如何为复刻的GitLab CE项目配置云端Ubuntu服务器持续部署（CD）？

Got it, let's walk through setting up a continuous deployment pipeline for your forked GitLab CE repo to a cloud Ubuntu server. I’ve tackled this exact scenario a few times, so here’s a practical, step-by-step breakdown that should work for your use case:

First, let’s get the basics sorted:

• Server Setup: Ensure your Ubuntu cloud server has the necessary ports open (80, 443, 22 for GitLab and SSH access). Create a dedicated deploy user (e.g., gitlab-deploy) with sudo access—never use root for deployments. Set up passwordless SSH access from your CI/CD runner to this user (generate an SSH key pair, add the public key to ~gitlab-deploy/.ssh/authorized_keys on the server).
• Repo Setup: Make sure your forked GitLab CE repo is hosted on a platform that supports CI/CD (GitLab.com, GitHub, or your own GitLab instance) and you have admin rights to configure pipeline variables.

I recommend two low-friction options, depending on your hosting platform:

• GitLab CI/CD (if hosted on GitLab): Built-in, no extra infrastructure needed—perfect for quick setup.
• GitHub Actions (if hosted on GitHub): Just as straightforward, with a huge library of pre-built actions.
• Jenkins (for custom workflows): Only use this if you need super complex, on-premise-compliant pipelines—it’s more maintenance-heavy.

We’ll focus on GitLab CI/CD first, since it’s native to GitLab CE, then cover the GitHub alternative.

Step 1: Add CI/CD Variables #

In your repo’s settings, go to CI/CD > Variables and add these sensitive values (mark them as "protected" and "masked"):

• SSH_PRIVATE_KEY: The private key for your gitlab-deploy user.
• SERVER_IP: Your cloud Ubuntu server’s public IP.

Step 2: Write the .gitlab-ci.yml File #

Drop this file in your repo root—tweak it to match your specific code changes:

stages:
  - build
  - deploy

# Stage 1: Compile GitLab assets and dependencies
build_gitlab:
  stage: build
  image: ruby:2.7  # Match the Ruby version your GitLab CE branch uses
  before_script:
    - apt-get update && apt-get install -y build-essential libyaml-dev libgdbm-dev libreadline-dev libncurses5-dev libffi-dev zlib1g-dev
    - gem install bundler
    - bundle install --without development test  # Skip dev/test dependencies for speed
  script:
    - bundle exec rake gitlab:assets:compile RAILS_ENV=production  # Compile frontend assets
    # Add any custom build commands here if your changes require extra compilation
  artifacts:
    paths:
      - public/  # Compiled static assets
      - app/     # Modified Ruby code (adjust this based on your changes)
    expire_in: 1 week  # Clean up old artifacts automatically

# Stage 2: Deploy to the Ubuntu server
deploy_to_server:
  stage: deploy
  image: alpine:latest  # Lightweight image for SSH/rsync
  before_script:
    - apk add --no-cache openssh-client rsync
    - mkdir -p ~/.ssh
    - echo "$SSH_PRIVATE_KEY" > ~/.ssh/id_rsa
    - chmod 600 ~/.ssh/id_rsa  # Critical for SSH security
    - ssh-keyscan -H "$SERVER_IP" >> ~/.ssh/known_hosts  # Avoid host key prompts
  script:
    # 1. Stop GitLab to prevent file conflicts during deployment
    - ssh gitlab-deploy@"$SERVER_IP" "sudo gitlab-ctl stop"
    # 2. Sync compiled assets and modified code to the server
    - rsync -avz --exclude='.git' ./public/ gitlab-deploy@"$SERVER_IP":/opt/gitlab/embedded/service/gitlab-rails/public/
    - rsync -avz --exclude='.git' ./app/ gitlab-deploy@"$SERVER_IP":/opt/gitlab/embedded/service/gitlab-rails/app/
    # 3. Reconfigure and restart GitLab to apply changes
    - ssh gitlab-deploy@"$SERVER_IP" "sudo gitlab-ctl reconfigure && sudo gitlab-ctl start"
    # 4. Verify deployment success
    - ssh gitlab-deploy@"$SERVER_IP" "sudo gitlab-ctl status"
  only:
    - main  # Only trigger deployment when pushing to your main branch

If your repo is on GitHub, create .github/workflows/deploy.yml instead:

name: Deploy Forked GitLab CE to Ubuntu Server

on:
  push:
    branches: [ main ]

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout code
        uses: actions/checkout@v4

      - name: Set up Ruby
        uses: ruby/setup-ruby@v1
        with:
          ruby-version: '2.7'
          bundler-cache: true  # Cache dependencies to speed up builds

      - name: Compile GitLab assets
        run: |
          sudo apt-get install -y build-essential libyaml-dev libgdbm-dev libreadline-dev libncurses5-dev libffi-dev zlib1g-dev
          bundle exec rake gitlab:assets:compile RAILS_ENV=production

      - name: Deploy to Ubuntu server
        uses: appleboy/ssh-action@v1.0.3
        with:
          host: ${{ secrets.SERVER_IP }}
          username: gitlab-deploy
          key: ${{ secrets.SSH_PRIVATE_KEY }}
          script: |
            sudo gitlab-ctl stop
            rsync -avz --exclude='.git' $GITHUB_WORKSPACE/public/ /opt/gitlab/embedded/service/gitlab-rails/public/
            rsync -avz --exclude='.git' $GITHUB_WORKSPACE/app/ /opt/gitlab/embedded/service/gitlab-rails/app/
            sudo gitlab-ctl reconfigure && sudo gitlab-ctl start
            sudo gitlab-ctl status

• Backup First: Always run sudo gitlab-backup create on the server before deploying—this saves you if something breaks during sync.
• Version Matching: Ensure the GitLab CE version in your fork matches the one installed on the server. Mismatched versions will cause weird bugs.
• Incremental Sync: Use rsync instead of full copies—it only transfers changed files, cutting down deployment time.
• Monitor Logs: After deployment, check GitLab logs with sudo gitlab-ctl tail to catch any errors early.
• Rollback Plan: Keep a copy of the previous working version on the server, or use the backup to roll back if deployment fails.

内容的提问来源于stack exchange，提问作者Scott Ge