Ticket 10120

Summary:	QOS flags=partitiontimelimit does not override default MaxTime of partition
Product:	Slurm	Reporter:	Bom <bom.singiali>
Component:	Accounting	Assignee:	Jacob Jenson <jacob>
Status:	RESOLVED FIXED	QA Contact:
Severity:	6 - No support contract
Priority:	---	CC:	bom.singiali
Version:	19.05.3
Hardware:	Linux
OS:	Linux
Site:	-Other-	Alineos Sites:	---
Atos/Eviden Sites:	---	Confidential Site:	---
Coreweave sites:	---	Cray Sites:	---
DS9 clusters:	---	HPCnow Sites:	---
HPE Sites:	---	IBM Sites:	---
NOAA SIte:	---	OCF Sites:	---
Recursion Pharma Sites:	---	SFW Sites:	---
SNIC sites:	---	Linux Distro:	CentOS
Machine Name:		CLE Version:
Version Fixed:	19.05.03	Target Release:	---
DevPrio:	---	Emory-Cloud Sites:	---

Description Bom 2020-11-03 10:56:12 MST

Dear Team,

I am trying to implement solution provided in:
https://bugs.schedmd.com/show_bug.cgi?id=4681

However, user job is going to pending state with following status:

[root@vicb-submit-01 ~]# squeue -u louis.kuemmerle
             JOBID PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
            679046   icb_gpu  test.sh louis.ku PD       0:00      1 (QOSMaxWallDurationPerJobLimit)


==== Partiton(icb_gpu) max time limit ===

# scontrol show part | grep MaxTime
   MaxNodes=UNLIMITED MaxTime=7-00:00:00 MinNodes=0 LLN=NO MaxCPUsPerNode=UNLIMITED

# qos (icb_long)max time limit is 6 days.

## Goal: to allow specific user to run job for 20 days with partition maxtime overwrite features, without modifying existing partition and qos.

Please advice, thank you for your help.

--
Best Regards
Bomay

Comment 1 Bom 2020-11-03 11:03:14 MST

Also,

Status keep switching between: 

NODELIST(REASON) 
(Priority) <->(QOSMaxWallDurationPerJobLimit)

Comment 2 Bom 2021-05-19 08:05:50 MDT

PartitionTimeLimit flag on qos resolved this issue.