Hi Navin, navin srivastava <navin.alt...@gmail.com> writes:
> Hi team, Not sure what team you are referring to here - people here tend to be individuals from individual institutions. > i have extracted the %utilization report and found that the idle time > is at the higher end so wanted to check is there any way we can find > the node based utilization? > > it will help us to figure out what are the nodes are unutilized. Are jobs waiting, despite nodes being idle? If so, you need to look at what resources are being requested by the jobs to find out why there are not starting. If jobs are not waiting, then perhaps just not enough jobs are being submitted to fill your cluster. Cheers, Loris -- Dr. Loris Bennett (Mr.) ZEDAT, Freie Universität Berlin Email loris.benn...@fu-berlin.de
