Ticket 4896

Summary:	scontrol "Invalid node state specified" when using x[xxx,xxx] form of host name
Product:	Slurm	Reporter:	Jenny Williams <jennyw>
Component:	User Commands	Assignee:	Director of Support <support>
Status:	RESOLVED INFOGIVEN	QA Contact:
Severity:	4 - Minor Issue
Priority:	---	CC:	tim
Version:	17.02.3
Hardware:	Linux
OS:	Linux
Site:	University of North Carolina at Chapel Hill	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:		Target Release:	---
DevPrio:	---	Emory-Cloud Sites:	---

Description Jenny Williams 2018-03-09 14:22:16 MST

[root@longleaf-sched slurmmon]# scontrol -ddd update nodename=b1016,c0812,b[1021,1029],c[0507,0535,0539],b1021 state=idle
slurm_update error: Invalid node state specified
[root@longleaf-sched slurmmon]# scontrol -ddd update nodename=b1016 state=idle                                        [root@longleaf-sched slurmmon]# scontrol -ddd update nodename=c0812,b[1021,1029] state=idle     
slurm_update error: Invalid node state specified
[root@longleaf-sched slurmmon]# scontrol -ddd update nodename=c0812 state=idle             
[root@longleaf-sched slurmmon]# scontrol -ddd update nodename=b[1021,1029] state=idle     
slurm_update error: Invalid node state specified

Comment 2 Isaac Hartung 2018-03-12 11:19:58 MDT

Hi Jenny,

I'm looking into your problem and will have something shortly.

--Isaac

Comment 3 Jenny Williams 2018-03-12 11:22:15 MDT

Isaac –

I have not been able to reproduce the issue since; lets just close this one for now.
Thanks!
Jenny

From: bugs@schedmd.com [mailto:bugs@schedmd.com]
Sent: Monday, March 12, 2018 1:20 PM
To: Williams, Jenny Avis <jennyw@email.unc.edu>
Subject: [Bug 4896] scontrol "Invalid node state specified" when using x[xxx,xxx] form of host name

Comment # 2<https://bugs.schedmd.com/show_bug.cgi?id=4896#c2> on bug 4896<https://bugs.schedmd.com/show_bug.cgi?id=4896> from Isaac Hartung<mailto:ihartung@schedmd.com>

Hi Jenny,



I'm looking into your problem and will have something shortly.



--Isaac

________________________________
You are receiving this mail because:

  *   You reported the bug.

Comment 4 Isaac Hartung 2018-03-12 11:30:45 MDT

OK, if you do experience this problem in the future it may be explained by the following:

I've lifted the these excerpts from the scontrol man page:
______________________________________________________________________

"RESUME" is not an actual node state, but will change a node state from "DRAINED", "DRAINING", "DOWN" or "REBOOT" to either "IDLE" or "ALLOCATED" state as appropriate. 

While all of the above states are valid, some of them are not valid new node states given their prior state.
______________________________________________________________________

We typically set a node's state to RESUME when we want it to return to idle or allocated, though IDLE also works.  We suspect that the nodes that you are having trouble updating are not in one of the valid prior states listed above.

Regards