Ticket 721

Summary:	mismatch between mysql password and munge socket
Product:	Slurm	Reporter:	vecsys.msf
Component:	Accounting	Assignee:	David Bigagli <david>
Status:	OPEN ---	QA Contact:
Severity:	6 - No support contract
Priority:	---	CC:	tevend
Version:	2.6.7
Hardware:	Linux
OS:	Linux
Site:	-Other-	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:
Target Release:	---	DevPrio:	---
Emory-Cloud Sites:	---

Description vecsys.msf 2014-04-15 21:14:20 MDT

Hi
I setup the mysql user slurm password to PASSWORD.
When running sbatch, Munge uses the word PASSWORD as the communication socket !!!
To make thing works I set mysql slurm user password to "/var/run/munge/munge.socket.2", AccountingStoragePass to the same value and everything works fine now. Mysql & Munge are happy.
Regards
Marc


config:
AccountingStorageType=accounting_storage/mysql
AccountingStorageHost=msfdev-msf.AAA.dmz
AccountingStorageEnforce=associations,limits,qos
AccountingStorageUser=slurm
AccountingStoragePass=PASSWORD

command
sbatch -o /home/MSF/slurm/%j.log -e /home/MSF/slurm/%j.log -M msf -A transcription  /home/MSF/slurm/test.sh
sbatch: error: Munge encode failed: Failed to access "PASSWORD": No such file or directory (retrying ...)
sbatch: error: Munge encode failed: Failed to access "PASSWORD": No such file or directory (retrying ...)
sbatch: error: Munge encode failed: Failed to access "PASSWORD": No such file or directory
sbatch: error: authentication: Socket communication error
sbatch: error: Batch job submission failed: Protocol authentication error

Comment 1 David Bigagli 2014-04-25 06:41:58 MDT

Cannot reproduce.

David

Comment 2 Steven DuChene 2021-07-26 22:26:53 MDT

I am able to reproduce this exact condition.
I have to set my StoragePass=/var/run/munge/munge.socket.2 in slurmdb.conf and and AccountingStoragePass=/var/run/munge/munge.socket.2 in slurm.conf the user password in mariadb to the same value to get sacct to function.

If I set these two value in the conf files to an actual password that matches the user password in mariadb I get the following error when I run sacct

lockhart-login1:~ # sacct
sacct: error: If munged is up, restart with --num-threads=10
sacct: error: Munge encode failed: Failed to access "PASSWORD": No such file or directory
sacct: error: slurm_send_node_msg: g_slurm_auth_create: REQUEST_PERSIST_INIT has authentication error: Invalid authentication credential
sacct: error: slurm_persist_conn_open: failed to send persistent connection init message to localhost:6819
sacct: error: Sending PersistInit msg: Protocol authentication error
sacct: error: Problem talking to the database: Protocol authentication error