Hello all,
I've been using maui and torque for a few years now very successfully
and happily. However, I've just installed torque (v2.1.6) and maui
(v3.2.6p18) on a small test cluster here and am having a little trouble
with submitting jobs.
As soon as I submit jobs (both mpi and serial jobs) to the cluster using
qsub maui crashes although the job is actually submitted to one of the
cluster nodes and runs to completion (as shown by qstat). The test job
that I am using is simply: 'echo "sleep 60" | qsub' as suggested in the
pbs quickstart manual.
Running showq before submitting a job works fine and shows the expected
output. However as soon as the job has been submitted showq results in
the following output:
$ showq
ERROR: cannot send request to server aerie:43559 (server may not be
running)
ERROR: cannot request service (status)
Maui cannot be restarted until the job has completed (no error message
is given, but the server doesn't start up). However, as soon as the job
has finished the server can be started and showq runs happily.
The maui log file does not show anything that seems to relate to an
error even when I've changed the LOGLEVEL to 7 within the maui.cfg file.
Is this a known problem that anyone else has seen? I'm stumped as to
where to go from here.
Thanks in advance for any help that you can give me
Rich
-----------------------------------------
Rich Bruin
Materials Grid project
Department of Earth Sciences
University of Cambridge
[EMAIL PROTECTED]
http://www.materialsgrid.org
_______________________________________________
mauiusers mailing list
[email protected]
http://www.supercluster.org/mailman/listinfo/mauiusers
