[Beowulf] Running on headnode only.

David Havery dhavery at pssclabs.com
Fri Apr 15 15:55:35 PDT 2005


Good day and thank you all very much for your time,

I have a single node cluster, that is, a headnode and a slave node.  I'd 
like to be able to use qsub to queue a job to use only the headnode.  
I've attached my queue information at the bottom of this mail as well as 
the job script.

The changes that I've made so far are:
create node master np=2,ntype=time-shared,properties="headnode"
changed to:
create node master np=2,ntype=cluster,properties="headnode"

and in the jobscript changed:
#PBS -l nodes=1:ppn=1
to
#PBS -l nodes=1:ppn=1:headnode

Now when I run the attached job script, the error I get is:
qsub: Job exceeds queue resource limits

This is my job script (much of it is unimportant but here's the whole 
thing):
----------------------------------
#!/bin/sh
### Number of nodes 1 nodes using 1 Processor Per Node
#PBS -l nodes=1:ppn=1:headnode
### Example: to request 2 VPs on each of 3 nodes and 1 VPs on 2 more nodes
## #PBS -l nodes=3:ppn=2+2:ppn=1
### Output files. If not specified PBS uses the job name and ID.
###PBS -o output-mytorquejob.log
###PBS -e output-mytorquejob.err
### Merge stderr with stdout
#PBS -j oe
### Mail to user
#PBS -m eb
### Queue name
#PBS -q default
### Job name
#PBS -N PBS_AOUT_Test
### Declare job-non-rerunable
#PBS -r n
# This job's working directory
echo Job ID: $PBS_JOBID
echo Working directory is $PBS_O_WORKDIR
cd $PBS_O_WORKDIR
echo Running on host `hostname`
echo Time is `date`
echo Directory is `pwd`
echo This job runs on the following processors:
echo `cat $PBS_NODEFILE`

#Define number of processors
NPROCS=`wc -l < $PBS_NODEFILE`
echo This job has allocated $NPROCS nodes/processors.

#
for AOUT in a02knp-61MB.x86_64.fc2 ; do
#for AOUT in a06knp-549MB.x86_64.fc2 ; do
#for AOUT in a15knp-3.4GB.x86_64.fc2 a16knp-3.9GB.x86_64.fc2 ; do
        EXEC_CMD="time /home/mpiadmin/bin/${AOUT}"
        echo -e "\n$EXEC_CMD\n"
        time $EXEC_CMD &
done

wait

# Sleep for 2 minutes to give us plenty of time to check out the operation
# of the queuing system.
#sleep 120
------------------------------------------


Here is the queue information
--------------------------------------------
create queue default
set queue default queue_type = Execution
set queue default Priority = 50
set queue default max_running = 4
set queue default max_user_run = 4
set queue default resources_min.cput = 48:00:00
set queue default resources_max.cput = 49:00:00
set queue default resources_default.cput = 48:00:00
set queue default enabled = True
set queue default started = True
#
# Set server attributes.
#
set server scheduling = True
set server node_pack = True
set server max_running = 4
set server max_user_run = 4
set server acl_host_enable = True
set server acl_hosts = master.host.com
set server acl_hosts += *.host.com
set server acl_hosts += localhost.localdomain
set server managers = mpiadmin at master.host.com
set server managers += mpiadmin at localhost.localdomain
set server managers += root at master.host.com
set server managers += root at localhost.localdomain
set server default_queue = default
set server log_events = 127
set server mail_from = pbsadmin
set server query_other_jobs = True
set server resources_default.neednodes = 1
set server resources_default.nodect = 1
set server resources_default.nodes = 1
set server scheduler_iteration = 150
set server node_ping_rate = 150
set server node_check_rate = 300
set server comment = 'Torque Server @ master.host.com'
#
# Define cluster nodes
#
create node master np=2,ntype=cluster,properties="headnode"
create node n01 np=2,ntype=cluster,properties="odd"

print server
--------------------------------------------


-------------------------
David Havery
dhavery at psslabs.com
http://www.pssclabs.com
(949) 380-7288
-------------------------





More information about the Beowulf mailing list