"Lipari, Don" <lipa...@llnl.gov> writes: > Raw usage is captured and saved locally to the cluster in the > assoc_usage file in your configured StateSaveLocation. The values > saved there are subject to half life decay based on your > PriorityDecayHalfLife setting. These values can be forcibly reset to > zero by running “sacctmgr modify account <xxx> set RawUsage=0”. > > Usage values saved to the SLURM db do not decay. sreport retrieves > these values from the db to generate its reports. > > I’m not sure what's going on with sprio.
Without the configuration having be changed, sprio now seems to be working. If I had to make a wild guess, I would assume the problem earlier was connected to some slightly pathological initial situation in which the pending jobs were all from one user and would all have had the same priority. Cheers Loris > Don > >> -----Original Message----- >> From: owner-slurm-...@lists.llnl.gov [mailto:owner-slurm- >> d...@lists.llnl.gov] On Behalf Of Loris Bennett >> Sent: Friday, February 10, 2012 6:30 AM >> To: slurm-dev@lists.llnl.gov >> Subject: [slurm-dev] Problem with setting up fairshare >> >> Hi, >> >> I am trying to set up fairshare scheduling and can send details of the >> configuration, but I wondered whether there was a simple explanation of >> the following: >> >> sshare shows "Raw Usage" of zero for all accounts, whereas sreport >> shows >> non-zero CPU usage. Also sprio on its own returns nothing and the >> following if given an existing job ID: >> >> # sprio -j 12345 >> Unable to find jobs matching user/id(s) specified >> >> Any ideas? >> >> Cheers >> >> Loris >> -- >> Dr. Loris Bennett (Mr.) >> ZEDAT, Freie Universität Berlin Email loris.bennett@fu- >> berlin.de >> > -- Dr. Loris Bennett (Mr.) ZEDAT, Freie Universität Berlin Email loris.benn...@fu-berlin.de
