capper

Author:	Benedikt Böhm
Web:	http://github.com/zenops/capper
Git:	git clone https://github.com/zenops/capper

Capper is a collection of opinionated Capistrano recipes.

Rationale

Most web applications are deployed the same way. While capistrano itself is already quite opinionated, maintaining a multitude of applications feels like copy&paste very fast.

Capper provides sane defaults and many recipes for technologies typically used with Ruby and Python deployments to make config/deploy.rb much more declarative and terse.

Release Notes

v1.0.0 (14/07/2012)

capper has been in production for about 1 year and its API is now considered stable.

Usage

Capper recipes should be loaded via Capfile like this:

require "capper"

load "capper/bundler"
load "capper/rails"
load "capper/rvm"
load "capper/unicorn"

load "config/deploy"

load "capper/multistage"

Note: capper does not support capistranos default deploy recipe, instead an enhanced copy is shipped directly with capper and enabled by default.

Multistage

Capper provides a clean DSL syntax for multistage support. In contrast to capistrano-ext you don't have to create a bunch of tiny files in config/deploy/ but instead use DSL sugar in config/deploy.rb:

stage :production do
  server "app1.example.com", :app, :db, :primary => true
  server "app2.example.com", :app
end

stage :staging do
  server "staging.example.com", :app, :db, :primary => true
  set :rails_env, "staging"
  set :branch, "staging"
end

Note: the multistage recipe must be loaded after config/deploy otherwise it cannot know the defined stages.

Templates

Capper provides a powerful yet simple mechanism to upload scripts and config files rendered from templates during deploy. Template files are stored in config/deploy/templates but are rarely needed since capper recipes already contain required template files.

To upload a custom template create an ERb file and upload it with upload_template_file:

upload_template_file("myscript.sh", "#{bin_path}/myscript", :mode => "0755")

The above command will upload config/deploy/templates/myscript.sh.erb to the specified path. Inside the template all capistrano options (current_path, application, etc) are available.

systemd integration

Capper provides integration with a systemd user session running for the user that is being deployed to. Capper will upload service units and enable/start these accordingly.

systemd is disabled by default, If you use systemd simply activate it like this:

set :use_systemd, true

Recipes

The following recipes are included in capper.

base

The base recipe is an enhanced version of capistranos default deploy recipe. It is loaded automatically and provides the basic opinions capper has about deployment:

• The application is deployed to a dedicated user account with the same name as the application.
• The parent path for these user accounts defaults to /var/app.
• No sudo privileges are ever needed.
• Releases are cleaned up by default.
• The application is deployed via remote_cache and git.

The deploy:finalize_update task has been enhanced to make symlinks declarative in config/deploy.rb:

set :symlinks, {
  "config/database.yml" => "config/database.yml",
  "shared/uploads" => "public/uploads"
}

The above snippet will create symlinks from #{shared_path}/config/database.yml to #{release_path}/config/database.yml and from #{shared_path}/uploads to #{release_path}/public/uploads after deploy:update_code has run.

airbrake

The airbrake recipe is merely a copy of airbrakes native capistrano integration without after/before hooks, so airbrake notifications can be enabled on-demand in stage blocks:

stage :production do
  ...
  after "deploy", "airbrake:notify"
end

bundler

The bundler recipe is an extension of bundlers native capistrano integration:

• During bundle:install it is ensured that a known-to-work bundler version (specified via bundler_version) is installed.
• When used together with the rvm recipe bundles are not installed globally to shared/bundle but instead a gemset specific location is used (shared/bundle/#{gemset}).
• The option ruby_exec_prefix is set to bundle exec for convenience. (see ruby recipe for details)

delayed_job

The delayed_job recipe provides integration with DelayedJob. A script to start/stop delayed job workers is uploaded to #{bin_path}/delayed_job. The script supports multiple instances and priority ranges.

If monit integration has been enabled via capper/monit workers are automatically (re)started during deploy and can be specified via delayed_job_workers:

set :delayed_jobs_workers, {
  :important => 0..1,
  :worker1 => 2..10,
  :worker2 => 2..10
}

django

The django recipe provides setup, migrate and collectstatic tasks for Django.

python

The python recipe provides basic support for Python applications. It will create a symlink from #{current_path}/#{application} to #{current_path} for Python namespace support.

rails

The rails recipe sets the default rails_env to production and includes tasks for deploying the asset pipeline for rails 3.1 applications. It also provdes a migrate task for Rails applications.

rvm

The rvm recipe is an extension to RVMs native capistrano integration. The recipe forces the rvm_type to :user and will automatically determine the ruby version and gemset via the projects .ruby-version and .ruby-gemset files.

A deploy:setup hook is provided to ensure the correct rvm, ruby and rubygems versions are installed on all servers.

ruby

The ruby recipe provides basic support for Ruby applications. It will setup a gemrc file and and variables for ruby_exec_prefix (such as bundler).

thin

The thin recipe provides integration with Thin. A script to manage the thin process is uploaded to #{bin_path}/thin.

unicorn

The unicorn recipe provides integration with Unicorn. A script to manage the unicorn process is uploaded to #{bin_path}/unicorn. Additionally this recipe also manages the unicorn configuration file (in config/unicorn.rb).

The following configuration options are provided:

unicorn_worker_processes

Number of unicorn workers (default: 4)

unicorn_timeout

Timeout after which workers are killed (default: 30)

uwsgi

The uwsgi recipe provides integration with uWSGI. A script to manage the uwsgi process is uploaded to #{bin_path}/uwsgi. Additionally this recipe also manages the uwsgi configuration file (in config/uwsgi.xml).

The following configuration options are provided:

uwsgi_worker_processes

Number of uwsgi workers (default: 4)

virtualenv

The virtualenv recipe provides deploy:setup hooks for virtualenv support. In addition required Python libraries are installed via pip into this environment.

The following configuration options are provided:

python_requirements_file

Relative path to requirements file (default: requirements.txt)

whenever

The whenever recipe is a simplified version of whenevers native capistrano integration. With one application per user account the whole crontab can be used for whenever. Additionally this recipe take the ruby_exec_prefix setting into account. To define the target servers user the cron role.

server "app1.example.com", :app, :cron

node deployment

read http://big-elephants.com/2012-07/deploying-node-with-capistrano/ about the the use case.

nave

The nave recipe sets up nave Virtual Environments for Node:

set :use_nave, true
set :nave_dir, '~/.nave'
set :node_version, '0.8.1'

npm

The npm recipe runs npm install after deploy:update_code. When used with the nave recipe npm install runs nave use <ver> npm install. Not it is recommended to add npm-shrinkwrap.json into version control to manage npm dependencies:

set :npm_cmd, "npm"

forever

The forever recipe starts your app as daemon in the background. When used with the nave recipe it runs nave use <ver> forever [action]:

set :forever_cmd, "forever" # e.g. "./node_modules/.bin/forever"
set :node_env, "production" # the NODE_ENV environment variable used to start the script
set :main_js, "index.js" # e.g. "./build/main.js" the script you want to start

Contributing to capper

• Check out the latest master to make sure the feature hasn't been implemented or the bug hasn't been fixed yet
• Check out the issue tracker to make sure someone already hasn't requested it and/or contributed it
• Fork the project
• Start a feature/bugfix branch
• Commit and push until you are happy with your contribution

Copyright

Copyright (c) 2011-2012 Benedikt Böhm. See LICENSE for further details.