Swallow

Swallow is a library that helps make it easier to deploy projects with Capistrano to AWS.

New Server Setup

1. Have systems create a new vanilla web image.
2. Run $ cap deploy:init to initialize the server. This will:
   • Install RBEnv (rbenv:init)

New Project Setup

1. Copy the default Capfile from the Swallow project to the project's root.
2. Copy the default config/deploy.yml to the project.
3. Copy and configure the default config/newrelic.yml, or in config/deploy.yml set use_newrelic: false.
4. Edit deploy.yml to match the project's needs. See the config section to get an understanding for what is needed and what is not.
5. If the project will use unicorn, follow the setting up unicorn instructions.
6. Run $ cap deploy:setup and select the targeted environment. This will:
   • Create the project's main directory structure
   • Install the project's specified ruby version (rbenv:setup_ruby)
   • Create the shared unicorn (unicorn:setup_sockets_dir)
7. Run $ cap deploy:cold. This will:
   • Install the project's version of ruby if it's not already installed (rbenv:setup_ruby)
   • Pull down the latest code from GitHub
   • Properly link the standard shared directories to the project (i.e. /log, /tmp/pids)
   • Link the unicron sockets directory (unicorn:create_socket_dir)
   • Install bundler if its not already installed (bundler:setup)
   • Install required gems (bundler:install)
   • Copy over the database and any other config files (deploy:copy_database_configuration and deploy:copy_configs)
   • Create the project's deploy tag (deploy:tag)
   • Copy previously generated & deployed assets if there are no asset changes, otherwise it will precompile & sync assets (assets:check)
   • Run the airbrake & New Relic deploy notifications (airbrake:notice_deployment and newrelic:notice_deployment)
   • Run the database migrations (deploy:migrate)