[torqueusers] torque setting ulimit -l

[Jan Ploski]

torqueusers-bounces at supercluster.org schrieb am 10/31/2007 04:35:29 AM:

> Hi,
> 
> Experiencing an odd issue that I can seemingly only blame on torque.
> Would love some assistance.
> 
> [brett at tango ping]$ ssh tango010 ulimit -l
> unlimited
> [brett at tango ping]$ qsub -l nodes=tango010:ppn=2 -l walltime=0:5:0 -I
> qsub: waiting for job 22605.tango-m.vpac.org to start
> qsub: job 22605.tango-m.vpac.org ready
> 
> [brett at tango010 ~]$ ulimit -l
> 32
> [brett at tango010 ~]$ grep -i unlimited /etc/security/limits.conf
> * soft memlock unlimited
> * hard memlock unlimited
> 
> Why is torque limiting the memlock so low, and how can I stop it from
> doing this?
> 
> what is even stranger:
> 
> [brett at tango ping]$ qsub -l nodes=tango001:ppn=2 -l walltime=0:5:0 -I
> qsub: waiting for job 22606.tango-m.vpac.org to start
> uqsub: job 22606.tango-m.vpac.org ready
> 
> [brett at tango001 ~]$ ulimit -l
> unlimited
> 
> why is it only doing it on SOME nodes ?
> I'm quite baffled by this.

Check/set ulimit in the script from which pbs_mom is started - not just 
/etc/security/limits.conf.

Regards,
Jan Ploski