What does condor_status -af Name TotalLoadAvg say? TotalLoadAvg is as measured by the condor classad
Steve
From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of Dimitri Maziuk via HTCondor-users <htcondor-users@xxxxxxxxxxx>
Sent: Friday, November 2, 2018 4:25:56 PM To: htcondor-users@xxxxxxxxxxx Cc: Dimitri Maziuk Subject: [HTCondor-users] kill vs. load avg (was: scheduling delay) Hi everyone,
I put KILL = ( TotalLoadAvg > 4.5 ) on a host, and I am now watching top on it, w/ 32 condor jobs keeping load average in the 30s for the last 5 minutes or so. So, what's wrong with my KILL _expression_? Hopefully I'm missing something glaringly obvious... TIA -- Dimitri Maziuk Programmer/sysadmin BioMagResBank, UW-Madison -- http://www.bmrb.wisc.edu |