<div dir="ltr">I've not parsed your message very far, but...<div><br></div><div>for i in `cat limit_users` ; do </div><div><br></div><div>sacctmgr where user=$i partition=foo account=bar set grptresrunmins=cpu=Nlimit</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Dec 16, 2021 at 6:01 PM Ross Dickson <<a href="mailto:ross.dickson@ace-net.ca">ross.dickson@ace-net.ca</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr">It would like to impose a time limit stricter than the partition limit on a certain subset of users. I should be able to do this with a QOS, but I can't get it to work. What am I missing?<br><br>At <a href="https://slurm.schedmd.com/resource_limits.html" target="_blank">https://slurm.schedmd.com/resource_limits.html</a> it says, <br>"Slurm's hierarchical limits are enforced in the following order ...:<br><br>1. Partition QOS limit<br>2. Job QOS limit<br>3. User association<br>4. Account association(s), ascending the hierarchy<br>5. Root/Cluster association<br>6. Partition limit<br>7. None<br><br>Note: If limits are defined at multiple points in this hierarchy, the point in this list where the limit is first defined will be used." <div><br></div><div>And there's a little more later about the Partition limit being an upper bound on everything.<br><br>This says to me that if:<br>* there is a large time limit on a partition,<br>* there is a smaller time limit on the job QOS, and<br>* the partition has no associated QOS,<br>then the MaxWall on the Job QOS should have effect. <br><br>But that's not what I observe. I've created a QOS 'nonpaying' with MaxWall=1-0:0:0, and set MaxTime=7-0:0:0 on partition 'general'. I set the association on user1 so that their job will get QOS 'nonpaying', then submit a job with --time=7-0:0:0, and it runs: <br><br>$ scontrol show partition general | egrep 'QoS|MaxTime'<br> AllocNodes=ALL Default=YES QoS=N/A<br> MaxNodes=UNLIMITED MaxTime=7-00:00:00 MinNodes=0 LLN=NO MaxCPUsPerNode=UNLIMITED<br>$ sacctmgr show qos nonpaying format=name,flags,maxwall<br> Name Flags MaxWall<br>---------- -------------------- -----------<br> nonpaying 1-00:00:00<br>$ scontrol show job 33 | egrep 'QOS|JobState|TimeLimit'<br> Priority=4294901728 Nice=0 Account=acad1 QOS=nonpaying<br> JobState=RUNNING Reason=None Dependency=(null)<br> RunTime=00:00:40 TimeLimit=7-00:00:00 TimeMin=N/A<br>$ scontrol show config | grep AccountingStorageEnforce<br>AccountingStorageEnforce = associations,limits,qos<br clear="all"><div><br></div><div>Help!?</div><div><br></div>-- <br><div dir="ltr"><div dir="ltr"><div>Ross Dickson, Computational Research Consultant<br></div>ACENET -- Compute Canada -- Dalhousie University<br></div></div></div></div>
</blockquote></div>