rq-scheduler Multischedulers don't work

It seems that now you can't run two schedulers on two servers for high availability.

This happens because before the call of acquire_lock() it still runs register_birth() which raises ValueError("There's already an active RQ scheduler"). Initially register_birth() and register_death() were deleted in #143, but then they were reverted back for some reason.

Sep 07 '17 06:09 yetanotherape

True. Also the current locking mechanism doesn't seem to be as robust and may give rise to a possibility of multiple schedulers acquiring the lock.

This is a very important feature for us as we're hoping to run the scheduler on multiple servers for a failover. (Also makes our deployment easier as each server remain exactly identical)

Oct 30 '18 09:10 oxalorg

We are experiencing same issue in our rolling updates, so I'm just wondering if there is progress/work being made on this topic?

Feb 24 '20 20:02 jsam

@jsam This has been fixed by #212 (which is now merged to master). Could you perhaps give it a try and let us know if it solved your use case?

Apr 24 '20 06:04 oxalorg