Labor

A wrapper for "gearman-ruby" which provides an easy mechanism for managing your jobs. Jobs are just Ruby classes which respond to the perform method.

gem install labor

Basic Rakefile:

require 'labor'
require 'labor/tasks'

class TestJob < Labor::Ability
  def perform
    puts "Hello, world!"

    true
  end
end

class AnotherTest < Labor::Ability
  def perform
    puts "Hello, world... again!"

    true
  end
end

Labor.servers = ["10.0.0.1:4370", "10.0.0.2:4370"]

Let's run the worker:

$ export ABILITIES=test-job,another-test
$ rake labor:work

The ABILITIES environment variable can accept multiple job names; each seperated by a comma.

Job names can also carry meta data associated with that particular worker. Meta data is any information that shouldn't affect what you name the job class in your system, but indicate to Gearman that this worker is unique. You can do this by putting your meta data inside [ and ] in the ability name. For example: email[123]. Everything within and including the [] symbols will be removed when referencing your worker class' perform method.

Labor also allows you to load in a configuration file along with your worker. This configuration file is written in Ruby and offers a simple DSL (similar to the one present in Capistrano) for setting variables.

Configs look like this:

# settings.rb

set :foo, "bar"
set :i_am, "at the #{bar}"

group :credentials do
  set :username, "murkturgler"
  set :password, "supersecret"
end

The config file would then be loaded in like this:

Labor.config File.join(File.dirname(__FILE__), 'settings.rb')

When the worker is started up, Labor will load this config into the @config instance variable of your job class/module. You'll then be able to retrieve keys from the config within your job.

class TestJob < Labor::Ability
  def perform
    puts "Hey! I'm #{@config[:i_am]}"
    #=> "Hey! I'm at the bar"

    puts "Here's my username: #{@config.credentials.username}"
    #=> "Here's my username: murkturgler"

    true
  end
end