[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [condor-users] Myrinet - similar problem



Hi,

I don't have a solution but I think I have the same problem.

I have recently installed Condor on a Beowulf cluster to do workload
management on the nodes. The central manager is on the front end. If the
nodes are all busy but there are more jobs in the queue I want them to flock
to another Condor pool. The problem is that the nodes on the beowulf are in
a different IP address space than the rest of the computers here. The
central manager on the front end is configured to use the NIC on the network
of the nodes of the cluster. But in order to flock to a different pool
outside of the beowulf it would have to use a different NIC.

So I am also looking for a way to make the central manager of Condor
communicate over two NICs. Does anyone know how to configure it like that?

Thanks in advance,
Michael.



----- Original Message -----
From: "Joel Hernandez" <joelh@xxxxxxxxx>
To: <condor-users@xxxxxxxxxxx>
Sent: Thursday, March 04, 2004 11:25 PM
Subject: [condor-users] Myrinet


> We have two eight node dual cpu clusters (louie and duey).  Users submit
> their jobs on louie and when all the nodes are busy, they start to flock
> and run on duey.  This was working fine until we recently installed
> Myrinet on louie but not on duey.  In order to get Condor to run using
> IP over Myrinet, I modified the condor_config.local files on each of the
> louie nodes to reflect the Myrinet IP address for the
> NETWORK_INTERFACE.  However, since we do not have Myrinet on the duey
> cluster, the head node on duey and the head node on louie must still use
> the 100 Mbps ethernet address to comunicate.  This would require two
> values for the NETWORK_INTERFACE on the head node on louie.  It appears
> that Condor does not support such a thing.  Is there another way to get
> the head node on a cluster to communicate via two network interfaces?
>
> Thanks,
> Joel
> -----------------------------------------------------------------
> Joel Hernandez
> Systems Programmer / Analyst
> MCNC-RDI Center for Networked Information Discovery and Retrieval
> joelh@xxxxxxxxx
> http://www.cnidr.org
>
>
>
>
> Condor Support Information:
> http://www.cs.wisc.edu/condor/condor-support/
> To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
> unsubscribe condor-users <your_email_address>
>
>

Condor Support Information:
http://www.cs.wisc.edu/condor/condor-support/
To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
unsubscribe condor-users <your_email_address>