Ruby cartridge, which is used in Openshift, supports by default only Passenger server running on Apache. But this Advanced Ruby cartridge allows you to use other popular servers, to be specific puma, unicorn, thin, rainbows and passenger. And you are allowed to switch Ruby platform to JRuby.

This for example means you can take advantage of websockets or better performance for your specific application. Not speaking about advantages of JRuby, especially for production servers and enterprise applications.

Make sure you have newer version of rhc installed, if not update your rhc

gem update rhc

It's standalone cartridge, so you can install it like this

rhc app create YOUR_APP_NAME http://cartreflect-claytondev.rhcloud.com/reflect?github=openshift-cartridges/advanced-ruby-cartridge

By default Passenger webserver is used. In order to change it, we will use new feature of rhc which allows us to change environment variables. Name of chosen webserver has to be inserted in OPENSHIFT_RUBY_SERVER variable.

If you are missing some webserver, you can take advantage of support for adding custom webserver, more information in next section.

# supported are (for ruby MRI (by default)): puma, unicorn, thin, passenger
# supported for jruby are: puma

# example of selecting webserver puma
rhc env set OPENSHIFT_RUBY_SERVER=puma -a YOUR_APP_NAME

To take effect, application must be restated or you must deploy your code again.

rhc app restart YOUR_APP_NAME

You can check which server is running by using this command

rhc ssh YOUR_APP_NAME '~/advanced-ruby/bin/control server'

If you are using Gemfile you need to add gem with your selected webserver. In this case it would be

group :production do
	gem 'puma'
end

For using webserver, which is not included in this cartridge, you can create custom server control script which defines operations like start, stop and restart of used webserver. This script must be located in application repository in .openshift/action_hooks/server_control and must be executable. Script will be invoked with context of repository.

For creating your own custom control script use this template.

#!/bin/bash -e
source "${OPENSHIFT_RUBY_DIR}/lib/servers"

function start() {
  # start action here
}

function stop() {
  # stop action here
}

function restart() {
  # restart action here
}

case "$1" in
  start)    start ;;
  stop)     stop ;;
  restart)  restart ;;
  *)        exit 0 ;;
esac

Variable OPENSHIFT_RUBY_SERVER must be set to custom by running this command rhc env set OPENSHIFT_RUBY_SERVER=custom -a YOUR_APP_NAME.

For inspiration, you can take a look to servers directory, to be specific into control file of some webserver.

It's highly advised to run any ruby command with ruby_context function (for example ruby_context "puma --help"), use it even for installing gems or you will experienced a lot of problems, especially if you are using JRuby.

If you have problem with debugging I recommend you to ssh into your app (rhc ssh APP_NAME) and test scripts manually by invoking them to see all error messages.

By default classic ruby implementation is used, but this cartridge also supports JRuby which can be switched as easy as changing webserver. You can switch to JRuby and vice versa in anytime, not just before first deploy.

# for default ruby implementation
rhc env set OPENSHIFT_RUBY_PLATFORM=default -a YOUR_APP_NAME

# for jruby
rhc env set OPENSHIFT_RUBY_PLATFORM=jruby -a YOUR_APP_NAME

Ruby platform is automatically changed only after new deploy of your code.

Check which Ruby platform is currently running by using

rhc ssh YOUR_APP_NAME '~/advanced-ruby/bin/control ruby-implementation'

If you have any problems with running application using this cartridge, take a look at my example Rails application which supports JRuby and also default Ruby environment, https://github.com/pbrazdil/test-app.

If you want to contact me, send me an email on [email protected]