Hello Pascual and list
Pascual, did you configure your Torque/PBS server?
Your nodes file,
/opt/torque/server_priv/nodes (or equivalent path)
should have the nodes you want to use, and how many CPUs (or cores) they
have.
Something more or less like this:
my_master_node.local np=2 #(only if you want the master
node to run jobs also)
my_compute_node_01.local np=2
my_compute_node_02.local np=2
...
You also need to restart the Torque/PBS server after you change the
nodes file.
See:
http://www.clusterresources.com/torquedocs21/1.2basicconfig.shtml
http://www.clusterresources.com/torquedocs21/1.3advconfig.shtml
http://www.clusterresources.com/torquedocs21/1.1installation.shtml
http://www.clusterresources.com/torquedocs21/
Gus Correa
--
---------------------------------------------------------------------
Gustavo J. Ponce Correa, PhD - Email: [EMAIL PROTECTED]
Lamont-Doherty Earth Observatory - Columbia University
P.O. Box 1000 [61 Route 9W] - Palisades, NY, 10964-8000 - USA
---------------------------------------------------------------------
Pascual Muñoz Muñoz wrote:
Hi all,
We are assembling a cluster following:
http://debianclusters.cs.uni.edu/
Right now we have a master/head node and a cluster/compute node. Both
are listed in the configuration files (torque pbs nodes and maui
machines.LINUX) to have 2 processors.
When we send jobs to the queue through a script like:
#!/bin/bash
#PBS -l nodes=2
cd /home/justme/
mpirun -machinefile $PBS_NODEFILE -np 4 /usr/bin/whatever
The 4 processes run on the master node! And the cluster node is free!
On other clusters, this works fine, so the problem is not in the
program being run, rather in the torque/maui (we guess) configuration.
So we did something wrong in configuration, but can't guess what. All
your ideas will be welcomed.
Thanks in advance.
_______________________________________________
mauiusers mailing list
[email protected]
http://www.supercluster.org/mailman/listinfo/mauiusers
