We're currently experiencing some issues with "endless" work item cooking with a fairly simple PDG rendering setup in Solaris.
We are using Renderman v25.1 and a PDG network inside Solaris that's setting off a USD Render ROP via a RopFetch node.
The specific problem we're facing is that ropfetch work items seem to get stuck in an endless "cooking" state, even after the frame has been rendered and written to disk.
Unfortunately this issue has proven difficult to replicate consistently and appears to occur sporadically, leading to varying results each time.
We have tried giving the task more time to complete, even 40 hrs at one point, for a frame that'd normally complete in 5/10 minutes.
On the work items that endlessly cook, we get an ALF_PROGRESS 100% output, along with memory statistics. The last lines in the output are "Running postframe procedurals script...." and "Running postrender procedurals script....".
On the work items that do "complete", we get the same, but with additional PDG_RESULT and PDG_SUCCESS lines.
We're wondering if anyone has encountered similar issues in the past and could offer insights or potential solutions to try?
Also if anyone is able provide advice on what might be causing work items to remain incomplete, or generally cook endlessly, it would be greatly appreciated
Thanks for your time!