Runner

The Runner class encapsulates all of the responsibilities of the worker into a class that can be instantiated and executed in numerous environments including custom scripts or even a Ruby on Rails application.

A simple runner can be created with the following code

runner = Wamp::Worker::Main.new uri: 'ws://127.0.0.1:8080/ws', realm: 'realm1'
runner.start

When "start" is called, the runner will execute infinitely until either

• Something calls runner.stop, or
• An "INT" or "TERM" signal is sent to the script (from pressing ctrl-c for example)

The main thread of the runner is responsible for maintaining the socket connection to the WAMP router (see wamp_client) and performs the duties of a dispatcher proxy (see Proxy).

The runner spawns a side thread that will watch Redis queues for requests from requestors and backgrounders. To minimize the access to Redis, the Redis "brpop" method is used which will block until a descriptor is added to the Redis queues being watched. Upon receiving a descriptor, the thread will place it in a Ruby Queue object which is designed to allow information to be passed between threads.

While the main Runner thread is executing, it will check the Queue on each tick of the Event Machine to see if there are descriptors to execute and process them accordingly.

Note that the main Runner thread will also increment a "tick" counter every second which signals to the requestors if the worker is running in the event of a timeout.

Starting in a Custom Script

To start the runner in a custom script, do the following

require "wamp/worker"

Wamp::Worker.configure do
  connection uri: 'ws://127.0.0.1:8080/ws', realm: 'realm1'
end

Wamp::Worker.run

If you don't want the runner to block, you can also spawn a thread

Thread.new do
  Wamp::Worker.run
end

Runners with different configurations are also supported

require "wamp/worker"

Wamp::Worker.configure :runner1 do
  connection uri: 'ws://127.0.0.1:8080/ws', realm: 'realm1'
end

Wamp::Worker.configure :runner2 do
  connection uri: 'ws://127.0.0.1:8080/ws', realm: 'realm2'
end

Thread.new do
  Wamp::Worker.run :runner1
end

Thread.new do
  Wamp::Worker.run :runner2
end

Starting from Command Line

The Wamp::Worker GEM includes the script "bin/wamp-worker" that is intended to encapsulate the runner. It has the following options

• "-l" (default: "info"): logging level (debug, info, warn, error)
• "-n" (default: "default"): name of the worker
• "-e" (default: "development"): application environment to load
• "-r" (default: "./": the path to the file to require. Note that is "-r" is a directory, the script will assume you are loading a "Ruby on Rails" application

To start the runner from the command line, do the following

$ wamp-worker -r "<path to require file>" -l info

Ruby on Rails

To start the runner in a Ruby on Rails environment, create a config file

app/config/initializers/wamp-worker.rb

Wamp::Worker.configure do
  connection uri: 'ws://127.0.0.1:8080/ws', realm: 'realm1'
end

and call the script

$ cd "<top of Rails application>"
$ wamp-worker -e development -l debug

Some notes about running in Rails

• Make sure have config.eager_load = true set in your config/environments/* files
• Multiple workers can be created with different configurations by passing the name to the configure call (Wamp::Worker.configure :other do) and passing the "-n" option to the script ($ wamp-worker -n other)

To run as a Heroku worker, add the following to your Procfile

app/Procfile

web: bundle exec puma -C config/puma.rb
worker: bundle exec wamp-worker -e production

Multiple workers are also supported

app/config/initializers/wamp-worker.rb

Wamp::Worker.configure :worker1 do
  connection uri: 'ws://127.0.0.1:8080/ws', realm: 'realm1'
end

Wamp::Worker.configure :worker2 do
  connection uri: 'ws://127.0.0.1:8080/ws', realm: 'realm2'
end

app/Procfile

web: bundle exec puma -C config/puma.rb
worker1: bundle exec wamp-worker -n worker1 -e production
worker2: bundle exec wamp-worker -n worker2 -e production