Re: [HTCondor-users] Setting walltime for SLURM job

[Seren,Ãmit <uemit.seren@xxxxxxxxxxxxxx>]

Good news:

I think I found the issue and it was my fault. I made a copy of the old job routes config file (without the set_remote_BatchRuntime setting) inside of the /etc/condor-ce/condor.d folder and this file instead of the modified file was picked up. After removing the copy, the walltime was proplery applied.

 

Best

Ãmit

 

 

From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of Uemit Seren <uemit.seren@xxxxxxxxxxxxxx>
Reply-To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>
Date: Friday, 24. July 2020 at 14:43
To: Brian Lin <blin@xxxxxxxxxxx>, HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>
Subject: Re: [HTCondor-users] Setting walltime for SLURM job

 

Hi Brian,

 

[root@ce-1 config.d]# rpm -q condor htcondor-ce blahp

condor-8.8.8-1.el7.x86_64

htcondor-ce-3.2.1-1.el7.noarch

package blahp is not installed

 

The blahp files seem to come from a different repository:

 

[root@ce-1 config.d]# rpm -qf /usr/libexec/condor/glite/bin/blah_check_config

condor-externals-8.8.8-1.el7.x86_64

 

FYI: We installed following packages from UMD4 repos:

htcondor-ce-client

htcondor-ce-view

htcondor-ce

htcondor-ce-slurm

 

Best

 

 

-- 

Ãmit Seren Msc

HPC Engineer

+4369910269552

Vienna BioCenter (GMI, IMP, IMBA)

 

From: Brian Lin <blin@xxxxxxxxxxx>
Date: Friday, 24. July 2020 at 14:36
To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>, Uemit Seren <uemit.seren@xxxxxxxxxxxxxx>
Subject: Re: [HTCondor-users] Setting walltime for SLURM job

 

Hi Ãmit,

What's the output of `rpm -q condor htcondor-ce blahp`?

Thanks,
Brian

On 7/24/20 2:50 AM, Seren,Ãmit wrote:

Hi,

 

We have a compute element with condor-8.8.8-1/htcondor-ce-3.2.1-1 submitting to a site local SLURM batch system.

Submission and routing of jobs to SLURM work just fine, however we cannot managed to set the walltime for the slurm jobs.

Based on this thread https://www-auth.cs.wisc.edu/lists/htcondor-users/2020-March/msg00011.shtml we have following job route config:

 

 

JOB_ROUTER_ENTRIES @=jre

[

    GridResource = "batch slurm";

    TargetUniverse = 9;

    name = "GRID jobs";

 

    set_default_queue = "grid";

    set_default_maxWallTime = 2880;

    # in 8.8.4 <= this is required. see: https://www-auth.cs.wisc.edu/lists/htcondor-users/2020-March/msg00011.shtml

    set_remote_BatchRuntime = 2880;

    set_default_xcount = 1;

    set_default_maxMemory = 8000;

    set_remote_JobDirectory = Undefined;

]

@jre

 

However the BLAHP generated job submission script for SLURM is missing the -t stanza:

 

#!/bin/bash

# SLURM job wrapper generated by slurm_submit.sh

# on Fri Jul 24 09:23:02 CEST 2020

#

# SLURM directives:

#SBATCH -o /dev/null

#SBATCH -e /dev/null

#SBATCH --mem=8000

#SBATCH -p grid

 

Anybody has some pointers how to properly set the walltime ?

 

Thanks in advance

Best

Ãmit

 

This body part will be downloaded on demand.