Hi everyone,

I have a job running on a single node with 32 cores (and 32 cpus requested):

# sstat -P -j 124654.batch
JobID|MaxVMSize|MaxVMSizeNode|MaxVMSizeTask|AveVMSize|MaxRSS|MaxRSSNode|MaxRSSTask|AveRSS|MaxPages|MaxPagesNode|MaxPagesTask|AvePages|MinCPU|MinCPUNode|MinCPUTask|AveCPU|NTasks
124654.batch|1405072K|calc06|0|1405072K|1061412K|calc06|0|1061412K|0|calc06|0|0|298-05:39:38|calc06|0|298-05:39:38|1

It looks like MinCPU/AveCPU stopped accumulating exactly at 
298-05:39:38, as if there's some sort of overflow.

I'm running slurm 2.4.4. NEWS doesn't mention relevant changes.
(I have plans to upgrade, but cannot do a test right now).

Reply via email to