[slurm-users] How to submit simple job on Master and Compute nodes

[Loris Bennett]

Hi Zainul,

Zainul Abiddin <zainul1114 at gmail.com> writes:

> Hi All,
>
> Please help me to submit a simple job on master and compute nodes.
> Here is the my commands
>
> [root at smaster ~]# sinfo -Nl
> Thu Feb 04 16:54:58 2021
> NODELIST   NODES PARTITION       STATE CPUS    S:C:T MEMORY TMP_DISK WEIGHT AVAIL_FE REASON
> smaster        1      hpc*        idle 4       4:1:1   1024        0      1   (null) none
> snode          1     debug        idle 4       4:1:1   1024        0      1   (null) none
> [root at smaster ~]# srun -n 4 -N 1 --mem=100 --time 0:0:10 /bin/hostname
> smaster
> smaster
> smaster
> smaster
> [root at smaster ~]# srun -n 4 -N 2 --mem=100 --time 0:0:10 /bin/hostname
> srun: Requested partition configuration not available now
> srun: job 25 queued and waiting for resources
> srun: Job allocation 25 has been revoked
> srun: Force Terminated job 25
> [root at smaster ~]#

This job probably can't start, because you only have one node in the
default partition 'hpc' but the job requests two.

> [root at smaster ~]# cat /etc/slurm/slurm.conf
> # slurm.conf file generated by configurator easy.html.
> # Put this file on all nodes of your cluster.
> # See the slurm.conf man page for more information.
> #
> .
> .
> .
> #
> # COMPUTE NODES
> NodeName=smaster NodeAddr=192.168.1.195 CPUs=4 RealMemory=1024 State=UNKNOWN
> NodeName=snode NodeAddr=192.168.1.196 CPUs=4 RealMemory=1024 State=UNKNOWN
> PartitionName=debug Nodes=snode Default=YES MaxTime=INFINITE State=UP
> PartitionName=hpc Nodes=smaster Default=YES MaxTime=INFINITE State=UP
>
> [root at smaster ~]#
>
> and how do I configure a single partition for two nodes.

 PartitionName=hpc Nodes=smaster,snode Default=YES MaxTime=INFINITE State=UP

Cheers,

Loris

-- 
Dr. Loris Bennett (Hr./Mr.)
ZEDAT, Freie Universität Berlin         Email loris.bennett at fu-berlin.de