So I narrowed down the issue, it happends when a worker detects he should be the ticker and does an update on all other workers to set them false, an oracle error happens (about deadlock) and it fails, the error "Error coordinating ticker" appears and then it continues, but lines later it COMMITS and makes itself a ticker... and everyone wants to be a ticker and scheduler dies and fire starts and i suffer in pain, so the fix should be removing the catch that prints the message "Error coordinating ticker" so that when the whole operation fails it rolls back and not commits itself to be a ticker.. anyways issue is open now:
https://github.com/web2py/web2py/issues/1787 El miércoles, 4 de octubre de 2017, 8:49:22 (UTC-6), Boris Aramis Aguilar Rodríguez escribió: > > Hi, I've been recently using lots of tasks scheduled on a server so I had > to use several workers to deal with the queue, currently I'm using 45 > workers. > > Somewhere between 12 hours after starting the 45 workers, suddenly tasks > start accumulating in the queue and workers are alive (they report a > heartbeat), and the only thing i've noticed is that from the 45 workers > more than 1 becomes the ticker (the flag is_ticker is True for several of > them), so if I discovered that if I set to False the flag for all workers > except for one then tasks start being assigned again correctly, but it > happends again after some time. > > I'm not sure how to find where the bug is, so that is why I come here > asking for your guidance :) > > Thanks! > -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
