Capistrano AWS

This gem is slightly based on cap-ec2, which is not maintained for a while, becoming outdated.

The purpose of this gem is to provide a flexible and simple integration to AWS EC2, exposing the aws-sdk connection in order to allow any customization.

Requirements

• Capistrano 3

Installation

Add to your Gemfile and run bundler:

gem 'capistrano-aws'

Or install the gem system-wide:

gem install cap-ec2

In your Capfile:

require 'capistrano/aws'

Configuration

# AWS regions to use.
set :aws_ec2_regions, ['us-east-1']

# Application name to match application tag.
set :aws_ec2_application, (proc { fetch(:application) })

# Stage to match stage tag.
set :aws_ec2_stage, (proc { fetch(:stage) })

# Tag to be used for Capistrano stage.
set :aws_ec2_stage_tag, 'Stage'

# Tag to be used to match the application.
set :aws_ec2_application_tag, 'Application'

# Tag to be used for Capistrano roles of the server (the tag value can be a comma separated list).
set :aws_ec2_roles_tag, 'Roles'

# Extra filters to be used to retrieve the instances. See the README.md for more information.
set :aws_ec2_extra_filters, []

# Tag to be used as the instance name in the instances table (aws:ec2:instances task).
set :aws_ec2_name_tag, 'Name'

# How to contact the instance (:public_ip, :public_dns, :private_ip).
set :aws_ec2_contact_point, :public_ip

The AWS credentials are loaded from your system. Check https://github.com/aws/aws-sdk-ruby#configuration for more information.

Usage

The instances must be registered in each stage. In your config/deploy/<stage_name>.rb, add the following line:

aws_ec2_register

It will use the instance tags to call the server function in capistrano. You can pass a hash as an argument, and it will be merged into the function call:

aws_ec2_register user: 'hello', port: 2222

Custom EC2 Filters

If you need to identify the instances based on more information, you can specify extra filters to be used in the filters option in the Aws::EC2::Resource.instances call.

Imagine you have some split test servers for your production environment, so you can use another tag for specifying the variation of some servers:

Name	Application	Environment	Roles	SplitTestVariation
foo-app01	foo-app	production	app,db	0
foo-app02	foo-app	production	app	0
foo-app03	foo-app	production	app	1
foo-app04	foo-app	production	app	1

In your environment files:

# production.rb

set :aws_ec2_extra_filters, [
  {
    name: "tag:SplitTestVariation",
    values: ["0"],
  },
]

# production-b.rb

set :aws_ec2_stage, "production" # Optional. Use if you have the same environment for the B servers.
set :aws_ec2_extra_filters, [
  {
    name: "tag:SplitTestVariation",
    values: ["1"],
  },
]

The :aws_ec2_stage variable is needed in order to override the default value of the stage fielter(:stage). If you really have a different environment for your B servers, you can just use the name of the environment as the file name and remove this line.

Utility tasks

aws:ec2:instances

List all the instances found for with the current configuration.

cap production aws:ec2:instances

Contributing

Open an issue or make a PR, feel free to contribute!