[torquedev] Re: [torqueusers] Bizzare $PBS_O_WORKDIR bug

Dave Jackson jacksond at clusterresources.com
Tue Sep 4 15:32:59 MDT 2007


Chris,

  This issue was fixed some time ago in Moab.  Please feel free to
upgrade to any of the more recent patch releases and let us know if you
see any further issues.  We did validate this once again against Moab
5.1.0p6 and found that all behaved properly.

Thanks,
Dave

On Tue, 2007-09-04 at 17:29 +1000, Chris Samuel wrote:
> On Tue, 4 Sep 2007, Garrick Staples wrote:
> 
> > Which scheduler are you using?  If using moab, kill it, submit the
> > "bad" job, look at 'qstat -f $jobid', start the scheduler, and look
> > at 'qstat -f $jobid' again and compare the difference.
> 
> Bingo - good call Garrick!
> 
> Without Moab, and with Moab started in -P mode (to begin paused) we
> have this at the end of the Variable_List:
> 
>         PBS_O_WORKDIR=/home/csamuel/tmp/tango022,PBS_O_QUEUE=run_1_day
> 
> As soon as Moab starts the job that gets corrupted to:
> 
>       PBS_O_WORKDIR=/home/csamuel/tmp/tango022UEUE=run_1_day;NODES=tango022
> 
> Moab is 5.1.0p4 on AMD64, compiled for libtorque thus:
> 
> [root at tango-m ~]# ldd /usr/local/sbin/moab
>         libpthread.so.0 => /lib64/libpthread.so.0 (0x0000003a11400000)
>         libm.so.6 => /lib64/libm.so.6 (0x0000003a11000000)
>         libtorque.so.0 => /usr/local/torque-2.1.8/lib/libtorque.so.0 (0x00002aaaaaabe000)
>         libc.so.6 => /lib64/libc.so.6 (0x0000003a10800000)
>         /lib64/ld-linux-x86-64.so.2 (0x0000003a10400000)
> 
> cheers,
> Chris
> _______________________________________________
> torquedev mailing list
> torquedev at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torquedev



More information about the torqueusers mailing list