Ticket 737

Summary:	More detail in emails from slurm
Product:	Slurm	Reporter:	Josko Plazonic <plazonic>
Component:	Other	Assignee:	Moe Jette <jette>
Status:	RESOLVED FIXED	QA Contact:
Severity:	5 - Enhancement
Priority:	---	CC:	da
Version:	14.03.0
Hardware:	Linux
OS:	Linux
Site:	Princeton (PICSciE)	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:	---
Machine Name:		CLE Version:
Version Fixed:	14.03.2	Target Release:	---
DevPrio:	---	Emory-Cloud Sites:	---

Description Josko Plazonic 2014-04-22 07:33:24 MDT

Our users have been complaining that they do not get much info from slurm's emails.  E.g. this is what a user reported:

The emails from slurm are much less informative, with empty body and a
title like "SLURM Job_id=82771 Name=ovlp-gf-P12-7 Ended, Run time
13:13:32". More importantly, it does NOT report the fact that the job
actually failed, as shown in log:
slurmstepd: Job 82771 exceeded memory limit (41943176 > 41943040), being
killed
slurmstepd: Exceeded job memory limit
slurmstepd: *** JOB 82771 CANCELLED AT 2014-04-22T14:19:04 ***
slurmstepd: Exceeded step memory limit at some point. Step may have been
partially swapped out to disk.

They would like at the very least job status/exit code or something that they can look at and see which jobs need a 2nd look - ideally as much as possible.

I don't see any way of achieving that as is, any suggestions or workarounds?

Comment 1 David Bigagli 2014-04-22 07:36:01 MDT

Josko let me look into this and get back to you.

David

Comment 2 Moe Jette 2014-04-25 09:02:05 MDT

I've added a job's exit state (COMPELTED, FAILED, NODE_FAIL, etc.) plus its exit code to the email message, for example:

SLURM Job_id=200 Name=tmp Failed, Run time 00:00:01, FAILED, ExitCode 3

This change will be in v14.03.2 when released, probably next week.

https://github.com/SchedMD/slurm/commit/e8dce1a7ce6d812a8ca0c52a9d4e76e1a6f2c02e

Comment 3 Josko Plazonic 2014-04-28 01:23:34 MDT

Perfect, thanks!