SparkleFormation CLI ServerSpec callback
Ruby
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
lib
.gitignore
.rubocop.yml
CHANGELOG.md
CONTRIBUTING.md
Gemfile
LICENSE
README.md
Rakefile
sfn-serverspec.gemspec

README.md

SparkleFormation ServerSpec Callback

sfn-serverspec is a callback for sfn which executes Serverspec assertions against stack compute resources after successful stack creation.

Usage

Enable

Callbacks are configured via the .sfn configuration file. First, the callback must be enabled:

Configuration.new do
  callbacks do
    # requiring sfn-serverspec here is optional when using bundler
    require ['sfn-serverspec']
    default ['serverspec_validator']
  end
end

Configure

Some configuration of callback behavior is available via the .sfn file:

Configuration.new do
  sfn_serverspec do
    global_spec_patterns [File.join(Dir.pwd, 'spec/base/*_spec.rb')]
    ssh_proxy_command 'ssh ec2-user@server.example.com nc %h %p'
    ssh_user 'ubuntu'
  end
end

All of the following settings are optional:

  • global_spec_patterns - Array of strings specifying file paths for specs, defaults to []
  • ssh_proxy_command - String passed as the proxy command for remote SSH connection to the target compute resource, defaults to nil
  • ssh_user - Username for remote SSH connection, defaults to ec2-user
  • ssh_port - Port for remote SSH connection, defaults to 22
  • ssh_key_paths - Array of strings describing paths to one or more ssh private keys, defaults to []
  • ssh_key_passphrase - String used as passphrase for any encrypted ssh private keys defined in ssh_key_paths, defaults to nil

Additional configuration may be provided in a SparkleFormation template at the resource level:

resources(:my_cool_app_asg) do
  type 'AWS::AutoScaling::AutoScalingGroup'
  properties do
    # ...
  end
  serverspec do
    spec_patterns [File.join(Dir.pwd, '../spec/my_cool_app/*_spec.rb')]
  end
end

For each resource with a serverspec block defined, the resource-level value of spec_patterns is combined with the value of global_spec_patterns from the .sfn config file to yield a list of specs which will be executed against a given resource.

The following configuration options specified at the resource level will override the same options specified in the .sfn config file:

  • ssh_user
  • ssh_port
  • ssh_proxy_command
  • ssh_key_paths
  • ssh_key_passphrase

On-demand validation

Provided that you are using Bundler, this callback also adds a serverspec command to sfn, enabling on-demand validation of a running stack. You'll want to add sfn-serverspec to your Gemfile thusly:

group :sfn do
  gem 'sfn-serverspec'
end

Note that placing sfn and its friends in the :sfn group ensures that they'll be automatically require-d by sfn at run time.

The serverspec command requires both a stack name and a template to use as the source of Serverspec configuration. The template may be provided via command line flag or interactive file path prompt:

$ bundle exec sfn serverspec example-stack -f sparkleformation/example.rb
[Sfn]: Callback template serverspec_validator: starting
[Sfn]: Callback template serverspec_validator: complete
[Sfn]: Serverspec validating stack example-stack with template sparkleformation/example.rb:
[Sfn]: Callback after_serverspec serverspec_validator: starting
[Sfn]: Serverspec validating i-55836892 (10.101.100.15)

base
Port "22"
should be listening

hello world
displays a custom message on the index page
Port "80"
should be listening

Finished in 2.92 seconds (files took 10.49 seconds to load)
3 examples, 0 failures

[Sfn]: Serverspec validating i-52836895 (10.101.100.19)

Port "22"
should be listening

hello world
displays a custom message on the index page
Port "80"
should be listening

Finished in 0.35415 seconds (files took 13.41 seconds to load)
3 examples, 0 failures

[Sfn]: Callback after_serverspec serverspec_validator: complete

Caveats

Providing serverspec configuration on compute resources works in a manner similar to sfn's built-in Stack Policy callback. Specifically, the serverspec key and its contents are removed from the template during compile, cached by the callback and executed after completion of stack creation.

As of this writing, the callback only processes Serverspec configuration on AWS::AutoScaling::AutoScalingGroup and AWS::EC2::Instance resources.

Non-compute resources with serverspec configuration will not be processed by this callback, and will probably yield a validation error when trying to validate or create a stack from the template.

Info