Cap-EC2

Cap-EC2 is used to generate Capistrano namespaces and tasks from Amazon EC2 instance tags, dynamically building the list of servers to be deployed to.

Notes

Cap-EC2 is only compatible with Capistrano 3.x or later; if you want the Capistrano 2.x version, use Capify-EC2. Note that the configuration file (config/ec2.yml) is not compatible between versions either.

This documentation assumes familiarity with Capistrano 3.x.

A number of features that are in Capify-EC2 are not yet available in Cap-EC2, due to the architectural changes in Capistrano 3.x. The following features are missing (this is not an exhaustive list!):

• rolling deploy (this should be implemented via SSHKit)
• ELB registration/de-registration (not widely used)
• Variables set by EC2 tags
• Connecting to instances via SSH using a convenience task

Pull requests for these would be welcomed, as would sending feedback via the Issues on this project about features you would like.

Installation

gem install cap-ec2

or add the gem to your project's Gemfile.

You also need to add the gem to your Capfile:

require "cap-ec2/capistrano"

Configuration

Configurable options, shown here with defaults:

set :ec2_config, 'config/ec2.yml'

set :ec2_project_tag, 'Project'
set :ec2_roles_tag, 'Roles'
set :ec2_stages_tag, 'Stages'

set :ec2_access_key_id, nil
set :ec2_secret_access_key, nil
set :ec2_region, %w{}
set :ec2_contact_point, nil

set :ec2_filter_by_status_ok?, nil

Order of inheritance

cap-ec2 supports multiple methods of configuration. The order of inheritance is: YAML File > User Capistrano Config > Default Capistrano Config > ENV variables.

Regions

:ec2_region is an array of AWS regions, if not present all regions will be checked for matching instances - this the default behavior and can be slow, if speed is an issue consider setting your required regions.

If:ec2_access_key_id or :ec2_secret_access_key are not set in any configuration the environment variables AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY and AWS_REGION will be checked. If running on EC2 the IAM instance profile credentials will be used if credentials are not set by any other method.

Misc settings

• project_tag

  Cap-EC2 will look for a tag with this name when searching for instances that belong to this project. Cap-EC2 will look for a value which matches the :application setting in your deploy.rb. The tag name defaults to "Project" and must be set on your instances.

• stages_tag

  Cap-EC2 will look for a tag with this name to determine which instances belong to a given stage. The tag name defaults to "Stages".

• roles_tag

  Cap-EC2 will look for a tag with this name to determine which instances belong to a given role. The tag name defaults to "Roles".

• filter_by_status_ok?

  If this is set to true, then Cap-EC2 will not return instances which do not have both EC2 status checks as OK. By default this is set to nil, so Cap-EC2 can return you instances which don't have OK status checks. Be warned that just-launched instances take a while to start returning OK.

YAML Configuration

If you'd prefer do your configuration via a YAML file config/ec2.yml can be used, (or an alternative name/location set via set :ec2_config):

If so YAML file will look like this:

access_key_id: "YOUR ACCESS KEY"
secret_access_key: "YOUR SECRET KEY"
regions:
 - 'eu-west-1'
project_tag: "Project"
roles_tag: "Roles"
stages_tag: "Stages"

Your config/ec2.yml file can contain (access_key_id, secret_access_key, regions) - if a value is omitted then the order of inheritance is followed.

Usage

Imagine you have four servers on EC2 named and tagged as follows:

'Name' Tag	'Roles' Tag	'Stages' Tag
server-1	web	production
server-2	web,app	production
server-3	app,db	production
server-4	web,db,app	staging

Imagine also that we've called our app "testapp", as defined in config/deploy.rb like so:

set :application, "testapp"

Defining the roles in config/deploy/[stage].rb

To define a role, edit config/deploy/[stage].rb and add the following:

ec2_role :web

Let's say we edited config/deploy/production.rb. Adding this configuration to the file would assign the role :web to any instance that has the following properties:

• has a tag called "Roles" that contains the string "web"
• has a tag called "Project" that contains the string "testapp"
• has a tag called "Stages" that contains the current stage we're executing (in this case, "production")

Looking at the above table, we can see we would match server-1 and server-2. (You can have multiple roles in tag separated by commas.)

Now we can define the other roles:

ec2_role :app
ec2_role :db

In the "production" stage, the :app role would apply to server-2 and server-3, and the :db role would apply to server-3.

In the "staging" stage, all roles would apply only to server-4.

Servers belonging to multiple projects

If you require your servers to have multiple projects deployed to them, you can simply specify all the project names you want to the server to be part of in the 'Projects' tag, separated by commas. For example, you could place a server in the testapp and myapp projects by setting the 'Projects' tag to testapp,myapp.

Servers in multiple stages

If your use-case requires servers to be in multiple stages, simply specify all the stages you want the server to be in 'Stages' tag, separated by commas. For example, you could place a server in the production and staging stages by setting the 'Stages' tag to production,staging.

Passing options to roles

You can pass options when defining your roles. The options are exactly the same as the options that the Capistrano native role definition takes, since they are passed straight through to Capistrano. For example:

ec2_role :app,
  user: 'user_name',
  ssh_options: {
    user: 'user_name', # overrides user setting above
    keys: %w(/home/user_name/.ssh/id_rsa),
    forward_agent: false,
    auth_methods: %w(publickey password)
    password: 'please use keys'
  }

See the example config files Capistrano builds for you in config/deploy for more details.

Note that at the moment there's no way to pass variables in from EC2 tags - but it would be trivial to add.

Tasks and deployment

You can now define your tasks for these roles in exactly the same way as you would if you weren't using this gem.

Contacting instances

By default, Cap-EC2 will attempt to communicate with the EC2 instance using the following instance interfaces in order:

1. Public DNS (:public_dns)
2. Public IP (:public_ip)
3. Private IP (:private_ip)

This can be configured using the Capistrano variable :ec2_contact_point, and supplying one of the above symbols. For example:

set :ec2_contact_point, :private_ip

This would cause Cap-EC2 to try communicating with the instance on its private IP address. If you leave this variable unset, the behaviour is as in previous Cap-EC2 instances (falling through the lookup list as specified above).

Utility tasks

Cap-EC2 adds a few utility tasks to Capistrano for displaying information about the instances that you will be deploying to. Note that unlike Capistrano 2.x, all tasks require a stage.

View instances

This command will show you information all the instances your configuration matches for a given stage.

cap [stage] ec2:status

Example:

$ cap production ec2:status

Num  Name                          ID          Type      DNS              Zone        Roles         Stage
00:  server-1-20131030-1144-0      i-abcdefgh  m1.small  192.168.202.248  us-west-2c  banana,apple  production
01:  server-2-20131118-1839-0      i-hgfedcba  m1.small  192.168.200.60   us-west-2a  banana        production

View server names

This command will show you the server names of the instances matching the given stage:

cap [stage] ec2:server_names

Example:

$ cap production ec2:server_names
server-1-20131030-1144-0
server-2-20131118-1839-0

View server instance IDs

This command will show the instance IDs of the instances matching the given stage:

cap [stage] ec2:instance_ids

Example:

$ cap production ec2:instance_ids
i-abcdefgh
i-hgfedcba

Acknowledgements

Thanks to Rylon for maintaining Capify-EC2 and reviewing my thought processes for this project.

Contributing

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