rustfs/.github/workflows/rustfs-tier-test.yml

name: RustFS Tier Test

on:
  workflow_dispatch:
    inputs:
      rustfs_version:
        description: 'RustFS release tag to test (e.g. 1.0.0-rc.4-preview.1)'
        required: false
        default: '1.0.0-rc.4-preview.1'
      package_url:
        description: 'Direct .deb URL (nightly/R2/dev). Overrides rustfs_version.'
        required: false
        type: string
      rc_sha256:
        description: 'Optional SHA-256 for the preinstalled rc binary; mismatch is an infrastructure failure.'
        required: false
        type: string
      force_case_failure:
        description: 'Diagnostic only: rewrite single-single/TIER-101 to FAIL after execution to verify artifact and final-gate behavior.'
        required: false
        default: false
        type: boolean
  repository_dispatch:
    # Chain handoff: dispatched when the KMS suite finishes.
    types: [rustfs-chain-tier]

permissions:
  contents: read

concurrency:
  group: rustfs-shared-functional-tests
  cancel-in-progress: false

defaults:
  run:
    shell: bash

env:
  RUSTFS_ACCESS_KEY: ${{ secrets.RUSTFS_ACCESS_KEY }}
  RUSTFS_SECRET_KEY: ${{ secrets.RUSTFS_SECRET_KEY }}
  RUSTFS_NODES: ${{ secrets.RUSTFS_NODES || vars.RUSTFS_NODES }}
  RUSTFS_SSH_USER: ${{ secrets.RUSTFS_SSH_USER || vars.RUSTFS_SSH_USER }}
  RUSTFS_NIGHTLY_PACKAGE_URL: ${{ vars.RUSTFS_NIGHTLY_PACKAGE_URL || 'https://dl.rustfs.com/artifacts/rustfs/packages/nightly/rustfs-nightly-latest.deb' }}
  RUSTFS_EXPECTED_RC_SHA256: ${{ inputs.rc_sha256 || vars.RUSTFS_TIER_RC_SHA256 }}
  PF_TESTING_GH_TOKEN: ${{ secrets.PF_TESTING_GH_TOKEN }}
  TIER_ARTIFACTS_DIR: /tmp/rustfs-tier-artifacts-${{ github.run_id }}-${{ github.run_attempt }}

jobs:
  tier-test:
    runs-on: smoke-testing
    # Requirement: a failing suite must not fail the workflow; failures
    # are filed to rustfs/backlog and the chain continues.
    continue-on-error: true
    timeout-minutes: 420
    if: ${{ github.event_name == 'workflow_dispatch' || github.event_name == 'repository_dispatch' }}
    steps:
      # auto-testing is private: clone it with the dedicated PF token (not
      # GITHUB_TOKEN) and retry transient GitHub/network failures.
      - name: Checkout auto-testing scripts (with retry)
        env:
          GH_TOKEN: ${{ secrets.PF_TESTING_GH_TOKEN }}
        run: |
          set -euo pipefail
          rm -rf auto-testing
          for attempt in 1 2 3 4 5; do
            if gh repo clone rustfs/auto-testing auto-testing -- --depth 1 --quiet; then
              echo "auto-testing cloned (attempt ${attempt})"
              exit 0
            fi
            rm -rf auto-testing
            echo "clone attempt ${attempt} failed; retrying in $((attempt * 15))s" >&2
            sleep $((attempt * 15))
          done
          echo "ERROR: unable to clone rustfs/auto-testing after 5 attempts" >&2
          exit 1

      - name: Show environment
        run: |
          uname -a
          jq --version
          openssl version
          df -h /data | tail -1

      - name: Cleanup environment (before)
        run: |
          set -euo pipefail
          sudo docker rm -f rustfs-test-mqtt >/dev/null 2>&1 || true
          sudo rm -f /tmp/rustfs-mosquitto.conf
          read -r -a NODES <<< "${RUSTFS_NODES:-vm000 vm001 vm002}"
          SSH_USER="${RUSTFS_SSH_USER:-azureuser}"
          for node in "${NODES[@]}"; do
            ssh -o BatchMode=yes -o ConnectTimeout=10 -o StrictHostKeyChecking=accept-new "${SSH_USER}@${node}" '
              set -euo pipefail
              SUDO=""; [ "$(id -u)" -ne 0 ] && SUDO="sudo -n"
              ${SUDO} systemctl stop rustfs 2>/dev/null || true
              if ${SUDO} dpkg -l rustfs 2>/dev/null | grep -q "^ii"; then
                ${SUDO} dpkg -P rustfs
              fi
              for i in 1 2 3 4; do ${SUDO} rm -rf /data/rustfs${i}/mnmd; done
              ${SUDO} rm -rf /var/log/rustfs /var/lib/rustfs/kms
            '
          done

      - name: Ensure MQTT broker + clients
        run: |
          set -euo pipefail
          if ! command -v mosquitto_sub >/dev/null 2>&1; then
            sudo apt-get update
            sudo apt-get install -y mosquitto-clients
          fi
          command -v docker >/dev/null 2>&1 || { echo 'docker not found on runner'; exit 1; }
          sudo docker rm -f rustfs-test-mqtt >/dev/null 2>&1 || true
          cat <<'EOF' | sudo tee /tmp/rustfs-mosquitto.conf >/dev/null
          listener 1883 0.0.0.0
          allow_anonymous true
          EOF
          sudo docker run -d --name rustfs-test-mqtt -p 1883:1883 \
            -v /tmp/rustfs-mosquitto.conf:/mosquitto/config/mosquitto.conf:ro \
            eclipse-mosquitto:2 >/dev/null
          for _ in {1..10}; do
            if ss -tln 2>/dev/null | grep -q ':1883'; then
              break
            fi
            sleep 1
          done
          ss -tln 2>/dev/null | grep -q ':1883' || {
            echo 'mosquitto container is not listening on 1883'
            sudo docker logs rustfs-test-mqtt || true
            exit 1
          }

      - name: Run tier suite
        id: test
        continue-on-error: true
        env:
          LOG_FILE: /tmp/rustfs-tier.log
          PACKAGE_URL_INPUT: ${{ inputs.package_url }}
          RUSTFS_VERSION_INPUT: ${{ inputs.rustfs_version }}
        run: |
          set -euo pipefail
          chmod +x auto-testing/rustfs-tier-test.sh
          RC_BIN="$(command -v rc)"
          PACKAGE_URL="${PACKAGE_URL_INPUT}"
          RUSTFS_VERSION="${RUSTFS_VERSION_INPUT}"
          ARGS=(
            --all-topologies
            -y
            --log-file "${LOG_FILE}"
            --rc-bin "${RC_BIN}"
            --artifacts-dir "${TIER_ARTIFACTS_DIR}"
          )
          if [ -n "${RUSTFS_EXPECTED_RC_SHA256}" ]; then
            ARGS+=(--expected-rc-sha256 "${RUSTFS_EXPECTED_RC_SHA256}")
          fi
          if [ -n "${PACKAGE_URL}" ]; then
            ARGS+=(--package-url "${PACKAGE_URL}")
          elif [ -n "${RUSTFS_VERSION}" ]; then
            ARGS+=(--version "${RUSTFS_VERSION}")
          else
            ARGS+=(--package-url "${RUSTFS_NIGHTLY_PACKAGE_URL}")
          fi
          ./auto-testing/rustfs-tier-test.sh "${ARGS[@]}"

      - name: Inject diagnostic case failure
        if: ${{ always() && inputs.force_case_failure }}
        run: |
          set -euo pipefail
          RESULT_FILE="${TIER_ARTIFACTS_DIR}/cases/single-single--TIER-101.json"
          test -s "${RESULT_FILE}"
          TMP_FILE="$(mktemp "${TIER_ARTIFACTS_DIR}/cases/.forced.XXXXXX")"
          jq '.status = "FAIL" | .case_rc = 97' "${RESULT_FILE}" > "${TMP_FILE}"
          mv "${TMP_FILE}" "${RESULT_FILE}"

      - name: Generate report
        if: always()
        env:
          LOG_FILE: /tmp/rustfs-tier.log
          REPORT_FILE: /tmp/rustfs-tier-report.md
          CASE_TABLE: /tmp/rustfs-tier-cases.md
          GATE_RC_FILE: /tmp/rustfs-tier-gate.rc
          PACKAGE_URL_INPUT: ${{ inputs.package_url }}
          RUSTFS_VERSION_INPUT: ${{ inputs.rustfs_version }}
          TEST_OUTCOME: ${{ steps.test.outcome }}
          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
          TRIGGER_NAME: ${{ github.event_name }}
        run: |
          set -euo pipefail
          PACKAGE_URL="${PACKAGE_URL_INPUT}"
          RUSTFS_VERSION="${RUSTFS_VERSION_INPUT}"
          if [ -n "${PACKAGE_URL}" ]; then
            PACKAGE_SOURCE="${PACKAGE_URL}"
          elif [ -n "${RUSTFS_VERSION}" ]; then
            PACKAGE_SOURCE="version ${RUSTFS_VERSION}"
          else
            PACKAGE_SOURCE="${RUSTFS_NIGHTLY_PACKAGE_URL}"
          fi
          set +e
          python3 auto-testing/rustfs_tier_report.py \
            --results-dir "${TIER_ARTIFACTS_DIR}/cases" \
            --provenance "${TIER_ARTIFACTS_DIR}/provenance.json" \
            --output "${CASE_TABLE}"
          CASE_GATE_RC=$?
          set -e
          printf '%s\n' "${CASE_GATE_RC}" > "${GATE_RC_FILE}"
          if [ ! -s "${CASE_TABLE}" ]; then
            {
              echo "## Case Summary"
              echo ""
              echo "Structured report generation failed before producing output (exit ${CASE_GATE_RC})."
            } > "${CASE_TABLE}"
          fi
          {
            echo "# RustFS tier test report"
            echo ""
            echo "- Run: ${RUN_URL}"
            echo "- Trigger: ${TRIGGER_NAME}"
            echo "- Package: ${PACKAGE_SOURCE}"
            echo "- Test Step Outcome: ${TEST_OUTCOME}"
            echo "- Structured Gate Exit: ${CASE_GATE_RC}"
            echo ""
            cat "${CASE_TABLE}"
            echo ""
            echo "## Log tail"
            echo '```text'
            tail -n 200 "${LOG_FILE}" || true
            echo '```'
          } | tee "${REPORT_FILE}"
          cat "${REPORT_FILE}" >> "${GITHUB_STEP_SUMMARY}"

      - name: Upload functional report to dashboard
        if: always()
        continue-on-error: true
        env:
          GH_TOKEN: ${{ env.PF_TESTING_GH_TOKEN }}
          REPORT_FILE: /tmp/rustfs-tier-report.md
          SUITE: tier
        run: |
          set -euo pipefail
          if [ -z "${GH_TOKEN:-}" ]; then
            echo "PF_TESTING_GH_TOKEN is not configured; skipping dashboard upload"
            exit 0
          fi
          DATE="$(date -u +%Y-%m-%d)"
          REPORT_PATH="functional-reports/${SUITE}/${DATE}.md"
          CONTENT="$(python3 -c 'import base64,sys;print(base64.b64encode(open(sys.argv[1],"rb").read()).decode())' "${REPORT_FILE}")"
          SHA="$(gh api "repos/rustfs/dashboard/contents/${REPORT_PATH}" -q '.sha' 2>/dev/null || true)"
          if [ -n "${SHA}" ]; then
            jq -n --arg msg "report(${SUITE}): ${DATE}" --arg content "${CONTENT}" --arg sha "${SHA}" \
              '{message:$msg, content:$content, sha:$sha}' \
              | gh api --method PUT "repos/rustfs/dashboard/contents/${REPORT_PATH}" --input - >/dev/null
          else
            jq -n --arg msg "report(${SUITE}): ${DATE}" --arg content "${CONTENT}" \
              '{message:$msg, content:$content}' \
              | gh api --method PUT "repos/rustfs/dashboard/contents/${REPORT_PATH}" --input - >/dev/null
          fi

      - name: File failure issue in rustfs/backlog
        if: ${{ always() && (failure() || steps.test.outcome == 'failure' || steps.test.outcome == 'cancelled') }}
        continue-on-error: true
        env:
          GH_TOKEN: ${{ secrets.PF_TESTING_GH_TOKEN }}
          SUITE: 'tier'
          SUITE_LABEL: 'Tier'
          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
          REPORT_FILE: '/tmp/rustfs-tier-report.md'
          LOG_FILE: '/tmp/rustfs-tier.log'
        run: |
          set -euo pipefail
          if [ -z "${GH_TOKEN:-}" ]; then
            echo "PF_TESTING_GH_TOKEN is not configured; skipping backlog issue"
            exit 0
          fi
          TITLE="[functional][${SUITE}] ${SUITE_LABEL} suite failed (run ${GITHUB_RUN_ID})"
          EXISTING="$(gh issue list -R rustfs/backlog --state all \
            --search "in:title \"run ${GITHUB_RUN_ID}\"" \
            --json number --jq '.[].number' || true)"
          if [ -n "${EXISTING}" ]; then
            echo "backlog issue already exists for run ${GITHUB_RUN_ID}; skipping"
            exit 0
          fi
          redact() {
            sed -E \
              -e 's/(RUSTFS_(ACCESS_KEY|SECRET_KEY)[=: ]+)[^[:space:]]+/\1[REDACTED]/Ig' \
              -e 's/(Authorization:).*/\1 [REDACTED]/Ig' \
              -e 's/(X-Amz-Signature=)[^&[:space:]]+/\1[REDACTED]/Ig' \
              -e 's/^.*(password|secret|token)[=: ].*/[REDACTED SENSITIVE LINE]/Ig'
          }
          BODY_FILE="$(mktemp)"
          {
            echo "The **${SUITE_LABEL}** functional suite failed."
            echo ""
            echo "- Suite: \`${SUITE}\`"
            echo "- Run: ${RUN_URL}"
            echo "- Trigger: ${GITHUB_EVENT_NAME}"
            echo "- Date: $(date -u +%Y-%m-%d)"
            echo ""
            echo "## Report (errors and symptoms)"
            echo ""
            if [ -s "${REPORT_FILE}" ]; then
              redact < "${REPORT_FILE}"
            elif [ -s "${LOG_FILE:-}" ]; then
              echo "(report file missing; log tail below)"
              echo ""
              tail -n 200 "${LOG_FILE}" | redact
            else
              echo "(no report or log file was produced)"
            fi
          } | head -c 55000 > "${BODY_FILE}"
          gh label create functional-test -R rustfs/backlog --color d73a4a 2>/dev/null || true
          if ! gh issue create -R rustfs/backlog --title "${TITLE}" \
              --body-file "${BODY_FILE}" --label functional-test; then
            gh issue create -R rustfs/backlog --title "${TITLE}" --body-file "${BODY_FILE}"
          fi
          echo "filed backlog issue for suite ${SUITE}"

      - name: Upload report and logs
        if: always()
        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f # v6
        with:
          name: rustfs-tier-test-${{ github.run_id }}-${{ github.run_attempt }}
          path: |
            /tmp/rustfs-tier.log
            /tmp/rustfs-tier-report.md
            /tmp/rustfs-tier-cases.md
            /tmp/rustfs-tier-gate.rc
            ${{ env.TIER_ARTIFACTS_DIR }}/
          if-no-files-found: error

      - name: Cleanup environment (after)
        if: always()
        run: |
          set -euo pipefail
          sudo docker rm -f rustfs-test-mqtt >/dev/null 2>&1 || true
          sudo rm -f /tmp/rustfs-mosquitto.conf
          read -r -a NODES <<< "${RUSTFS_NODES:-vm000 vm001 vm002}"
          SSH_USER="${RUSTFS_SSH_USER:-azureuser}"
          for node in "${NODES[@]}"; do
            ssh -o BatchMode=yes -o ConnectTimeout=10 -o StrictHostKeyChecking=accept-new "${SSH_USER}@${node}" '
              set -euo pipefail
              SUDO=""; [ "$(id -u)" -ne 0 ] && SUDO="sudo -n"
              ${SUDO} systemctl stop rustfs 2>/dev/null || true
              if ${SUDO} dpkg -l rustfs 2>/dev/null | grep -q "^ii"; then
                ${SUDO} dpkg -P rustfs
              fi
              for i in 1 2 3 4; do ${SUDO} rm -rf /data/rustfs${i}/mnmd; done
              ${SUDO} rm -rf /var/log/rustfs /var/lib/rustfs/kms
            '
          done

      - name: Enforce tier suite result
        if: always()
        env:
          TEST_OUTCOME: ${{ steps.test.outcome }}
          GATE_RC_FILE: /tmp/rustfs-tier-gate.rc
        run: |
          set -euo pipefail
          failed=0
          if [ "${TEST_OUTCOME}" != "success" ]; then
            echo "tier suite step outcome is ${TEST_OUTCOME}, expected success" >&2
            failed=1
          fi
          if [ ! -s "${GATE_RC_FILE}" ]; then
            echo "structured gate result is missing" >&2
            failed=1
          else
            GATE_RC="$(tr -d '[:space:]' < "${GATE_RC_FILE}")"
            if ! [[ "${GATE_RC}" =~ ^[0-9]+$ ]] || [ "${GATE_RC}" -ne 0 ]; then
              echo "structured 56-case gate failed with exit ${GATE_RC:-invalid}" >&2
              failed=1
            fi
          fi
          [ "${failed}" -eq 0 ]

      - name: "Continue functional chain (next: Storage engine)"
        # Only chain-triggered runs forward to the next suite; standalone
        # workflow_dispatch runs stop after their own cleanup.
        if: ${{ always() && github.event_name == 'repository_dispatch' }}
        continue-on-error: true
        env:
          GH_TOKEN: ${{ secrets.PF_TESTING_GH_TOKEN }}
        run: |
          set -euo pipefail
          if [ -z "${GH_TOKEN:-}" ]; then
            echo "PF_TESTING_GH_TOKEN is not configured; cannot dispatch the next suite" >&2
            exit 1
          fi
          echo "Dispatching next functional suite: Storage engine"
          gh api --method POST repos/rustfs/rustfs/dispatches \
            -f event_type='rustfs-chain-storage' \
            -F 'client_payload[from_suite]=tier'

      - name: Notify on failure
        if: failure()
        run: |
          echo "RustFS tier suite failed"
          echo "See the uploaded report and log artifacts for details."