GenericNQS batch system

[Lars Henriksen]

Dear beowulfers

I'm having some problems with the Generic NQS batch system.

Creating and using queues on a single host works fine,, but when i try
to submit jobs to queues on remote hosts, it does not work. Does anyone
have experience with that kind of operation?

Here is what i've done:

On the scheduling host (host1):

# qmgr create pipe sched-queue destination = exe-in at host2
# qmgr set lb_out sched-queue
# qmgr enable queue sched-queue

On the host that has to do the job execution (host2):

# qmgr create batch exe-queue pipeonly
# qmgr create pipe exe-in pipeonly destination exe-queue
# qmgr set lb_in exe-in
# qmgr enable queue exe-queue
# qmgr enable queue run-in
# qmgr set scheduler host1

In 'nmapmgr' on both host, entries has been added both for principal
names and aliases.

/etc/hosts.nqs looks like this on both hosts:
* *

So when i try to submit at job to the system on host1:
(top of job description file:)
-------
#QSUB-q sched-queue             
#QSUB-eo                        
#QSUB-r test                 
<snipped shell code to be executed>
-------

nothing happens :-(

edited syslog from the host where submission is made:

host1 NQS daemon[7467]: psc_spawn: Rqst not scheduled due to none there.
host1 NQS daemon[7467]: psc_spawn: Rqst not scheduled due to none there.
host1 NQS Pipeclient[5899]: Process logging started at Tue Nov 18
10:24:36 2003
host1 NQS Netdaemon[5900]: Netdaemon: Connection from host1
host1 NQS Pipeclient[5899]: Unable to deliver request 31 to a
destination
host1 NQS Pipeclient[5899]: Msg #2:Scheduling request for retry at a
later time
host1 NQS Pipeclient[5899]: Msg #2:Request rescheduled; exiting

A 'qstat -x' shows this:

<SNIP>
Destset = {exe-in at host2 [RETRY]
                 <Unreachable since: Wed Nov 12 12:35:10 CET 2003>
                 <Next retry at: Tue Nov 18 10:30:07 CET 2003>
             };
</SNIP>

I'm kinda baffled by this... 

Well thanks for your patience in reading this. I hope some of you can
give me some pointers...

best regards

Lars
-- 
Lars Henriksen                  | MESH-Technologies A/S
Systems Manager & Consultant    | Forskerparken 10
www.meshtechnologies.com        | DK-5230 Odense M, Denmark
lars at meshtechnologies.com       | mobile: +45 2291 2904
direct: +45 6315 7310	 	| fax:	  +45 6315 7314


_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf