Rna gem

Rna is a ruby gem that provides simple DSL for generating node.json files required by chef-solo.

Requirements

$ gem install rna

Usage

$ mkdir rna
$ cd rna
$ rna init

This will create starter config/rna.rb and config/s3.yml files.

Example:

config/rna.rb file

# This is starter example rna template.
# This is meant be be modified to your needs.
default_includes 'base'
# Pre processing rules that run at the beginning
before do
  if role != 'base'
    node[:application] = nil
    node[:deploy_code] = false
    node[:framework_env] = 'production'
    node[:repository] = nil
  end

  node[:before] = 1
  node[:chef_branch] = 'prod' if role =~ /^prod/
  node[:chef_branch] = 'master' if role =~ /^stag/
end

settings[:sendgrid][:relayhost] = "smtp.sendgrid.net"

# Roles
role 'base' do
  role_list ['base']
end

# api
role 'prod-api-redis', 'stag-api-redis' do
  run_list ['base','api_redis']
end
role 'prod-api-app', 'stag-api-app' do
  run_list ['base','api_app']
  node[:application] = 'api'
  node[:deploy_code] = true
  node[:database][:adapter] = "mysql"
  node[:database][:host] = "127.0.0.1"
  node[:database][:user] = "user"
  node[:database][:pass] = "pass"
  node[:repository] = 'git@github.com:owner/repo.git/api.git'
end
role 'prod-api-resque', 'stag-api-resque' do
  includes 'prod-api-app'
  run_list ['base','api_resque']
  node[:workers] = 8
end


# Post processing rules that run at the end
after do
  node[:after] = 2
  node[:framework_env] = 'production' if role =~ /^prod/
  node[:framework_env] = 'staging' if role =~ /^stag/

  list = role.split('-')
  if list.size == 3
    env, repo, role = list
    role_list ['base', "#{repo}_#{role}"]
    node[:application] = repo
  end
end

Shared Settings

You might want a shared settings hash that you can use in only some of your roles.

settings[:foo][:bar] = 1
settings[:foo][:baz] = 2

You can use this any where in your roles.

role 'role1' do
  node[:foo][:bar] = settings[:foo][:bar]
end

role 'role2' do
  node[:foo][:bar] = settings[:foo][:bar]
end

role 'role3' do
  node[:foo][:baz] = settings[:foo][:baz]
end

Breaking up config/rna.rb

If you have a lot of roles, the config/rna.rb file can get unwieldy long. You can break up the rna.rb file and put role defintions in the config/rna directory. Any file in this directory will be automatically loaded.

An example is in the spec/project folder:

config/rna/api.rb
config/rna/blog.rb

Generating the json files

$ rna generate

Here is the example of the output looks like:

base.json:

{
  "before": 1,
  "role": "base",
  "run_list": [
    "role[base]"
  ],
  "after": 2
}

prod-api-app.json:

{
  "before": 1,
  "role": "prod-api-app",
  "run_list": [
    "role[base]",
    "role[api_app]"
  ],
  "application": "api",
  "deploy_code": true,
  "database": {
    "adapter": "mysql",
    "host": "127.0.0.1",
    "user": "user",
    "pass": "pass"
  },
  "repository": "git@github.com:owner/repo.git/api.git",
  "after": 2,
  "framework_env": "production"
}

prod-api-redis.json:

{
  "before": 1,
  "role": "prod-api-redis",
  "run_list": [
    "role[base]",
    "role[api_redis]"
  ],
  "after": 2,
  "framework_env": "production",
  "application": "api"
}

prod-api-resque.json:

{
  "before": 1,
  "role": "prod-api-resque",
  "run_list": [
    "role[base]",
    "role[api_resque]"
  ],
  "application": "api",
  "deploy_code": true,
  "database": {
    "adapter": "mysql",
    "host": "127.0.0.1",
    "user": "user",
    "pass": "pass"
  },
  "repository": "git@github.com:owner/repo.git/api.git",
  "workers": 8,
  "after": 2,
  "framework_env": "production"
}

stag-api-app.json:

{
  "before": 1,
  "role": "stag-api-app",
  "run_list": [
    "role[base]",
    "role[api_app]"
  ],
  "application": "api",
  "deploy_code": true,
  "database": {
    "adapter": "mysql",
    "host": "127.0.0.1",
    "user": "user",
    "pass": "pass"
  },
  "repository": "git@github.com:owner/repo.git/api.git",
  "after": 2,
  "framework_env": "staging"
}

stag-api-redis.json:

{
  "before": 1,
  "role": "stag-api-redis",
  "run_list": [
    "role[base]",
    "role[api_redis]"
  ],
  "after": 2,
  "framework_env": "staging",
  "application": "api"
}

stag-api-resque.json:

{
  "before": 1,
  "role": "stag-api-resque",
  "run_list": [
    "role[base]",
    "role[api_resque]"
  ],
  "application": "api",
  "deploy_code": true,
  "database": {
    "adapter": "mysql",
    "host": "127.0.0.1",
    "user": "user",
    "pass": "pass"
  },
  "repository": "git@github.com:owner/repo.git/api.git",
  "workers": 8,
  "after": 2,
  "framework_env": "staging"
}

Uploading to S3

$ rna build -o s3 # saves s3 based on config/s3.yml settings

The config/s3.yml should look like this:

access_key_id: hocuspocus
secret_access_key: opensesame
bucket: my-bucket
folder: chef/rna

rna

Development

Runtime