fix: nightly jsonbench test (#8212)

* fix: nightly jsonbench test Signed-off-by: luofucong <luofc@foxmail.com> * Apply suggestions from code review Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> --------- Signed-off-by: luofucong <luofc@foxmail.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
2026-06-03 13:50:40 +00:00 · 2026-06-02 14:16:08 +08:00
parent c1b0418377
commit c9f6f67ae9
1 changed files with 109 additions and 48 deletions
--- a/.github/workflows/nightly-jsonbench.yaml
+++ b/.github/workflows/nightly-jsonbench.yaml
@@ -1,19 +1,81 @@
 name: Nightly JSONBench

 on:
-  schedule:
-    # Trigger at 00:00(Asia/Shanghai) on every weekday.
-    - cron: "0 16 * * 0-4"
+  workflow_run:
+    workflows: [ "GreptimeDB Nightly Build" ]
+    types: [ completed ]
  workflow_dispatch:
+    inputs:
+      run_id:
+        description: The nightly build workflow run id to download GreptimeDB artifacts from
+        required: true
+        type: string
+
+permissions:
+  actions: read
+  contents: read

 concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
  cancel-in-progress: true

 jobs:
+  resolve-artifact:
+    name: Resolve GreptimeDB nightly artifact
+    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success') }}
+    runs-on: ubuntu-latest
+    outputs:
+      artifact-name: ${{ steps.find-artifact.outputs.artifact-name }}
+      run-id: ${{ steps.resolve-run-id.outputs.run-id }}
+    steps:
+      - name: Resolve nightly build run id
+        id: resolve-run-id
+        shell: bash
+        env:
+          EVENT_NAME: ${{ github.event_name }}
+          WORKFLOW_RUN_ID: ${{ github.event.workflow_run.id }}
+          INPUT_RUN_ID: ${{ inputs.run_id }}
+        run: |
+          set -euo pipefail
+
+          if [[ "${EVENT_NAME}" == "workflow_dispatch" ]]; then
+            run_id="${INPUT_RUN_ID}"
+          else
+            run_id="${WORKFLOW_RUN_ID}"
+          fi
+
+          if [[ ! "${run_id}" =~ ^[0-9]+$ ]]; then
+            echo "Invalid workflow run id: ${run_id}"
+            exit 1
+          fi
+
+          echo "run-id=${run_id}" >> "${GITHUB_OUTPUT}"
+
+      - name: Find GreptimeDB nightly artifact
+        id: find-artifact
+        shell: bash
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          RUN_ID: ${{ steps.resolve-run-id.outputs.run-id }}
+        run: |
+          set -euo pipefail
+
+          artifact_name=$(gh api "repos/${GITHUB_REPOSITORY}/actions/runs/${RUN_ID}/artifacts" --paginate \
+            --jq '.artifacts[] | select(.name | test("^greptime-linux-arm64-nightly-[0-9]{8}-[0-9a-f]+$")) | .name' \
+            | head -n 1)
+
+          if [[ -z "${artifact_name}" ]]; then
+            echo "Cannot find linux arm64 nightly artifact in workflow run ${RUN_ID}."
+            exit 1
+          fi
+
+          echo "Download GreptimeDB artifact: ${artifact_name}"
+          echo "artifact-name=${artifact_name}" >> "${GITHUB_OUTPUT}"
+
  allocate-runner:
    name: Allocate runner
-    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
+    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success') }}
+    needs: [ resolve-artifact ]
    runs-on: ubuntu-latest
    outputs:
      linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
@@ -43,55 +105,50 @@ jobs:

  jsonbench:
    name: Run JSONBench
-    if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
-    needs: [ allocate-runner ]
+    if: ${{ github.repository == 'GreptimeTeam/greptimedb' && (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success') }}
+    needs: [ resolve-artifact, allocate-runner ]
    runs-on: ${{ needs.allocate-runner.outputs.linux-arm64-runner }}
    timeout-minutes: 120
    env:
-      JSONBENCH_DATA_DIR: /home/runner/data/bluesky
-      JSONBENCH_OUTPUT_PREFIX: _ubuntu-latest
+      JSONBENCH_OUTPUT_PREFIX: _linux-arm64
    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
+      - name: Download GreptimeDB nightly artifact
+        uses: actions/download-artifact@v4
        with:
-          fetch-depth: 0
-          persist-credentials: false
+          name: ${{ needs.resolve-artifact.outputs.artifact-name }}
+          path: greptimedb-artifact
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+          run-id: ${{ needs.resolve-artifact.outputs.run-id }}

-      - uses: arduino/setup-protoc@v3
-        with:
-          repo-token: ${{ secrets.GITHUB_TOKEN }}
-
-      - uses: actions-rust-lang/setup-rust-toolchain@v1
-
-      - name: Rust Cache
-        uses: Swatinem/rust-cache@v2
-        with:
-          shared-key: "nightly-jsonbench"
-          cache-all-crates: "true"
-          save-if: ${{ github.ref == 'refs/heads/main' }}
-
-      - name: Build GreptimeDB
-        run: cargo build --profile nightly --bin greptime
-
-      - name: Reclaim disk space
+      - name: Prepare GreptimeDB binary
        shell: bash
        run: |
          set -euo pipefail

-          mkdir -p "${RUNNER_TEMP}/greptimedb-bin"
-          cp ./target/nightly/greptime "${RUNNER_TEMP}/greptimedb-bin/greptime"
-          chmod +x "${RUNNER_TEMP}/greptimedb-bin/greptime"
-
-          rm -rf ./target
+          tar -xzf "greptimedb-artifact/${{ needs.resolve-artifact.outputs.artifact-name }}.tar.gz"
+          cp "${{ needs.resolve-artifact.outputs.artifact-name }}/greptime" ./greptime
+          chmod +x ./greptime
+          rm -rf greptimedb-artifact "${{ needs.resolve-artifact.outputs.artifact-name }}"

      - name: Run JSONBench
+        env:
+          # TODO(LFC): Change to "3" (100m) when JSON2 ingestion performance is optimized.
+          JSONBENCH_DATASET: 2
        shell: bash
        run: |
          set -euo pipefail

-          cd "${RUNNER_TEMP}"
-          cp "${RUNNER_TEMP}/greptimedb-bin/greptime" ./greptime
-          chmod +x ./greptime
+          export JSONBENCH_DATA_DIR="/root/data/bluesky"
+          echo "Use JSONBench data directory ${JSONBENCH_DATA_DIR}"
+
+          echo "Cloning JSONBench"
+          git clone --branch greptimedb-new-json --depth 1 https://github.com/GreptimeTeam/JSONBench.git JSONBench
+
+          echo "Downloading JSONBench dataset choice ${JSONBENCH_DATASET} to ${JSONBENCH_DATA_DIR}"
+          mkdir -p "${JSONBENCH_DATA_DIR}"
+          printf "${JSONBENCH_DATASET}\n" | ./JSONBench/download_data.sh
+          downloaded_files=$(find "${JSONBENCH_DATA_DIR}" -type f | wc -l)
+          echo "Downloaded JSONBench dataset files: ${downloaded_files}"

          export GREPTIMEDB_STANDALONE__WAL__DIR=greptimedb_data/wal
          export GREPTIMEDB_STANDALONE__STORAGE__DATA_HOME=greptimedb_data
@@ -100,10 +157,12 @@ jobs:
          export GREPTIMEDB_STANDALONE__HTTP__BODY_LIMIT=1GB
          export GREPTIMEDB_STANDALONE__HTTP__TIMEOUT=500s

+          echo "Starting GreptimeDB standalone"
          ./greptime standalone start > greptimedb.log 2>&1 &
          greptime_pid=$!
          trap 'kill "${greptime_pid}" 2>/dev/null || true' EXIT

+          echo "Waiting for GreptimeDB health check"
          until curl -s --fail -o /dev/null http://localhost:4000/health; do
            if ! kill -0 "${greptime_pid}" 2>/dev/null; then
              cat greptimedb.log
@@ -111,12 +170,14 @@ jobs:
            fi
            sleep 1
          done
+          echo "GreptimeDB is ready"

-          git clone --branch greptimedb-new-json --depth 1 https://github.com/GreptimeTeam/JSONBench.git JSONBench
          cp ./greptime JSONBench/greptimedb/greptime

          cd JSONBench/greptimedb
-          ./main.sh 3 "${JSONBENCH_DATA_DIR}" success.log error.log "${JSONBENCH_OUTPUT_PREFIX}" false
+          echo "Running JSONBench main.sh with dataset choice ${JSONBENCH_DATASET} and install=false"
+          ./main.sh ${JSONBENCH_DATASET} "${JSONBENCH_DATA_DIR}" success.log error.log "${JSONBENCH_OUTPUT_PREFIX}" false
+          echo "JSONBench finished"

      - name: Upload JSONBench results
        if: always()
@@ -124,21 +185,21 @@ jobs:
        with:
          name: jsonbench-results
          path: |
-            ${{ runner.temp }}/greptimedb.log
-            ${{ runner.temp }}/JSONBench/greptimedb/*.log
-            ${{ runner.temp }}/JSONBench/greptimedb/*.total_size
-            ${{ runner.temp }}/JSONBench/greptimedb/*.data_size
-            ${{ runner.temp }}/JSONBench/greptimedb/*.index_size
-            ${{ runner.temp }}/JSONBench/greptimedb/*.count
-            ${{ runner.temp }}/JSONBench/greptimedb/*.results_runtime
-            ${{ runner.temp }}/JSONBench/greptimedb/*.query_results
+            ./greptimedb.log
+            ./JSONBench/greptimedb/*.log
+            ./JSONBench/greptimedb/*.total_size
+            ./JSONBench/greptimedb/*.data_size
+            ./JSONBench/greptimedb/*.index_size
+            ./JSONBench/greptimedb/*.count
+            ./JSONBench/greptimedb/*.results_runtime
+            ./JSONBench/greptimedb/*.query_results
          if-no-files-found: ignore
          retention-days: 7

  stop-linux-arm64-runner:
    name: Stop Linux ARM64 runner
    # It's always run as the last job in the workflow to make sure that the runner is released.
-    if: ${{ always() }}
+    if: ${{ always() && needs.allocate-runner.outputs.linux-arm64-ec2-runner-instance-id != '' }}
    runs-on: ubuntu-latest
    needs: [
      allocate-runner,