We are currently running Qube as our Render manager and have a small (96 cores) render farm + workstations at night. The farm is running Windows and we’re running Maya and Mental Ray renders (CMDRANGE).
We have started having issues with “File not found” errors and failing frames. IT seems to work OK for some frames and then fail again for others…not limited to any particular machine or machine type, job type or anything else I can find. I have checked out and run some tests on my network…all seems to be OK there. I have made sure that all machines are syncing clocks with NTP. I sent in diags and PIRs for the BlueArc Mercury…all checked out OK…I’m now totally at a loss as to where to look. One of my friends mentioned that they had similar issues on their linux farm and it turned out to be the limit on number of open files. PipelineFX assures me that that only happens in linux.
Anybody seen anything similar or have any suggestions where else I can look?
Terry