DelayedJob and PG Error No Connection to Server

Recently I’ve got strange errors on the staging server. Delayed::Job workers started successfully. But when they were starting to lock the jobs, workers failed with PG::Error: no connection to server and PG::Error: FATAL: invalid frontend message type 60 errors.

Obviously, that is not what I’ve expected from Delayed::Job workers. So I took the shovel and started digging into git history. Since the last release the only significant modification has been made in the internationalization. We’ve moved to I18n-active_record backend to grant the privilege to modify translations not only to developers but also to highly-educated mere mortals.

So, the problem source was found, and I18n-active_record backend was temporary disabled.

After a while I’ve finally had some time to get around to this issue again. I’ve enabled locales database backend and started some serious debugging.

The first thing I’ve found was that the problem did not turn up when delayed job workers were started using rake jobs:work task. So this was the first solution.

After some research on DelayedJob internals I’ve found that the main difference between a rake task and a binstub was in the fork method that was invoked in the binstub version. It is seamlessly executed using Daemons#run_process method.

Delayed::Job lifecycle

Now let’s take a break and look into DelayedJob internals. Delayed::Job has systems of the hooks that could be used by plugin-writers and in our applications. All this cool events functionality is hidden in Delayed::Lifecycle class. Each worker has its own instance of that class.

So, what events do we have there?

Job-related events:

 :enqueue
 :perform
 :error
 :failure
 :invoke_job

Worker-related events:

:execute
:loop
:perform
:error
:failure

You can setup callbacks to be run on before, after or around events simply using Delayed::Worker.lifecycle.before, Delayed::Worker.lifecycle.after and Delayed::Worker.lifecycle.around methods.

The Solution

Ok, let’s move on to our problem. It has turned out that delayed_job_active_record is closing all database connections in before_fork hook and reestablishing them in after_fork hook. It was clear that I18n-active-record does not like it and does not want to play by the rules set by others. So, it needed special treatment and I provided it.

I’ve looked into DelayedJob lifecycle and chosen before :execute hook. It is being executed after all DelayedJob ActiveRecord backend connections manipulations.

So, basically my locales initializer for delayed_job workers looks like this:

require 'i18n/backend/active_record'

Translation  = I18n::Backend::ActiveRecord::Translation

Delayed::Worker.lifecycle.before :execute do
  if Translation.table_exists?
    I18n.backend = I18n::Backend::ActiveRecord.new

    I18n.backend = I18n::Backend::Chain.new(I18n::Backend::Simple.new, I18n.backend)
  end
end