[torqueusers] job schedule and run problem!

Georgios Zikos georgios.zikos at gmail.com
Thu Sep 20 13:46:27 MDT 2007


Hi,

it seems that the problem arises either from the firewall configuration or
the hostname resolution of your server.
Make sure that you have the following ports open in your firewall because
from my experience oscar does not
does not change them:

pbs                  15001/tcp
pbs_server      15001/udp
pbs_mom        15002/tcp
pbs_resmom   15003/tcp
pbs_resmom   15003/udp
pbs_sched      15004/tcp  # not used by oscar

if that does not fix your problem make sure that you can access all the
nodes and the oscar nodes with their
hostnames and without ssh asking for password. The easiest way is to run the
TEST CLUSTER (step 8) from
the installation suite and make sure that SSH ping test, SSH server -> node
, SSH node -> server are passed.

If not, you probably have to change the /etc/hosts on the server and/or on
the nodes.

Finally maui does not use port 15004, it is used from  the pbs scheduler
which is inactive in oscar 5.0

I hope this helps.

Regards,
George
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20070920/759e10b8/attachment.html


More information about the torqueusers mailing list