Capistrano::Stretcher

capistrano task for stretcher.

Installation

Add this line to your application's Gemfile:

gem 'capistrano-stretcher'

And then execute:

$ bundle

Or install it yourself as:

$ gem install capistrano-stretcher

Requirements

capistrano-stretcher requires target server for building to application assets. This server should be installed the following packages:

git
rsync
tar
gzip
awk
openssl
aws-cli
consul
pv

target server builds assets, uploads assets to AWS S3 and invokes consul event automatically. So target server can access AWS s3 via aws-cli and join your deployment consul cluster.

If you want to use non-s3 (e.g. private DC), upload assets to your server with rsync and download from http(s).

Usage

You need to add require "capistrano/stretcher" to Capfile and add config/deploy.rb following variables:

role :build, ['your-target-server.lan'], :no_release => true

# If your build server is accessible to consul cluster, then set the same server
# If different, set a server that can access consul cluster
# just to kick consul event!
role :consul, ['your-target-server.lan'], :no_release => true

set :application, 'your-application'
set :deploy_to, '/var/www'
set :deploy_roles, 'www,batch'
set :stretcher_hooks, 'config/stretcher.yml.erb'
set :local_tarball_name, 'rails-applicaiton.tar.gz'
set :stretcher_src, "s3://your-deployment-bucket/assets/rails-application-#{env.now}.tgz"
set :manifest_path, "s3://your-deployment-bucket/manifests/"
# Optional, if you want to use mv
set :stretcher_sync_strategy, "mv"

# Optinal, if you want to http(s) in stretcher_src, manifest_path
set :rsync_ssh_option, "-p 22"
set :rsync_ssh_user, "MY_USER" # if undefined, use current user on build server
set :rsync_host, "xxx.xxx.xxx.xxx"
set :rsync_stretcher_src_path, "/var/www/resource/assets/rails-application-#{env.now}.tgz"
set :rsync_manifest_path,      "/var/www/resource/manifests"

and write hooks for stretcher to config/stretcher.yml.erb

default: &default
  pre:
    -
  success:
    -
  failure:
    - cat >> /tmp/failure
www:
  <<: *default
  post:
    - ln -nfs <%= fetch(:deploy_to) %>/shared/data <%= fetch(:deploy_to) %>/current/data
    - sudo systemctl reload unicorn
batch:
  <<: *default
  post:
    - ln -nfs <%= fetch(:deploy_to) %>/shared/data <%= fetch(:deploy_to) %>/current/data

above hooks is extracted to manifest.yml for stretcher. If you have "www,batch" roles and stages named staging and production, capistrano-stretcher extract to following yaml from configuration.

manifest_www_staging.yml
manifest_batch_staging.yml

and invoke

consul event -name deploy_www_staging s3://.../manifest_www.yml
consul event -name deploy_batch_staging s3://.../manifest_batch.yml

with cap staging stretcher:deploy command on target server. When it's invoked with cap production stretcher:deploy, capistrano-stretcher replace suffix staging to production.

Development

After checking out the repo, run bin/setup to install dependencies. Then, run rake test to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/pepabo/capistrano-stretcher.

LICENSE

The MIT License (MIT)