I still get a error on some very long renders (not the 5 min bug). And no error reports on monitor. If the exception was at least pass to the log, it would be easier to debug.
Deadline 7.0.0.44R
2014-11-07 07:49:28: 0: INFO: Prepass 3 of 3…: done [00:03:13.7]
2014-11-07 09:19:07: Scheduler Thread - Task “0_1-1” could not be found because task has been modified:
2014-11-07 09:19:07: current status = Rendering, new status = Rendering
2014-11-07 09:19:07: current slave = 4arq-17, new slave = 4arq-20
2014-11-07 09:19:07: current frames = 1-1, new frames = 1-1
2014-11-07 09:19:07: Scheduler Thread - Cancelling task…
2014-11-07 09:19:09: 0: In the process of canceling current task: ignoring exception thrown by PluginLoader
2014-11-07 09:19:10: Scheduler Thread - In the process of canceling current tasks: ignoring exception thrown by render thread 0
Best regards,
Daniel
It looks like the task was requeued at some point, which could very well still be that bug. The bug wasn’t that the task would be requeued immediately after 5 minutes, it was that it could possibly be requeued any time after 5 minutes. Can you double check that all of your client machines are running 7.0.0.44 (including workstations and Pulse, if you’re running it)? All it takes is one older version to do housecleaning, resulting in the task being requeued.
Are you guys currently running Pulse?
Cheers,
Ryan
Thanks Ryan, we have Pulse running, manually installed using the lastest 44 installer. I will double check if there is a slave that i didn’t installed manually.
Best regards,
Daniel
You can check the slave list in the Monitor. There is a Version column so you can see if any slaves are running different versions.
They are all 44, due to automatic upgrades… but some were not really on version 44 (there was some missing files). I reinstalled them, lets see if we got no more task drops. I also enabled verbose logging so i can see what is happening (i hope), when the pluginloader gives an error.
Best regards,
Daniel
Hi Ryan, just to confirm that there were no problems during weekend renders. So it must have been some nodes that were not fully upgraded that cause the problem.
Thanks for the help 
.
Best regards,
Daniel
