Job doesn't run from static schedule

[Babdus]

Hello,
I have read similar issues but nothing helped me.

Here is my lib/tasks/resque.rake file:

require 'resque/tasks'
require 'resque/scheduler/tasks'

namespace :resque do
  task :setup do
    require 'resque'
    require 'resque-scheduler'
  end
  task :setup_schedule => :setup do
    require 'resque-scheduler'
    Resque.schedule = YAML.load_file("#{Rails.root}/config/resque_schedule.yml")
  end
  task :scheduler => :setup_schedule
end

And resque_schedule.yml:

TestCronJob:
  every: '1m'
  class: TestCronJob
  queue: test
  description: "This job puts something every 1 minute"

And this is test_cron_job.rb:

class TestCronJob < ApplicationJob
  queue_as :test

  def perform
    puts 'something'
  end

end

In command line:

QUEUE=* bundle exec rake environment resque:work&

and

rake environment resque:scheduler

The job is being queued, but never run (I tested on several jobs, not just putting something):

resque-scheduler: [INFO] 2017-05-29T02:36:07+04:00: Starting
resque-scheduler: [INFO] 2017-05-29T02:36:07+04:00: Loading Schedule
resque-scheduler: [INFO] 2017-05-29T02:36:07+04:00: Scheduling TestCronJob 
resque-scheduler: [INFO] 2017-05-29T02:36:07+04:00: Schedules Loaded
resque-scheduler: [INFO] 2017-05-29T02:37:07+04:00: queueing TestCronJob (TestCronJob)
resque-scheduler: [INFO] 2017-05-29T02:38:07+04:00: queueing TestCronJob (TestCronJob)

Can anybody help me find out what is going wrong?

[carsonreinke]

@Babdus Any chance this is using a different Redis instance or database? Could you try running resque:work with the VERBOSE=1

[Babdus]

I have only one Redis instance with one database, so it can't be the case.
After running resque:work with the VERBOSE=1 here is the output:

*** Starting worker main
*** Running before_first_fork hooks
*** got: (Job{test} | TestCronJob | [])
*** Running before_fork hooks with [(Job{test} | TestCronJob | [])]
*** Running after_fork hooks with [(Job{test} | TestCronJob | [])]
*** done: (Job{test} | TestCronJob | [])

But nothing ever happened. (I've put code in this TestCronJob for retrying failed jobs)

[carsonreinke]

This issue is unrelated to resque-scheduler, as clearly the job is being queued into the right place and resque is running it.

With that said, I think your problem is ApplicationJob defines self.perform, but in your TestCronJob that defines the instance method perform, not the class level, so try changing that to just def self.perform and it should work.

[sevos]

I have the same issue. It seems that resque-schedule does not play with ActiveJob adapter. When I enqueue job manually, I get the following output:

*** got: (Job{default} | ActiveJob::QueueAdapters::ResqueAdapter::JobWrapper | [{"job_class"=>"BrainTickJob", "job_id"=>"a4abb078-fc2e-4ce9-bdd0-d781861d0815", "queue_name"=>"default", "priority"=>nil, "arguments"=>[], "executions"=>0, "locale"=>"en"}])

vs resque-schedule from static schedule:

*** got: (Job{default} | BrainTickJob | [])

[canted]

I'm seeing the same issue. I did get it to work with these settings, as an ugly workaround

job = HelloJob.new
Resque.schedule = {
  hello: {
    every: "10s",
    class: "ActiveJob::QueueAdapters::ResqueAdapter::JobWrapper",
    queue: :default,
    args: {"job_class"=>job.class.to_s, "job_id"=>job.job_id, "queue_name"=>job.queue_name, "priority"=>nil, "arguments"=>[], "locale"=>"en"},
    description: "Hello World"
  },
}

[carsonreinke]

@sevos Any chance you can run that with the VERBOSE on? There should be no reason why either should not work.

[TimFletcher]

I was having the exact same problem as @sevos and @Babdus. Using the following worked for me:

CopyClickStatsJob:
  every: '10s'
  class: ActiveJob::QueueAdapters::ResqueAdapter::JobWrapper
  queue: default
  description: "Migrate click stats from the write table to the read table"
  args:
    job_class:  CopyClickStatsJob
    queue_name: default
    arguments: []

[mickaelmagniez]

@TimFletcher solution work for me too

[glebm]

resque-scheduler still doesn't work with ActiveJob out of the box (despite the issue being closed). Resque itself does work with ActiveJob out of the box.

Update: There is a gem that works around the compatibility issue: https://github.com/JustinAiken/active_scheduler

[jpcuevaslavin]

@TimFletcher solution worked for me too. Thanks!

[stuartluscombe]

The solution from @TimFletcher worked for me also.

I tried the active-scheduler gem but I ran in to issues with resque-scheduler not initialising correctly.

[MartinSugasti]

@TimFletcher your code worked for me. Thanks.

[justrenato]

@TimFletcher worked for me too. Cant believe that's still happenning ;-;

[jronallo]

If instead of a scheduled job you're using enqueue_at_with_queue you can do something like this which includes arguments:

time_to_send = Time.now + 120
Resque.enqueue_at_with_queue(
        :mail_updates,
       time_to_send,
        ActiveJob::QueueAdapters::ResqueAdapter::JobWrapper,
        {
          job_class: SubscriptionsJob,
          job_id: SecureRandom.uuid,
          queue_name: :mail_updates,
          priority: nil,
          arguments: [{arg1: "value", another_arg: "value2"}],
          locale: 'en'
        }
      )