Ticket 9036

Summary:	scancel does not honour SLURM_CLUSTERS
Product:	Slurm	Reporter:	Marco <marco.delapierre>
Component:	User Commands	Assignee:	Gavin D. Howard <gavin>
Status:	RESOLVED FIXED	QA Contact:
Severity:	4 - Minor Issue
Priority:	---
Version:	20.02.2
Hardware:	Linux
OS:	Linux
Site:	Pawsey	Slinky Site:	---
Alineos Sites:	---	Atos/Eviden Sites:	---
Confidential Site:	---	Coreweave sites:	---
Cray Sites:	---	DS9 clusters:	---
Google sites:	---	HPCnow Sites:	---
HPE Sites:	---	IBM Sites:	---
NOAA SIte:	---	NoveTech Sites:	---
Nvidia HWinf-CS Sites:	---	OCF Sites:	---
Recursion Pharma Sites:	---	SFW Sites:	---
SNIC sites:	---	Tzag Elita Sites:	---
Linux Distro:	---	Machine Name:
CLE Version:		Version Fixed:	20.02.4
Target Release:	---	DevPrio:	---
Emory-Cloud Sites:	---

Description Marco 2020-05-12 08:27:07 MDT

When working with multiple clusters, it is possible to use SLURM_CLUSTERS to set the cluster to be used with certain commands, e.g. sbatch and squeue.
It would be great if also scancel could make use of this variable, as it would make the user experience more consistent.

As a practical user case, SLURM_CLUSTERS is the only way to use multiple clusters with the popular workflow manager Nextflow. However usability is kind of broken by inability to issue scancel across clusters using the environment variable.

We have a community of tens of users potentially benefitting from this.

Comment 2 Gavin D. Howard 2020-05-13 10:34:51 MDT

Thank you for the report. I am looking into it.

Comment 5 Gavin D. Howard 2020-07-16 16:02:38 MDT

Good news! We have added this to Slurm, and it will be in 20.02.4. I am closing this bug, but feel free to reopen if you have questions about the changes.

Comment 6 Marco 2020-07-17 00:34:39 MDT

Thanks Gavin,

great to hear this feature will be in the next Slurm release!

Sincerely, Marco