Yeah, thanks Hugh.  I realized that the scripts were combined after I
looked at them more closely (and after I sent my message).  Thanks for
fixing the instructions.

I actually have a sort-of-working xcpu cluster with moab/torque and
all, and I am trying to iron out some quirks.  One of the important
ones is that I am still not able to run MPI jobs using the xmvapich
command, which Lucho claims to work.  For me it only works with a
single process but hangs with 2 or more, regardless of whether it is
on a single node or several.

On the moab/torque issue, is there a possibility that moab and/or
torque could do user management, such that when a job gets scheduled
the user (job owner) gets added to the requested node(s) (using
xgroupset and xuserset) dynamically, and then deleted if no other job
belonging to that user is running on the said node?  It would simplify
user-management tremendously.

Regards,
Daniel

On 10/30/08, Hugh Greenberg <[EMAIL PROTECTED]> wrote:
>
>  Actually, I this script was combined with another script, so the
>  directions are wrong.  I will fix the instructions.
>
>  -
>
> Hugh Greenberg
>  Los Alamos National Laboratory, CCS-1
>  Email: [EMAIL PROTECTED]
>  Phone: (505) 665-6471
>
>
>
>  On Thu, 2008-10-30 at 17:58 -0600, Hugh Greenberg wrote:
>  > Daniel,
>  >
>  > You are absolutely right.  I forgot to include that script.  I will be
>  > submitting it to the repository.  Thanks.
>  >
>
>

Reply via email to