[slurm-users] Effect of slurmctld and slurmdb going down on running/pending jobs

Josef Dvoracek jose at fzu.cz
Thu Jun 24 09:16:20 UTC 2021


just set the partitions to "DOWN" to avoid unexpected behavior for users 
and reboot slurm(ctl|dbd)+sql box. Running jobs are from my experience 
not affected.
No need to drain nodes.


On 24. 06. 21 0:54, Amjad Syed wrote:
> Hello all
> We have  a cluster  running centos  7 . Our slurm  scheduler is 
> running on a vm  machine and  we are running out  of disk  space for /var
>  The slurm innodb is taking most of space.  We intend to expand the 
> vdisk for slurm server. This will require a reboot  for changes to 
> take  effect.  Do we have to stop users  submitting  jobs by draining 
> all partitions and then restart the server. That is slurmctld.slurmdb 
> and mariadb? Or  will the restarting of slurm vm have  no effect on 
> running/pending iobs?
> Sincerely
> Amjad

Josef Dvoracek
Institute of Physics | Czech Academy of Sciences
cell: +420 608 563 558 | https://telegram.me/jose_d | FZU phone nr. : 2669

More information about the slurm-users mailing list