Last active

Unicorn.rb for configuration

  • Download Gist
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
worker_processes 3 # amount of unicorn workers to spin up
timeout 30 # restarts workers that hang for 30 seconds
preload_app true
GC.respond_to?(:copy_on_write_friendly=) and
GC.copy_on_write_friendly = true
before_fork do |server,worker|
defined?(ActiveRecord::Base) and
after_fork do |server,worker|
defined?(ActiveRecord::Base) and
if defined?(EventMachine)
unless EventMachine.reactor_running? && EventMachine.reactor_thread.alive?
if EventMachine.reactor_running?
end { }
Signal.trap("INT") { EventMachine.stop }
Signal.trap("TERM") { EventMachine.stop }

I'm using this on Heroku and occasionally seeing this error: An EventMachine loop must be running to use publish_async calls.

Have you seen anything like this? I assume that for some reason the EventMachine crashed but the dyno remained alive. Do you have any suggestions for how to handle that case? Could I restart the EventMachine from inside one of my unicorn forks?

I have also seen the same error. Did you find a solution? Also, do you see this: Error R12 (Exit timeout) -> At least one process failed to exit within 10 seconds of SIGTERM

This only started happening after I added this code.

The issues commented above were solved ? If yes, how ? Thanks in advance.

To @arches - it is possible that EM can crash while the dyno still is up (even when EM is restarted after forks). A quick way to avoid this is to ensure the loop is running before each call:

def ensure_em
  unless EventMachine.reactor_running? && EventMachine.reactor_thread.alive? { }
    sleep 1

# put this before calls

To @elihuu - hard to say what that might be. It could be a stray event loop, but stop should be getting called. If something is going wrong in the event loop, you might be able to find it by registering a global error handler.

EventMachine.error_handler{ |e|
  puts "Error raised during event loop: #{e.message}"

Please sign in to comment on this gist.

Something went wrong with that request. Please try again.