! define_region_function_test {
         ($name: ident, $func: ident) => {
@@ -97,8 +97,8 @@ mod tests {
                                      } if valid_types == ConcreteDataType::numerics()));
                 }
 
-                #[test]
-                fn [<test_ $name _missing_table_mutation>]() {
+                #[tokio::test]
+                async fn [<test_ $name _missing_table_mutation>]() {
                     let f = $func;
 
                     let args = vec![99];
@@ -108,15 +108,15 @@ mod tests {
                         .map(|arg| Arc::new(UInt64Vector::from_slice([arg])) as _)
                         .collect::<Vec<_>>();
 
-                    let result = f.eval(FunctionContext::default(), &args).unwrap_err();
+                    let result = f.eval(FunctionContext::default(), &args).await.unwrap_err();
                     assert_eq!(
                         "Missing TableMutationHandler, not expected",
                         result.to_string()
                     );
                 }
 
-                #[test]
-                fn [<test_ $name>]() {
+                #[tokio::test]
+                async fn [<test_ $name>]() {
                     let f = $func;
 
 
@@ -127,7 +127,7 @@ mod tests {
                         .map(|arg| Arc::new(UInt64Vector::from_slice([arg])) as _)
                         .collect::<Vec<_>>();
 
-                    let result = f.eval(FunctionContext::mock(), &args).unwrap();
+                    let result = f.eval(FunctionContext::mock(), &args).await.unwrap();
 
                     let expect: VectorRef = Arc::new(UInt64Vector::from_slice([42]));
                     assert_eq!(expect, result);

src/common/function/src/table/flush_compact_table.rs

@@ -33,6 +33,8 @@ use crate::handlers::TableMutationHandlerRef;
 
 /// Compact type: strict window.
 const COMPACT_TYPE_STRICT_WINDOW: &str = "strict_window";
+/// Compact type: strict window (short name).
+const COMPACT_TYPE_STRICT_WINDOW_SHORT: &str = "swcs";
 
 #[admin_fn(
     name = FlushTableFunction,
@@ -168,8 +170,12 @@ fn parse_compact_params(
     })
 }
 
+/// Parses compaction strategy type. For `strict_window` or `swcs` strict window compaction is chose,
+/// otherwise choose regular (TWCS) compaction.
 fn parse_compact_type(type_str: &str, option: Option<&str>) -> Result<compact_request::Options> {
-    if type_str.eq_ignore_ascii_case(COMPACT_TYPE_STRICT_WINDOW) {
+    if type_str.eq_ignore_ascii_case(COMPACT_TYPE_STRICT_WINDOW)
+        | type_str.eq_ignore_ascii_case(COMPACT_TYPE_STRICT_WINDOW_SHORT)
+    {
         let window_seconds = option
             .map(|v| {
                 i64::from_str(v).map_err(|_| {
@@ -204,7 +210,7 @@ mod tests {
     use session::context::QueryContext;
 
     use super::*;
-    use crate::function::{Function, FunctionContext};
+    use crate::function::{AsyncFunction, FunctionContext};
 
     macro_rules! define_table_function_test {
         ($name: ident, $func: ident) => {
@@ -224,8 +230,8 @@ mod tests {
                                      } if valid_types == vec![ConcreteDataType::string_datatype()]));
                 }
 
-                #[test]
-                fn [<test_ $name _missing_table_mutation>]() {
+                #[tokio::test]
+                async fn [<test_ $name _missing_table_mutation>]() {
                     let f = $func;
 
                     let args = vec!["test"];
@@ -235,15 +241,15 @@ mod tests {
                         .map(|arg| Arc::new(StringVector::from(vec![arg])) as _)
                         .collect::<Vec<_>>();
 
-                    let result = f.eval(FunctionContext::default(), &args).unwrap_err();
+                    let result = f.eval(FunctionContext::default(), &args).await.unwrap_err();
                     assert_eq!(
                         "Missing TableMutationHandler, not expected",
                         result.to_string()
                     );
                 }
 
-                #[test]
-                fn [<test_ $name>]() {
+                #[tokio::test]
+                async fn [<test_ $name>]() {
                     let f = $func;
 
 
@@ -254,7 +260,7 @@ mod tests {
                         .map(|arg| Arc::new(StringVector::from(vec![arg])) as _)
                         .collect::<Vec<_>>();
 
-                    let result = f.eval(FunctionContext::mock(), &args).unwrap();
+                    let result = f.eval(FunctionContext::mock(), &args).await.unwrap();
 
                     let expect: VectorRef = Arc::new(UInt64Vector::from_slice([42]));
                     assert_eq!(expect, result);
@@ -350,6 +356,17 @@ mod tests {
                     compact_options: Options::Regular(Default::default()),
                 },
             ),
+            (
+                &["table", "swcs", "120"],
+                CompactTableRequest {
+                    catalog_name: DEFAULT_CATALOG_NAME.to_string(),
+                    schema_name: DEFAULT_SCHEMA_NAME.to_string(),
+                    table_name: "table".to_string(),
+                    compact_options: Options::StrictWindow(StrictWindow {
+                        window_seconds: 120,
+                    }),
+                },
+            ),
         ]);
 
         assert!(parse_compact_params(

src/common/function/src/table/migrate_region.rs

@@ -123,7 +123,7 @@ mod tests {
     use datatypes::vectors::{StringVector, UInt64Vector, VectorRef};
 
     use super::*;
-    use crate::function::{Function, FunctionContext};
+    use crate::function::{AsyncFunction, FunctionContext};
 
     #[test]
     fn test_migrate_region_misc() {
@@ -140,8 +140,8 @@ mod tests {
                          } if sigs.len() == 2));
     }
 
-    #[test]
-    fn test_missing_procedure_service() {
+    #[tokio::test]
+    async fn test_missing_procedure_service() {
         let f = MigrateRegionFunction;
 
         let args = vec![1, 1, 1];
@@ -151,15 +151,15 @@ mod tests {
             .map(|arg| Arc::new(UInt64Vector::from_slice([arg])) as _)
             .collect::<Vec<_>>();
 
-        let result = f.eval(FunctionContext::default(), &args).unwrap_err();
+        let result = f.eval(FunctionContext::default(), &args).await.unwrap_err();
         assert_eq!(
             "Missing ProcedureServiceHandler, not expected",
             result.to_string()
         );
     }
 
-    #[test]
-    fn test_migrate_region() {
+    #[tokio::test]
+    async fn test_migrate_region() {
         let f = MigrateRegionFunction;
 
         let args = vec![1, 1, 1];
@@ -169,7 +169,7 @@ mod tests {
             .map(|arg| Arc::new(UInt64Vector::from_slice([arg])) as _)
             .collect::<Vec<_>>();
 
-        let result = f.eval(FunctionContext::mock(), &args).unwrap();
+        let result = f.eval(FunctionContext::mock(), &args).await.unwrap();
 
         let expect: VectorRef = Arc::new(StringVector::from(vec!["test_pid"]));
         assert_eq!(expect, result);

src/common/grpc-expr/src/error.rs

@@ -64,12 +64,6 @@ pub enum Error {
         location: Location,
     },
 
-    #[snafu(display("Invalid column proto: {}", err_msg))]
-    InvalidColumnProto {
-        err_msg: String,
-        #[snafu(implicit)]
-        location: Location,
-    },
     #[snafu(display("Failed to create vector"))]
     CreateVector {
         #[snafu(implicit)]
@@ -137,7 +131,6 @@ impl ErrorExt for Error {
             Error::DuplicatedTimestampColumn { .. }
             | Error::DuplicatedColumnName { .. }
             | Error::MissingTimestampColumn { .. } => StatusCode::InvalidArguments,
-            Error::InvalidColumnProto { .. } => StatusCode::InvalidArguments,
             Error::CreateVector { .. } => StatusCode::InvalidArguments,
             Error::MissingField { .. } => StatusCode::InvalidArguments,
             Error::InvalidColumnDef { source, .. } => source.status_code(),

src/common/macro/src/admin_fn.rs

@@ -187,7 +187,8 @@ fn build_struct(
         }
 
 
-        impl crate::function::Function for #name {
+        #[async_trait::async_trait]
+        impl crate::function::AsyncFunction for #name {
             fn name(&self) -> &'static str {
                 #display_name
             }
@@ -200,7 +201,7 @@ fn build_struct(
                 #sig_fn()
             }
 
-            fn eval(&self, func_ctx: crate::function::FunctionContext, columns: &[datatypes::vectors::VectorRef]) ->  common_query::error::Result<datatypes::vectors::VectorRef> {
+            async fn eval(&self, func_ctx: crate::function::FunctionContext, columns: &[datatypes::vectors::VectorRef]) ->  common_query::error::Result<datatypes::vectors::VectorRef> {
                 // Ensure under the `greptime` catalog for security
                 crate::ensure_greptime!(func_ctx);
 
@@ -212,51 +213,36 @@ fn build_struct(
                 };
                 let columns = Vec::from(columns);
 
-                // TODO(dennis): DataFusion doesn't support async UDF currently
-                std::thread::spawn(move || {
-                    use snafu::OptionExt;
-                    use datatypes::data_type::DataType;
+                use snafu::OptionExt;
+                use datatypes::data_type::DataType;
 
-                    let query_ctx = &func_ctx.query_ctx;
-                    let handler = func_ctx
-                        .state
-                        .#handler
-                        .as_ref()
-                        .context(#snafu_type)?;
+                let query_ctx = &func_ctx.query_ctx;
+                let handler = func_ctx
+                    .state
+                    .#handler
+                    .as_ref()
+                    .context(#snafu_type)?;
 
-                    let mut builder = store_api::storage::ConcreteDataType::#ret()
-                        .create_mutable_vector(rows_num);
+                let mut builder = store_api::storage::ConcreteDataType::#ret()
+                    .create_mutable_vector(rows_num);
 
-                    if columns_num == 0 {
-                        let result = common_runtime::block_on_global(async move {
-                            #fn_name(handler, query_ctx, &[]).await
-                        })?;
+                if columns_num == 0 {
+                    let result = #fn_name(handler, query_ctx, &[]).await?;
+
+                    builder.push_value_ref(result.as_value_ref());
+                } else {
+                    for i in 0..rows_num {
+                        let args: Vec<_> = columns.iter()
+                            .map(|vector| vector.get_ref(i))
+                            .collect();
+
+                        let result = #fn_name(handler, query_ctx, &args).await?;
 
                         builder.push_value_ref(result.as_value_ref());
-                    } else {
-                        for i in 0..rows_num {
-                            let args: Vec<_> = columns.iter()
-                                .map(|vector| vector.get_ref(i))
-                                .collect();
-
-                            let result = common_runtime::block_on_global(async move {
-                                #fn_name(handler, query_ctx, &args).await
-                            })?;
-
-                            builder.push_value_ref(result.as_value_ref());
-                        }
                     }
+                }
 
-                    Ok(builder.to_vector())
-                })
-                    .join()
-                    .map_err(|e| {
-                        common_telemetry::error!(e; "Join thread error");
-                        common_query::error::Error::ThreadJoin {
-                            location: snafu::Location::default(),
-                        }
-                    })?
-
+                Ok(builder.to_vector())
             }
 
         }

src/common/macro/src/lib.rs

@@ -76,6 +76,7 @@ pub fn range_fn(args: TokenStream, input: TokenStream) -> TokenStream {
 ///    - `&ProcedureServiceHandlerRef` or `&TableMutationHandlerRef` or `FlowServiceHandlerRef` as the first argument,
 ///    - `&QueryContextRef` as the second argument, and
 ///    - `&[ValueRef<'_>]` as the third argument which is SQL function input values in each row.
+///
 /// Return type must be `common_query::error::Result<Value>`.
 ///
 /// # Example see `common/function/src/system/procedure_state.rs`.

src/common/mem-prof/Cargo.toml

@@ -15,8 +15,8 @@ tempfile = "3.4"
 tokio.workspace = true
 
 [target.'cfg(not(windows))'.dependencies]
-tikv-jemalloc-ctl = { version = "0.5", features = ["use_std"] }
+tikv-jemalloc-ctl = { version = "0.6", features = ["use_std", "stats"] }
 
 [target.'cfg(not(windows))'.dependencies.tikv-jemalloc-sys]
 features = ["stats", "profiling", "unprefixed_malloc_on_supported_platforms"]
-version = "0.5"
+version = "0.6"

src/common/meta/Cargo.toml

@@ -6,6 +6,7 @@ license.workspace = true
 
 [features]
 testing = []
+pg_kvbackend = ["dep:tokio-postgres"]
 
 [lints]
 workspace = true
@@ -56,6 +57,7 @@ store-api.workspace = true
 strum.workspace = true
 table.workspace = true
 tokio.workspace = true
+tokio-postgres = { workspace = true, optional = true }
 tonic.workspace = true
 typetag = "0.2"