On 12/05/15 16:13, Ludovic Prevost wrote:
Hi,
Exact, could you share your slurm.conf, cgroup.conf (if use), at least an
archive of /etc/slurm/*.
Thank you by advance,
Ludovic PREVOST
________________________________________
From: Veronique Legrand [[email protected]]
Sent: Tuesday, May 05, 2015 10:30 AM
To: slurm-dev
Subject: [slurm-dev] Re: since version 14.11.6 srun takes 2 cpu by default
On 04/05/15 15:28, Ludovic Prevost wrote:
Hi,
Could you please share you configuration ?
Best Regards,
Ludovic
On 04/29/2015 02:22 PM, Veronique Legrand wrote:
Hello,
I am running the slurm non regression tests on a virtual machine that
has 2 CPU. This is how I noticed this strange behavior.
test1.69 fails because each srun job launched by it take 2 CPU by
default (I checked with sacct when the jobs were over). Of course,
because of that, 1 of the 2 jobs stays pending (waiting for
ressources) until the other has completed. The aim of test1.69 is to
check that 2 srun commands can run simultaneously so, of course, it
fails on my VM.
If I want srun to take only 1 cpu, I have to indicate explicitely
"srun -n 1 myscript.sh".
I am wondering if I possibly missed something in my slurm
configuration/installation or if this is a bug introduced in the new
version.
I checked on our test cluster that is running under slurm 14.11.5 and
srun takes 1 CPU by default.
Any hint is welcome.
Kind regards,
Véronique
Hello Ludovic,
Thank you for your interest. What configuration files would be useful to
you (apart from slurm.conf) of course?
Kind regards,
Véronique
--
Véronique Legrand
IT engineer - Scientific computing
IT center for biology - C3BI
Institut Pasteur, Paris
mail: [email protected]
Tel: 01 44 38 95 03
Hi Ludovic,
Please find attached my slurm.conf file. I have no cgroup.conf and there
is no slurm directory under /etc.
Best regards,
Véronique
--
Véronique Legrand
IT engineer - Scientific computing
IT center for biology - C3BI
Institut Pasteur, Paris
mail: [email protected]
Tel: 01 44 38 95 03
# 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.
#
ControlMachine=XXXXXX-CentOS
#ControlAddr=
#
#MailProg=/bin/mail
MpiDefault=none
#MpiParams=ports=#-#
ProctrackType=proctrack/pgid
ReturnToService=1
SlurmctldPidFile=/var/run/slurmctld.pid
#SlurmctldPort=6817
SlurmdPidFile=/var/run/slurmd.pid
#SlurmdPort=6818
SlurmdSpoolDir=/var/spool/slurmd
SlurmUser=slurm
#SlurmdUser=root
StateSaveLocation=/var/spool
SwitchType=switch/none
TaskPlugin=task/none
#
#
# TIMERS
#KillWait=30
#MinJobAge=300
#SlurmctldTimeout=120
#SlurmdTimeout=300
#
#
# SCHEDULING
FastSchedule=1
SchedulerType=sched/backfill
#SchedulerPort=7321
SelectType=select/linear
#
#
# LOGGING AND ACCOUNTING
## AuthType=none
AccountingStorageType=accounting_storage/slurmdbd
AccountingStorageEnforce=associations,limits,qos
ClusterName=cluster_montdor
JobAcctGatherFrequency=30
JobAcctGatherType=jobacct_gather/linux
#SlurmctldDebug=3
SlurmctldLogFile=/home/slurm/slurm_logs
#SlurmdDebug=3
SlurmdLogFile=/home/slurm/slurm_logs
#
#
# COMPUTE NODES
NodeName=XXXXXX-CentOS CPUs=2 RealMemory=3000
#NodeName=slurm-slave-1 CPUs=4 State=UNKNOWN
PartitionName=partition_test Nodes=XXXXXX-CentOS Default=YES MaxTime=INFINITE
State=UP DefMemPerCPU=400 MaxMemPerCPU=500
