For sure! But letting your jobs run under a cgroup is better. You can better
control the resource usage. With cgroups a job for 1 core gets only 1 core.
With linuxproc you can submit a job for 1 core, but then use 20.
Well if linuxproc is good enough, then it's good.
Best regards,
Geert
We are pleased to announce the availability of Slurm version 17.11.4.
This includes roughly 38 fixes made since 17.11.3 was released early
this month.
Slurm can be downloaded from https://www.schedmd.com/downloads.php
- Tim
* Changes in Slurm 17.11.4
==
-- Add fatal
Thanks Geert for your response. I was able to start slurmd by changing
ProctrackType setting to proctrack/linuxproc.
Thanks,
Yogesh Aggarwal
-Original Message-
From: slurm-users [mailto:slurm-users-boun...@lists.schedmd.com] On Behalf Of
Geert Geurts
Sent: Wednesday, February 28, 2018
I think you'll be fine with uncomenting cgroupaautomount=yes in cgoup.conf
file.
Could you try it like that?
Regards,
Geert
From: Yogesh Aggarwal
Sent: Wednesday, February 28, 2018 20:09
To: slurm-users@lists.schedmd.com
Subject: [slurm-users] slurm-17.11.3-2
Hi All,
I am trying to install slurm-17.11.3-2 on Redhat Linux 7.2 system. I have
completed the installation and configuration but not able to start
slurmd.service. Below are the logs from /var/log/slurmd.log. I have also pasted
slurm.conf, cgroup.conf and cgroup_allowed_devices_file.conf for q
