Sidekiq "ghost jobs" messing up unique check

[m...@salesprodigy.com]

Hi there,

First off: huge fan of Sidekiq, have been using it for the last couple of months and things have been pretty smooth. Thanks!

I am running into one issue though. One of my workers is crashing for some reason and is leaving a bunch of jobs set to "running" (Listed in Jobs on the Busy page of Sidekiq Web and also listed under the process-workers key in Redis). 

When I restart the worker, these ghost jobs do not disappear, but just sit there and I am pretty sure there are causing new jobs with identical arguments not to get run. What is the event in Sidekiq that would clear these jobs out?

Thanks!

Mik

[Mike Perham]

Hi Mik,

Sidekiq does not have any restrictions on jobs running with the same arguments.  If you are using some plugin like a unique plugin to get this behavior, you should open an issue with that project.

If your Sidekiq process has crashed, jobs listed on the Busy page will stay there for up to one minute.  After that time, the job data should disappear.

Mike

--
You received this message because you are subscribed to the Google Groups "Sidekiq" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sidekiq+u...@googlegroups.com.
To post to this group, send email to sid...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sidekiq/18a02aa6-ff31-4507-a5c8-5f980e7bc24e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Mike Perham – CEO, Contributed Systems

Smart, effective open source infrastructure for your apps.

[m...@salesprodigy.com]

Hey Mike, thanks for the reply.

Let me dig on that side!

Thanks,

Mik

[m...@salesprodigy.com]

Quick comment: what is process that would run after the minute? It keeps my Jobs on the Busy page for at least a couple of hours, and only clears them when I restart. My jobs crash with a full on Seg Fault though...

Mik

On Tuesday, June 30, 2015 at 8:50:38 AM UTC-7, Mike Perham wrote:

[Mike Perham]

The data in Redis has a TTL of 60 seconds.  The TTL is refreshed by the process heartbeat.  Are you using Heroku?  If so, the process heartbeat may never stop because Heroku reuses process names ("worker.1") and so that data would never be removed.

To view this discussion on the web visit https://groups.google.com/d/msgid/sidekiq/5e00d805-f6a6-4c75-a39b-ff78fb7eac91%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.