[torqueusers] Routing queue issues

Timothy E Miller millerte at wfu.edu
Thu Mar 3 14:29:50 MST 2005


Hey all,

I'm running Torque 1.0.1p5 (yes, admittedly old) and am having the
wierdest problem with trying to establish a routing queue.

I have two perfectly functioning queues that do exactly what I want
them to do named:  acl and rhel3.  The definitions are shown at the
end as is the job script.  (The magic is in the Maui config file of
course).

When I try to create a simple routing queue to automatically route
jobs to the appropriate queue, any job that is *not* "nodes=1" or
"nodes=2" fails with the classic "job exceeds queue resource
limits".  Any job that has a qualifier on the nodes (ppn, node
property, etc.) or nodes=3 (or greater) fails.  It doesn't matter
whether it is specified in job script or on command line.

Anyone have any ideas?

Thanks,
-Tim

create queue acl
set queue acl queue_type = Execution
set queue acl max_running = 244
set queue acl acl_group_enable = True
set queue acl acl_groups = A
set queue acl acl_groups += B
set queue acl acl_groups += C
set queue acl acl_groups += D
set queue acl enabled = True
set queue acl started = True
create queue rhel3
set queue rhel3 queue_type = Execution
set queue rhel3 max_running = 32
set queue rhel3 acl_host_enable = True
set queue rhel3 acl_hosts = deac126
set queue rhel3 acl_hosts += deac016
set queue rhel3 acl_hosts += deac125
set queue rhel3 acl_hosts += deac015
set queue rhel3 acl_hosts += deac124
set queue rhel3 acl_hosts += deac014
set queue rhel3 acl_hosts += deac123
set queue rhel3 acl_hosts += deac013
set queue rhel3 acl_hosts += deach03
set queue rhel3 acl_hosts += deac012
set queue rhel3 acl_hosts += deac011
set queue rhel3 acl_hosts += deac010
set queue rhel3 acl_group_enable = True
set queue acl acl_groups = A
set queue acl acl_groups += B
set queue acl acl_groups += C
set queue acl acl_groups += D
set queue rhel3 enabled = True
set queue rhel3 started = True
create queue new
set queue new queue_type = Route
set queue new route_destinations = rhel3
set queue new route_destinations += acl
set queue new enabled = True
set queue new started = True

set server scheduling = True
set server acl_host_enable = True
set server acl_hosts = (all my compute and head nodes)
set server managers = (all my secret identities)
set server default_queue = acl
set server log_events = 511
set server mail_from = adm
set server query_other_jobs = True
set server resources_default.cput = 01:00:00
set server resources_default.mem = 1024mb
set server resources_default.nodes = 1
set server resources_default.pmem = 64mb
set server resources_default.walltime = 01:15:00
set server resources_max.mem = 244000mb
set server resources_max.ncpus = 2
set server resources_max.nodes = 244
set server resources_max.pmem = 2200mb
set server scheduler_iteration = 600
set server node_ping_rate = 60
set server node_check_rate = 300
set server node_pack = True

#!/bin/sh
#PBS -q new
#PBS -j oe
#PBS -l cput=00:01:00
#PBS -l walltime=00:01:00
#PBS -l nodes=4
#PBS -W group_list=C

echo "Test job 1"
echo -n "Hostname is "
hostname

echo "PWD is $PWD"

echo "Environment is..."
set

echo "Job information is..."
qstat -f $PBS_JOBID


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Timothy E. Miller, PhD, RHCE                   voice:  (336)758-3257
Parallel Computing Systems Administrator         fax:  (336)758-7127
Wake Forest University                          cell:  (336)782-6987
Computer Science, Information Systems, Physics
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


More information about the torqueusers mailing list