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

name: RustFS Heal Test

on:
  workflow_dispatch:
    inputs:
      package_url:
        description: 'Direct .deb URL (nightly/R2). Defaults to the latest nightly deb.'
        required: false
        type: string
      stop_node_gb:
        description: 'Stop the outage node when surviving nodes reach N GiB'
        required: false
        default: '15'
      warp_stop_gb:
        description: 'Stop warp when surviving nodes reach N GiB'
        required: false
        default: '40'
      cleanup_before:
        description: 'Reset the nodes before the test (DESTROYS existing data/config)'
        type: boolean
        default: true
      cleanup_after:
        description: 'Reset the nodes after the test (DESTROYS test data/config)'
        type: boolean
        default: true

permissions:
  contents: read

# Only one test at a time: both this and the pool-expansion workflow mutate
# the same test environment, so they share one concurrency group.
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_API_ENDPOINT: ${{ secrets.RUSTFS_API_ENDPOINT || vars.RUSTFS_API_ENDPOINT || vars.RUSTFS_RC_ENDPOINT }}
  RUSTFS_NODES: ${{ secrets.RUSTFS_NODES || vars.RUSTFS_NODES }}
  RUSTFS_SSH_USER: ${{ secrets.RUSTFS_SSH_USER || vars.RUSTFS_SSH_USER }}
  PF_TESTING_GH_TOKEN: ${{ secrets.PF_TESTING_GH_TOKEN }}
  RUSTFS_NIGHTLY_PACKAGE_URL: ${{ vars.RUSTFS_NIGHTLY_PACKAGE_URL || 'https://dl.rustfs.com/artifacts/rustfs/packages/nightly/rustfs-nightly-latest.deb' }}

jobs:
  heal-test:
    runs-on: smoke-testing
    timeout-minutes: 480
    # Manual-only standalone run. Nightly chain already runs heal in
    # rustfs-pool-expand-test.yml to avoid duplicate heal executions.
    if: ${{ github.event_name == 'workflow_dispatch' }}
    steps:
      - name: Checkout auto-testing scripts
        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
        with:
          repository: rustfs/auto-testing
          ref: main
          path: auto-testing
          persist-credentials: false
          token: ${{ secrets.PF_TESTING_GH_TOKEN }}

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

      - name: Cleanup environment (before)
        if: ${{ inputs.cleanup_before != 'false' }}
        run: |
          set -euo pipefail
          read -r -a NODES <<< "${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 /var/lib/rustfs/kms-backup
            '
          done

      - name: Install RustFS package & start cluster
        run: |
          ARGS=(--steps "1,2" -y --endpoint "${{ env.RUSTFS_API_ENDPOINT }}")
          if [ -n "${{ inputs.package_url }}" ]; then
            ARGS+=(--package-url "${{ inputs.package_url }}")
          else
            ARGS+=(--package-url "${{ env.RUSTFS_NIGHTLY_PACKAGE_URL }}")
          fi
          ./auto-testing/rustfs_heal_test.sh "${ARGS[@]}"

      - name: Preflight checks
        run: |
          ARGS=(--preflight --endpoint "${{ env.RUSTFS_API_ENDPOINT }}")
          if [ -n "${{ inputs.package_url }}" ]; then
            ARGS+=(--package-url "${{ inputs.package_url }}")
          else
            ARGS+=(--package-url "${{ env.RUSTFS_NIGHTLY_PACKAGE_URL }}")
          fi
          ./auto-testing/rustfs_heal_test.sh "${ARGS[@]}"

      - name: Run heal test (write -> outage -> heal -> verify)
        id: test
        run: |
          ./auto-testing/rustfs_heal_test.sh \
            --steps "3,4,5,6,7" -y \
            --endpoint "${{ env.RUSTFS_API_ENDPOINT }}" \
            --stop-node-gb "${{ inputs.stop_node_gb }}" \
            --warp-stop-gb "${{ inputs.warp_stop_gb }}" \
            --log-file /tmp/rustfs-heal-test.log

      - name: Generate report
        if: always()
        env:
          LOG_FILE: /tmp/rustfs-heal-test.log
          REPORT_FILE: /tmp/rustfs-heal-report.md
        run: |
          set -euo pipefail
          PACKAGE_URL='${{ inputs.package_url }}'
          if [ -n "${PACKAGE_URL}" ]; then
            PACKAGE_SOURCE="${PACKAGE_URL}"
          else
            PACKAGE_SOURCE="${RUSTFS_NIGHTLY_PACKAGE_URL}"
          fi
          {
            echo "# RustFS heal test report"
            echo ""
            echo "- Run: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
            echo "- Trigger: ${{ github.event_name }}"
            echo "- Package: ${PACKAGE_SOURCE}"
            echo "- Test Step Outcome: ${{ steps.test.outcome }}"
            echo ""
            echo "## Log tail"
            echo '```text'
            tail -n 200 "${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: ${{ env.PF_TESTING_GH_TOKEN }}
          REPORT_FILE: /tmp/rustfs-heal-report.md
          SUITE: heal
        run: |
          set -euo pipefail
          if [ -z "${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

          cat > /tmp/rustfs-functional-index.html <<'EOF'
          <!doctype html>
          <html lang="en">
          <head>
            <meta charset="utf-8" />
            <meta name="viewport" content="width=device-width, initial-scale=1" />
            <title>RustFS Functional Test Reports</title>
            <style>
              :root { --bg:#f4f6fb; --card:#fff; --text:#1f2937; --muted:#6b7280; --line:#e5e7eb; --accent:#0f766e; }
              * { box-sizing: border-box; }
              body { margin: 0; font-family: ui-sans-serif, -apple-system, Segoe UI, Helvetica, Arial, sans-serif; background: var(--bg); color: var(--text); }
              .wrap { max-width: 980px; margin: 32px auto; padding: 0 16px; }
              .card { background: var(--card); border: 1px solid var(--line); border-radius: 14px; padding: 20px; }
              h1 { margin: 0 0 8px; font-size: 26px; }
              p { margin: 0 0 14px; color: var(--muted); }
              .tabs { display: flex; gap: 10px; margin: 14px 0 18px; flex-wrap: wrap; }
              button { border: 1px solid var(--line); background: #fff; color: var(--text); border-radius: 10px; padding: 8px 14px; cursor: pointer; }
              button.active { background: var(--accent); color: #fff; border-color: var(--accent); }
              .report-btn { border: 0; background: transparent; padding: 0; color: var(--accent); }
              ul { list-style: none; margin: 0; padding: 0; }
              li { padding: 10px 0; border-bottom: 1px solid var(--line); }
              a { color: var(--accent); text-decoration: none; }
              a:hover { text-decoration: underline; }
              .meta { margin-top: 16px; border-top: 1px solid var(--line); padding-top: 14px; }
              .kv { margin: 6px 0; color: var(--text); }
              .muted { color: var(--muted); }
            </style>
          </head>
          <body>
            <div class="wrap">
              <div class="card">
                <h1>RustFS Functional Test Reports</h1>
                <p>Select a suite and date to view the build version used in that run.</p>
                <div class="tabs" id="tabs"></div>
                <ul id="list"></ul>
                <div class="meta">
                  <div class="kv"><strong>Date:</strong> <span id="report-date" class="muted">N/A</span></div>
                  <div class="kv"><strong>RustFS Version:</strong> <span id="report-version" class="muted">N/A</span></div>
                  <div class="kv"><a id="report-link" href="#" target="_blank" rel="noreferrer">Open report</a></div>
                </div>
              </div>
            </div>
            <script>
              const suites = [
                { key: 's3', label: 'S3 Compatibility' },
                { key: 'kms', label: 'KMS' },
                { key: 'tier', label: 'Tier' },
                { key: 'heal', label: 'Heal' },
                { key: 'pool', label: 'Pool Expansion' },
                { key: 'security', label: 'Security' },
                { key: 'upgrade', label: 'Upgrade' },
              ];
              const tabs = document.getElementById('tabs');
              const list = document.getElementById('list');
              const reportDate = document.getElementById('report-date');
              const reportVersion = document.getElementById('report-version');
              const reportLink = document.getElementById('report-link');

              function parseVersion(markdown) {
                const m = markdown.match(/^- RustFS Version:\s*(.+)$/m);
                return m ? m[1].trim() : 'N/A';
              }

              async function showReport(report) {
                reportDate.textContent = report.name.replace('.md', '');
                reportVersion.textContent = 'Loading...';
                reportLink.href = report.html_url;
                try {
                  const res = await fetch(report.download_url, { cache: 'no-store' });
                  if (!res.ok) {
                    reportVersion.textContent = 'N/A';
                    return;
                  }
                  const text = await res.text();
                  reportVersion.textContent = parseVersion(text);
                } catch (_e) {
                  reportVersion.textContent = 'N/A';
                }
              }

              async function loadSuite(suite) {
                list.innerHTML = '<li>Loading...</li>';
                const api = `https://api.github.com/repos/rustfs/dashboard/contents/functional-reports/${suite}`;
                try {
                  const res = await fetch(api);
                  if (!res.ok) {
                    list.innerHTML = '<li>No reports yet.</li>';
                    return;
                  }
                  const data = await res.json();
                  const files = data.filter(f => f.type === 'file' && f.name.endsWith('.md')).sort((a,b) => b.name.localeCompare(a.name));
                  if (!files.length) {
                    list.innerHTML = '<li>No reports yet.</li>';
                    reportDate.textContent = 'N/A';
                    reportVersion.textContent = 'N/A';
                    reportLink.href = '#';
                    return;
                  }
                  list.innerHTML = '';
                  files.forEach((f) => {
                    const li = document.createElement('li');
                    const btn = document.createElement('button');
                    btn.className = 'report-btn';
                    btn.textContent = f.name.replace('.md', '');
                    btn.addEventListener('click', () => showReport(f));
                    li.appendChild(btn);
                    list.appendChild(li);
                  });
                  showReport(files[0]);
                } catch (_e) {
                  list.innerHTML = '<li>Failed to load reports.</li>';
                  reportDate.textContent = 'N/A';
                  reportVersion.textContent = 'N/A';
                  reportLink.href = '#';
                }
              }

              function setActive(key) {
                for (const btn of tabs.querySelectorAll('button')) {
                  btn.classList.toggle('active', btn.dataset.key === key);
                }
                loadSuite(key);
              }

              for (const suite of suites) {
                const btn = document.createElement('button');
                btn.textContent = suite.label;
                btn.dataset.key = suite.key;
                btn.addEventListener('click', () => setActive(suite.key));
                tabs.appendChild(btn);
              }
              setActive('s3');
            </script>
          </body>
          </html>
          EOF

          INDEX_PATH="functional/index.html"
          INDEX_CONTENT="$(python3 -c 'import base64;print(base64.b64encode(open("/tmp/rustfs-functional-index.html","rb").read()).decode())')"
          INDEX_SHA="$(gh api "repos/rustfs/dashboard/contents/${INDEX_PATH}" -q '.sha' 2>/dev/null || true)"
          if [ -n "${INDEX_SHA}" ]; then
            jq -n --arg msg "functional ui update" --arg content "${INDEX_CONTENT}" --arg sha "${INDEX_SHA}" \
              '{message:$msg, content:$content, sha:$sha}' \
              | gh api --method PUT "repos/rustfs/dashboard/contents/${INDEX_PATH}" --input - >/dev/null
          else
            jq -n --arg msg "functional ui init" --arg content "${INDEX_CONTENT}" \
              '{message:$msg, content:$content}' \
              | gh api --method PUT "repos/rustfs/dashboard/contents/${INDEX_PATH}" --input - >/dev/null
          fi

      - name: Upload test logs
        if: always()
        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f # v6
        with:
          name: rustfs-heal-test-${{ github.run_id }}
          path: |
            /tmp/rustfs-heal-test*.log
            /tmp/rustfs-warp.*.log
          if-no-files-found: warn

      - name: Cleanup environment (after)
        if: ${{ always() && inputs.cleanup_after != 'false' }}
        run: |
          set -euo pipefail
          read -r -a NODES <<< "${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 /var/lib/rustfs/kms-backup
            '
          done

      - name: Notify on failure
        if: failure()
        run: |
          echo "RustFS heal test failed"
          echo "Package source: ${{ inputs.package_url || 'nightly (R2 latest)' }}"
          echo "See the uploaded log artifact for details."