PipelineFX Forum
Qube! => General => Topic started by: rtaege on February 25, 2014, 05:42:26 PM
-
I have a farm of 10 render nodes. I usually submit jobs so that each node will dedicate the entire worker to each frame.
But with one particular node, it doesn't work, it will try to use each core to render a frame, so that I end up with 12 instances of "render10" trying to render multiple frames, - usually just resulting in the worker freezing.
Is there any setting on the indivisual workers that override the submission settings for this? I can't seem to figure out what changed that causes this...
-
Do the other 9 nodes have worker_cpus set to 1, but this node is using the default of 0, which means "auto-discover the number of cores"?
Check the "slots" column in the WrangerView Worker view, and see if this has "x/12" and the others have "x/1", where "x" is any number.
-
This one giving me trouble has 0/0 listed. All the others are 16/16 when they're rendering.
-
worker_cpus are set to 0 on all of them.
-
any other ideas about this? Guess I'll open a support ticket.