Re: [Condor-users] Problem with wheckpointing

[Jaime Frey <jfrey@xxxxxxxxxxx>]

On Oct 17, 2005, at 9:56 AM, Nicolas GUIOT wrote:

> I have a problem with my checkpoint server : Jobs don't run (well,  
> they do, until they want to checkpoint) : I have this error message  
> on the checkpoint server :
>
>
> 10/17 14:44:01 ******************************************************
> 10/17 14:44:01 ** condor_ckpt_server (CONDOR_CKPT_SERVER) STARTING UP
> 10/17 14:44:01 ** $CondorVersion: 6.7.10 Aug  3 2005 $
> 10/17 14:44:01 ** $CondorPlatform: I386-LINUX_RH9 $
> 10/17 14:44:01 ** PID = 1561
> 10/17 14:44:01 ******************************************************
> 10/17 14:44:01 CKPT_SERVER running in directory /checkpoint
> 10/17 14:44:08  Receiving store request from XXX.XXX.XXX.24
>                         Using descriptor 7 to handle request
> 10/17 14:44:08 ERROR: cannot make directory './XXX.XXX.XXX.24'
>
> I started condor with the "root" user, and submitted the jobs from  
> the XXX.XXX.XXX.24
>
> What did I miss ?

Try making /checkpoint owned and writable by user condor.

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