Rails部署至服务器报错求助：执行cap production deploy后异常

Hey there! I’ve worked through countless Capistrano deployment hiccups with Rails apps, so let’s walk through the most reliable fixes when cap production deploy throws an error. Even without seeing the exact error message, these steps usually get things back on track:

• Confirm SSH Access & Server Permissions
  First, test if your local machine can SSH into the production server without a password: run ssh deploy@your-server-ip (replace with your actual user and server IP). If this prompts for a password, double-check that your local public SSH key is added to the server’s ~/.ssh/authorized_keys file for the deploy user.
  Also, make sure the deploy user has write access to your app’s directory (typically /var/www/your-app-name). If not, SSH into the server and run sudo chown -R deploy:deploy /var/www/your-app-name to fix permissions.

• Double-Check Capistrano Config Files
  Typos or misconfigs in config/deploy.rb and config/deploy/production.rb are super common culprits. Look for:

  • Incorrect server details in the server line (e.g., wrong IP, user name, or roles)
  • A mismatched Git repository URL (it should match what you use to pull code locally)
  • Missing entries in linked_dirs or linked_files (like public/uploads or .env—these files/folders need to persist across deploys)

• Update Bundler & Sync Dependencies
  Outdated Bundler versions on the server often break deployment. SSH in and run gem install bundler to update it. Also, make sure your Gemfile.lock is committed to Git—Capistrano relies on this to install the exact gem versions your app needs. If you’ve updated your Gemfile locally, run bundle install, commit the new Gemfile.lock, and try deploying again.

• Troubleshoot Database Migrations
  If the error hits during migrations, verify your production database user has permission to run migrations. Also, check that the database.yml file on the server (usually a linked file) has the correct credentials. You can test this directly on the server by running RAILS_ENV=production rails db:migrate—this will show you if there’s a connection issue or migration error.

• Ensure Your Rails Server Restarts
  Sometimes deploys “succeed” but the app doesn’t update because the server process (Puma, Unicorn, etc.) isn’t restarted. Make sure you have the right Capistrano gem for your server (e.g., capistrano-puma for Puma) and that it’s required in deploy.rb. If you need to manually restart, SSH into the server and run something like sudo systemctl restart puma (adjust the service name to match your setup).

• Dig Into Logs for Exact Error Details
  If nothing else works, check the logs! Your local Capistrano logs are in log/capistrano.log, and the Rails production logs on the server are in /var/www/your-app-name/shared/log/production.log. The exact error message here will tell you exactly what’s wrong—whether it’s a missing gem, a misconfigured environment variable, or a server resource issue.

内容的提问来源于stack exchange，提问作者Adamlin0708