Hello Josef!
we have also looked at this (in our case it was AFS) and also had the
the impression that HPC and kerberized authentication are not an
optimal combination.
If users need to access AFS, we recommend our users to
create a keytab file in their home (and check the permissions before)
see
Hi!
I tried it (I normally would use sacct and ssh) but even with "srun bash
--login"
the "id" command gives me all groups. And we are using rfc2307bis.
And I don't think that slurm stores any group information. So it seems
to be the way the session starts on your node. But at this point
no ide
Hello!
this is a simple wrapper for sacct which prints the
output from sacct as table. So you can make a
"sacctml -j foo --long" even without two 8k displays ;-)
cheers#!/usr/bin/python3
def sacct_command(cmd):
cmd_list=["sacct","--parsable2"]
cmd_list.extend(cmd)
result=subproces
Hello!
I want to archive old parts of our Database, but I don't
understand the paragraph "archive functionality" in the
man-page from sacctmgr.
I want to archive everything older than one year
I guess I have to execute
sacctmgr archive dump Directory=/foo Events Jobs Steps Suspend
PurgeEventA
Hello!
a simple way is to set MailProg in slurm.conf like
MailProg = /usr/local/bin/myscript.sh
and write a simple script
#! /bin/bash
/bin/mail -r "SLURM Daemon " "$@"
best regards
> Von: "Andre Torres"
> An: slurm-users@lists.schedmd.com
> Gesendet: Donnerstag, 24. Januar 2019 17:10
Hello!
When I started getting angry with jobs that required a lot of cpu resources,
and then after that,
use only a fraction of it I wrote a small script which gives you a simple
report of user,partiton, group or
accounts ordered by the average efficieny or the amount of wasted cpu seconds
of
