Re: [Condor-users] starter failed to connect to collector

[Jaime Frey <jfrey@xxxxxxxxxxx>]

On Oct 14, 2005, at 1:44 AM, DeVoil, Peter wrote:

> I have a few machines in a 30 node winXP pool that refuse to start  
> jobs.
> I see these in the starter log:
>
> ...
> 10/14 14:59:38 vm2: Received match <192.168.0.162:1353>#4441711918
> 10/14 14:59:38 vm2: State change: match notification protocol  
> successful
> 10/14 14:59:38 vm2: Changing state: Unclaimed -> Matched
> 10/14 15:01:38 vm2: State change: match timed out
> 10/14 15:01:38 vm2: Changing state: Matched -> Owner
> 10/14 15:01:38 vm2: State change: IS_OWNER is false
> 10/14 15:01:38 vm2: Changing state: Owner -> Unclaimed
> ...
> 10/14 15:04:50 DaemonCore: Command received via TCP from host
> <192.168.0.98:3484>
> 10/14 15:04:50 DaemonCore: received command 442 (REQUEST_CLAIM),  
> calling
> handler (command_request_claim)
> 10/14 15:04:50 Error: can't find resource with capability
> (<192.168.0.162:1353>#4441711918)
> ....

It appears the schedd that was matched to this startd took over 5  
minutes to connect to it to start the job. We'd have to look at the  
schedd log to see why it took so long.

+----------------------------------+---------------------------------+
|            Jaime Frey            |  Public Split on Whether        |
|        jfrey@xxxxxxxxxxx         |  Bush Is a Divider              |
|  http://www.cs.wisc.edu/~jfrey/  |         -- CNN Scrolling Banner |
+----------------------------------+---------------------------------+