[torqueusers] Re: [Beowulf] job runs with mpirun on a node but not if submitted via Torque.

Rahul Nabar rpnabar at gmail.com
Wed Apr 1 10:40:42 PDT 2009


On Wed, Apr 1, 2009 at 9:53 AM, Ling C. Ho <ling at fnal.gov> wrote:
> We had a problem with resources_max.pmem accidentally set too low for the
> Torque queue, and the user login shell was getting segfault. Torque showed
> Exit_status of 267.
>

I think it is a ulimt issue. Many thanks to Don Holmgren who pointed
me to the solution. I've gotten it working now but am still testing
more. Things seem sensitive to where exactly I put the ulimit
directive.

{The stack was reported as unlimited but apparently I still need a directive}

Thanks for the leads guys!

-- 
Rahul



More information about the Beowulf mailing list