Ticket 14736

Summary:	22.05/21.08 srun running against 20.11 slurmd - "io_init_msg_unpack: unpack error"
Product:	Slurm	Reporter:	Tim Wickberg <tim>
Component:	User Commands	Assignee:	Tim Wickberg <tim>
Status:	RESOLVED WONTFIX	QA Contact:
Severity:	3 - Medium Impact
Priority:	---	CC:	bart
Version:	21.08.8
Hardware:	Linux
OS:	Linux
Site:	SchedMD	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 Tim Wickberg 2022-08-11 10:15:15 MDT

21.08/22.05 srun fails against a 20.11 slurmd with:

srun: job 16426 queued and waiting for resources
srun: job 16426 has been allocated resources
srun: error: io_init_msg_unpack: unpack error
srun: error: io_init_msg_read_from_fd: io_init_msg_unpack failed: rc=-1
srun: error: failed reading io init message

Comment 3 Tim Wickberg 2022-08-11 11:30:02 MDT

Unfortunately there's no way to readily restore this cross-version communication when dealing with 20.11. There were some structural changes to the I/O initialization going into 21.08 that caused this.

Going forward, however, this type of cross-version support is working correctly, and will be maintained. We've verified that a 22.05 srun vs. 21.08 slurmd/slurmstepd is working correctly, and will work to ensure this is not broken again in the future.