Re: [Condor-users] jobs only executing on their node

[Ian Chesal <ichesal@xxxxxxxxxxxxxxxxxx>]

On Tuesday, 22 November, 2011 at 8:11 PM, Tom Melendez wrote:

Hi Folks,

Working on setting up my own small two-node test cluster and have

worked through some issues thanks to the list. Now I'm seeing that my

job is only executing on the machine in which it was submitted. I see

from condor_status that all of the slots from both machines are

present. But when I submit the job, despite how many I queue up, they

are only processed on the one node. I haven't seen anything in the

logs to even signal to me that the job is being considered on another

node.

Any ideas?

I'd say 9 times out of 10 it's the FileSystemDomain setting that's blocking jobs from spreading out across machines in a new pool like this.

But you can use condor_q to get a better idea of what machines match your job requirements. Just run:

condor_q -better-analyze <cluterid>.<procid>

Against a job that's idle in your queue. Post the results if they're unclear.

Regards,

- Ian

---

Ian Chesal

Cycle Computing, LLC

Leader in Open Compute Solutions for Clouds, Servers, and Desktops

Enterprise Condor Support and Management Tools

http://www.cyclecomputing.com

http://www.cyclecloud.com

http://twitter.com/cyclecomputing