capistrano_recipes

Bunch of capistrano recipes

GitHub: https://github.com/rubydev/capistrano_recipes

Installation

Add capistrano_recipes to your application's Gemfile:

gem 'capistrano_recipes', :git => 'git://github.com/rubydev/capistrano_recipes.git'

And then install the bundle:

$ bundle

Usage

First, initialize capistrano:

capify .

This will create ./Capfile and ./config/deploy.rb. Edit ./config/deploy.rb and load recipes required for your application:

require "capistrano_recipes/nginx"
require "capistrano_recipes/mysql"
require "capistrano_recipes/host"

Then run cap host:setup deploy:initial to make your first deploy. Use cap deploy:migrations for next deploys.

Example of rails app deployment

config/deploy.rb:

default_run_options[:pty] = true     # Must be set for the password prompt to work
set :use_sudo, false                 # Don't use sudo (deploy user must have very limited permissions in system)

set :default_stage, "staging"        # Deploy on staging server by default

# Configure capistrano deploy strategy
set :repository, '.'
set :deploy_via, :copy
set :copy_exclude, [".git", "coverage", "results", "tmp", "public/system", "builder"]

set :root_user, "user"               # User with root privileges on server. You will need him only for host:setup
                                     # and *:setup_host tasks. This will not be used during deploy,
                                     # deploy:migrations and other common tasks.
set :user, "app_name"                # Deployer user
set :group, "app_name"               # Deployer group
set :user_home_path, "/home/www"     # Deployer home on target system (used in host:setup when creating new user)
set(:deploy_to) { "#{user_home_path}/#{application}" } # Path where to deploy your application

role(:web) { domain }                         # Your HTTP server, Apache/etc
role(:app) { domain }                         # This may be the same as your `Web` server
role(:db, :primary => true) { domain }        # This is where Rails migrations will run

require "capistrano_colors"                   # Colorized output in console (gem install capistrano_colors)
require "capistrano/ext/multistage"           # Enable multistage deployment
require "capistrano_recipes/nginx"            # Use nginx as web-server
require "capistrano_recipes/mysql"            # Use mysql as db-server
require "capistrano_recipes/host"             # Require host:setup task
require "capistrano_recipes/monit"            # Use monit as monitoring system
require "capistrano_recipes/ruby/thin"        # Use thing as app-server
require "capistrano_recipes/ruby/rails"       # Load rails-specific recipes
require "bundler/capistrano"                  # Use bundler on server
load "deploy/assets"                          # Use rails assets pipeline

config/deploy/production.rb

set :application, "app_name"                  # Application name for production
                                              # (it's used in some configs and paths)
set :application_domain, "app_domain.com"     # Domain for production server
set :domain, "192.168.1.1"                    # Server address where to deploy production

set :rails_env, 'production'
set :branch, 'production'
set :thin_port, 4300
set :app_server, "thin"
set :nginx_use_ssl, true
set :thin_servers, 3

config/deploy/staging.rb

set :application, "app_name_test"             # Application name for staging server
set :application_domain, "qa.app_domain.com"  # Domain for staging server
set :domain, "192.168.1.1"                    # Server address where to deploy staging

set :rails_env, 'staging'
set :branch, 'staging'
set :thin_port, 4400
set :app_server, "thin"
set :nginx_use_ssl, false
set :thin_servers, 3

Then execute cap host:setup deploy:initial nginx:enable to create user and config files and perform initial deploy, then symlink /etc/nginx/sites-available/app_name.conf to /etc/nginx/sites-enabled/app_name.conf. On next deploy just run cap deploy:migrations to deploy new application version.

cap host:setup will perform:

• host:create_user - creates :user on target system (if doesn't exist yet)
• host:ssh_copy_id - adds ~/.ssh/id_rsa.pub from local machine to authorized_keys on target machine for :user
• Then it performs *:setup_host for all loaded recipes:
  • nginx:setup_host - creates nginx config file in /etc/nginx/sites-available/app_name.conf, grants user permissions to modify it. Creates #{deploy_to}/shared/logs/nginx directory and grants nginx permissions to write there.
  • mysql:setup_host - creates mysql user with random password grants him administrative permissions for application database (:database_name). Then creates :database_config_template in #{deploy_to}/shared/config/database.yml, grants it 440 permissions (only user and group can read it). It will be symlinked to #{deploy_to}/current/config/database.yml on each deploy.
  • monit:setup_host - creates monit config file for thin in /etc/monit/conf.d/app_name-thin, grants user permissions to modify it.
  • thin:setup_host - creates directory for thin sockets (if :thin_socket is set), creates thin config file in #{deploy_to}/shared/config/thin.yml

Config templates

Capistrano_recipes uses ERB to parse nginx/monit/thin/etc templates. You can see default config templates at github: https://github.com/rubydev/capistrano_recipes/tree/master/templates

If you don't like any of this, you can replace it with you own:

• Put your template in config/deploy/templates/#{template_name}
• Or change *_template variable value to reflect your template path: set :monit_config_template, 'deploy/monit.erb'

Contributing

1. Fork
2. Create your feature branch (git checkout -b my_branch)
3. Commit your changes (git commit -am "my cool feature")
4. Push to the branch (git push origin my_branch)
5. Create new Pull Request