[torqueusers] Maximum number of nodes per queue
Denis.Charland at imi.cnrc-nrc.gc.ca
Tue Nov 30 07:32:53 MST 2004
"Resources_max.nodect" attribute sets the maximum number of nodes that can be requested
by a single job in a queue, not by all the jobs in a queue.
From: Stewart.Samuels at aventis.com [mailto:Stewart.Samuels at aventis.com]
Sent: Tuesday, November 30, 2004 8:38 AM
To: Charland, Denis
Subject: RE: [torqueusers] Maximum number of nodes per queue
You can use the qmgr queue command "resources_max.nodect" to set a maximum number of jobs of any kind in a queue. This will supersede the server resource_default.nodect resource.
From: torqueusers-bounces at supercluster.org
[mailto:torqueusers-bounces at supercluster.org]On Behalf Of Charland,
Sent: Monday, November 29, 2004 4:41 PM
To: torqueusers at supercluster.org
Subject: [torqueusers] Maximum number of nodes per queue
I'm running OpenPBS 2.3.16 on a 128 processor Beowulf cluster. I recently
created a seperate queue for guest users and I would like to limit the number of
nodes that can be allocated to all jobs running from this queue. To do so, I
used the "resources_available.resource_name=value" queue attribute. For
resource_name, I tried "nodes", "nodect", ncpus" and "neednodes" without
success. The man pages for pbs_queue_attributes (7B) defines this attribute as
being "The list of resource and amounts available to jobs running in this queue.
The sum of the resources of each type used by all jobs running from this queue
cannot exceed the total amount listed here".
Anybody has idea what can be the problem?
Note: Torque 1.1.0p4 behaves the same way.
Linux Cluster Administrator
National Research Council of Canada
torqueusers mailing list
torqueusers at supercluster.org
More information about the torqueusers