production-suite.yml

name: Zulip production suite

on:
  push:
    branches: ["*.x", chat.zulip.org, main]
    tags: ["*"]
  pull_request:
    paths:
      - .github/workflows/production-suite.yml
      - "**/migrations/**"
      - manage.py
      - pnpm-lock.yaml
      - puppet/**
      - requirements/**
      - scripts/**
      - tools/**
      - web/babel.config.js
      - web/postcss.config.js
      - web/third/**
      - web/webpack.config.ts
      - zerver/worker/queue_processors.py
      - zerver/lib/push_notifications.py
      - zerver/decorator.py
      - zproject/**
  workflow_dispatch:

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

defaults:
  run:
    shell: bash

permissions:
  contents: read

jobs:
  production_build:
    # This job builds a release tarball from the current commit, which
    # will be used for all of the following install/upgrade tests.
    name: Ubuntu 20.04 production build
    runs-on: ubuntu-latest

    # Docker images are built from 'tools/ci/Dockerfile'; the comments at
    # the top explain how to build and upload these images.
    # Ubuntu 20.04 ships with Python 3.8.10.
    container: zulip/ci:focal

    steps:
      - name: Add required permissions
        run: |
          # The checkout actions doesn't clone to ~/zulip or allow
          # us to use the path option to clone outside the current
          # /__w/zulip/zulip directory. Since this directory is owned
          # by root we need to change it's ownership to allow the
          # github user to clone the code here.
          # Note: /__w/ is a docker volume mounted to $GITHUB_WORKSPACE
          # which is /home/runner/work/.
          sudo chown -R github .

          # This is the GitHub Actions specific cache directory the
          # the current github user must be able to access for the
          # cache action to work. It is owned by root currently.
          sudo chmod -R 0777 /__w/_temp/

      - uses: actions/checkout@v3

      - name: Create cache directories
        run: |
          dirs=(/srv/zulip-{venv,emoji}-cache)
          sudo mkdir -p "${dirs[@]}"
          sudo chown -R github "${dirs[@]}"

      - name: Restore pnpm store
        uses: actions/cache@v3
        with:
          path: /__w/.pnpm-store
          key: v1-pnpm-store-focal-${{ hashFiles('pnpm-lock.yaml') }}

      - name: Restore python cache
        uses: actions/cache@v3
        with:
          path: /srv/zulip-venv-cache
          key: v1-venv-focal-${{ hashFiles('requirements/dev.txt') }}
          restore-keys: v1-venv-focal

      - name: Restore emoji cache
        uses: actions/cache@v3
        with:
          path: /srv/zulip-emoji-cache
          key: v1-emoji-focal-${{ hashFiles('tools/setup/emoji/emoji_map.json') }}-${{ hashFiles('tools/setup/emoji/build_emoji') }}-${{ hashFiles('tools/setup/emoji/emoji_setup_utils.py') }}-${{ hashFiles('tools/setup/emoji/emoji_names.py') }}-${{ hashFiles('package.json') }}
          restore-keys: v1-emoji-focal

      - name: Build production tarball
        run: ./tools/ci/production-build

      - name: Upload production build artifacts for install jobs
        uses: actions/upload-artifact@v3
        with:
          name: production-tarball
          path: /tmp/production-build
          retention-days: 1

      - name: Verify pnpm store path
        run: |
          set -x
          path="$(pnpm store path)"
          [[ "$path" == /__w/.pnpm-store/* ]]

      - name: Generate failure report string
        id: failure_report_string
        if: ${{ failure() && github.repository == 'zulip/zulip' && github.event_name == 'push' }}
        run: tools/ci/generate-failure-message >> $GITHUB_OUTPUT

      - name: Report status to CZO
        if: ${{ failure() && github.repository == 'zulip/zulip' && github.event_name == 'push' }}
        uses: zulip/github-actions-zulip/send-message@v1
        with:
          api-key: ${{ secrets.ZULIP_BOT_KEY }}
          email: "github-actions-bot@chat.zulip.org"
          organization-url: "https://chat.zulip.org"
          to: "automated testing"
          topic: ${{ steps.failure_report_string.outputs.topic }}
          type: "stream"
          content: ${{ steps.failure_report_string.outputs.content }}

  production_install:
    # This job installs the server release tarball built above on a
    # range of platforms, and does some basic health checks on the
    # resulting installer Zulip server.
    strategy:
      fail-fast: false
      matrix:
        include:
          # Docker images are built from 'tools/ci/Dockerfile'; the comments at
          # the top explain how to build and upload these images.
          - docker_image: zulip/ci:focal
            name: Ubuntu 20.04 production install and PostgreSQL upgrade with pgroonga
            os: focal
            extra-args: ""

          - docker_image: zulip/ci:jammy
            name: Ubuntu 22.04 production install
            os: jammy
            extra-args: ""

          - docker_image: zulip/ci:noble
            name: Ubuntu 24.04 production install
            os: noble
            extra-args: ""

          - docker_image: zulip/ci:bullseye
            name: Debian 11 production install with custom db name and user
            os: bullseye
            extra-args: --test-custom-db

          - docker_image: zulip/ci:bookworm
            name: Debian 12 production install
            os: bookworm
            extra-args: ""

    name: ${{ matrix.name  }}
    container:
      image: ${{ matrix.docker_image }}
      options: --init
    runs-on: ubuntu-latest
    needs: production_build

    steps:
      - name: Download built production tarball
        uses: actions/download-artifact@v3
        with:
          name: production-tarball
          path: /tmp

      - name: Add required permissions and setup
        run: |
          # This is the GitHub Actions specific cache directory the
          # the current github user must be able to access for the
          # cache action to work. It is owned by root currently.
          sudo chmod -R 0777 /__w/_temp/

          # Since actions/download-artifact@v2 loses all the permissions
          # of the tarball uploaded by the upload artifact fix those.
          chmod +x /tmp/production-upgrade-pg
          chmod +x /tmp/production-pgroonga
          chmod +x /tmp/production-install
          chmod +x /tmp/production-verify
          chmod +x /tmp/generate-failure-message

      - name: Create cache directories
        run: |
          dirs=(/srv/zulip-{venv,emoji}-cache)
          sudo mkdir -p "${dirs[@]}"
          sudo chown -R github "${dirs[@]}"

      - name: Install production
        run: sudo /tmp/production-install ${{ matrix.extra-args }}

      - name: Verify install
        run: sudo /tmp/production-verify ${{ matrix.extra-args }}

      - name: Install pgroonga
        if: ${{ matrix.os == 'focal' }}
        run: sudo /tmp/production-pgroonga

      - name: Verify install after installing pgroonga
        if: ${{ matrix.os == 'focal' }}
        run: sudo /tmp/production-verify ${{ matrix.extra-args }}

      - name: Upgrade postgresql
        if: ${{ matrix.os == 'focal' }}
        run: sudo /tmp/production-upgrade-pg

      - name: Verify install after upgrading postgresql
        if: ${{ matrix.os == 'focal' }}
        run: sudo /tmp/production-verify ${{ matrix.extra-args }}

      - name: Generate failure report string
        id: failure_report_string
        if: ${{ failure() && github.repository == 'zulip/zulip' && github.event_name == 'push' }}
        run: /tmp/generate-failure-message >> $GITHUB_OUTPUT

      - name: Report status to CZO
        if: ${{ failure() && github.repository == 'zulip/zulip' && github.event_name == 'push' }}
        uses: zulip/github-actions-zulip/send-message@v1
        with:
          api-key: ${{ secrets.ZULIP_BOT_KEY }}
          email: "github-actions-bot@chat.zulip.org"
          organization-url: "https://chat.zulip.org"
          to: "automated testing"
          topic: ${{ steps.failure_report_string.outputs.topic }}
          type: "stream"
          content: ${{ steps.failure_report_string.outputs.content }}

  production_upgrade:
    # The production upgrade job starts with a container with a
    # previous Zulip release installed, and attempts to upgrade it to
    # the release tarball built for the current commit being tested.
    #
    # This is intended to catch bugs that result in the upgrade
    # process failing.
    strategy:
      fail-fast: false
      matrix:
        include:
          # Docker images are built from 'tools/ci/Dockerfile.prod'; the comments at
          # the top explain how to build and upload these images.
          - docker_image: zulip/ci:focal-3.2
            name: 3.2 Version Upgrade
            os: focal
          - docker_image: zulip/ci:bullseye-4.2
            name: 4.2 Version Upgrade
            os: bullseye
          - docker_image: zulip/ci:bullseye-5.0
            name: 5.0 Version Upgrade
            os: bullseye
          - docker_image: zulip/ci:jammy-6.0
            name: 6.0 Version Upgrade
            os: jammy
          - docker_image: zulip/ci:bookworm-7.0
            name: 7.0 Version Upgrade
            os: bookworm
          - docker_image: zulip/ci:bookworm-8.0
            name: 8.0 Version Upgrade
            os: bookworm

    name: ${{ matrix.name  }}
    container:
      image: ${{ matrix.docker_image }}
      options: --init
    runs-on: ubuntu-latest
    needs: production_build

    steps:
      - name: Download built production tarball
        uses: actions/download-artifact@v3
        with:
          name: production-tarball
          path: /tmp

      - name: Add required permissions and setup
        run: |
          # This is the GitHub Actions specific cache directory the
          # the current github user must be able to access for the
          # cache action to work. It is owned by root currently.
          sudo chmod -R 0777 /__w/_temp/

          # Since actions/download-artifact@v2 loses all the permissions
          # of the tarball uploaded by the upload artifact fix those.
          chmod +x /tmp/production-upgrade
          chmod +x /tmp/production-verify
          chmod +x /tmp/generate-failure-message

      - name: Create cache directories
        run: |
          dirs=(/srv/zulip-{venv,emoji}-cache)
          sudo mkdir -p "${dirs[@]}"
          sudo chown -R github "${dirs[@]}"

      - name: Temporarily bootstrap PostgreSQL upgrades
        # https://chat.zulip.org/#narrow/stream/43-automated-testing/topic/postgres.20client.20upgrade.20failures/near/1640444
        # On Debian, there is an ordering issue with post-install maintainer
        # scripts when postgresql-client-common is upgraded at the same time as
        # postgresql-client and postgresql-client-15.  Upgrade just
        # postgresql-client-common first, so the main upgrade process can
        # succeed.  This is a _temporary_ work-around to improve CI signal, as
        # the failure does represent a real failure that production systems may
        # encounter.
        run: sudo apt-get update && sudo apt-get install -y --only-upgrade postgresql-client-common

      - name: Upgrade production
        run: sudo /tmp/production-upgrade

        # TODO: We should be running production-verify here, but it
        # doesn't pass yet.
        #
        # - name: Verify install
        #   run: sudo /tmp/production-verify

      - name: Generate failure report string
        id: failure_report_string
        if: ${{ failure() && github.repository == 'zulip/zulip' && github.event_name == 'push' }}
        run: /tmp/generate-failure-message >> $GITHUB_OUTPUT

      - name: Report status to CZO
        if: ${{ failure() && github.repository == 'zulip/zulip' && github.event_name == 'push' }}
        uses: zulip/github-actions-zulip/send-message@v1
        with:
          api-key: ${{ secrets.ZULIP_BOT_KEY }}
          email: "github-actions-bot@chat.zulip.org"
          organization-url: "https://chat.zulip.org"
          to: "automated testing"
          topic: ${{ steps.failure_report_string.outputs.topic }}
          type: "stream"
          content: ${{ steps.failure_report_string.outputs.content }}