[slurm-users] QoS and Preemption

Greetings, I will try and over simplify what we are trying to achieve in one of our clusters. The current setup has four QoS groups, A, B, C and D. QoS A can use the whole the cluster's resources and preempts over QoS B, C and D. QoS B can use half the cluster's resources and preempts

Re: [slurm-users] scontrol: can't hold by jobname

Hi Jessica, That did the trick. I have submitted a bug against the documentation: https://bugs.schedmd.com/show_bug.cgi?id=4946. Thanks, Loris Jessica Nettelblad writes: > Try this instead: > > scontrol hold name=g09 > > (Line 441: >

Re: [slurm-users] scontrol: can't hold by jobname

Try this instead: scontrol hold name=g09 (Line 441: https://github.com/SchedMD/slurm/blob/master/src/scontrol/update_job.c) On Mon, Mar 19, 2018 at 9:17 AM, Loris Bennett wrote: > Hi, > > The manpage for version 17.02.9 of 'scontrol' says the following about > the

[slurm-users] scontrol: can't hold by jobname

Hi, The manpage for version 17.02.9 of 'scontrol' says the following about the argument for 'hold': The job_list argument is a comma separated list of job IDs OR "jobname=" with the job’s name I am trying to put a number of jobs with the job name "g09" on hold. However, I can't get the

Re: [slurm-users] #SBATCH options as Bash script parameters

Hi Jessie, Jessie Poquérusse writes: > Hello, > > In trying to modularize and genericize my Bash scripts as much as > possible, I was wondering if there was a way to turn #SBATCH options > (mainly walltime, mem, and output and error file directories) into >