Introduction

Capistrano-S3 is a capistrano deployment strategy that will:

* svn export locally
* make a tarball (named after the svn revision) of the release
* push the tarball to S3
* each server pulls that tarball from S3

This allows one to use an svn repository not exposed to the outside world, save time pushing out of a poky cable modem, and have new EC2 instances pull directly from S3 on startup.

If the tarball is already in S3 (from a previous deployment, say to a staging instance) the checkout is skipped.

Installation

gem install capistrano_s3

To use it, specify properties in config/deploy.rb:

set :deploy_via, :s3_bucket
set :deploy_s3_bucket, 'com.example.releases' # The name of the S3 bucket that should get releases
s3_config = YAML::load(ERB.new(IO.read("secret/s3.yml")).result) # Follow this pattern and don't ckin your secrets
# s3_config = { 'AWS_ACCOUNT_NUMBER' => '1234-5678-9012', 'AWS_ACCESS_KEY_ID' => 'ABCDEFGHIJKLMNOPQRST', 'AWS_SECRET_ACCESS_KEY' => 'abcdefghijklmnopqrstuvwxyz01234567890ABC' }
set :s3_config, s3_config

Now regular capistrano deployment tasks will go through S3

About

Rubyforge Project: rubyforge.org/projects/capistrano-s3
Author: Bill Kirtley - bill at [nospam] virosity dt com
License: Distributed under MIT License