-
Type: Task
-
Status: Done (View Workflow)
-
Priority: Normal
-
Resolution: Done
-
Component/s: pfs_instdata
-
Labels:None
Slightly adapt drpActor config for current (2023-12) usage on pfsa-usr01:
- change logging directory to allow running as `pfs`. We currently run a hacked pfs_instdata pointing to `/work/mhs/logs`: make that official
- lower the number of processes in the process pool: that is being broken by the kernel's oom-killer. Currently 16. Certainly drop to 12, maybe have to go lower until we understand why this is happening. The individual drpActor process VM and RSS sizes are tiny compared to the jupyter sizes, but the drpActor processes are the ones being killed.