Delayed jobs are not being run

[danlo]

Greetings,

I am running resque-scheduler (4.0.0), resque (1.25.2), ruby (2.2.0p0)

I run the resque-scheduler standalone job via

LOGFILE=/[censored]/resque-scheduler.log VERBOSE=true bundle exec resque-scheduler -B

And, my Resque.enqueue_at() jobs are not being executed when they should be:

Do you have any thoughts on how to go about debugging why my jobs are not being executed.

I've tried restarting resque-scheduler several times, but it hasn't executed the new jobs. I have no "cron" like scheduled jobs.

(The first block of jobs, were removed and manually executed)

-daniel

[yaauie]

It looks like the "Job count" column is 0 for times in the past; for some reason that isn't getting cleaned up, but the jobs are getting removed from the delayed queue; looking at the code, I see that error handling in enqueueing a job happens after the job is popped from the delayed queue, and that exceptions raised here are squashed & logged. Do your logs indicate that anything is amiss?

The command-line utility less can be used to inspect your log file; the '+G' flag will cause it to start showing the end-of-file, which may be where the interesting stuff is happening, but you can scroll with the arrow keys:

$ less +G /[censored]/resque-scheduler.log

It's possible that your scheduler process doesn't have access to the job's class, which would cause it to raise an exception, or that something in the job class' before_enqueue_* hook is raising an exception; the logs will tell you more.

Your command above doesn't include --initializer-path or --environment flags, so I assume the process just doesn't know enough about the work to be able to put it on a queue (it needs the class definition for that). If this is the case, you'll need to point it to a ruby script that can load your application (without starting any server processes), at least far enough to have all the relevant job classes loaded.

Can you verify that the jobs are getting placed on their queues?

[danlo]

Ah! That was was the problem

There wasn't an initializer-path.

I like to store individual "types" of data in different redis databases, and resque work is stored in a database other than the default 0

Since there wasn't an initializing piece of code, resque-scheduler never knew to look in the right database.