Re: [Condor-users] logEvictEvent with unknown reason (108)

[Dan Bradley <dan@xxxxxxxxxxxx>]

David,

The example below shows that the startd is refusing to "activate" a 
resource claim.  Are most of the evictions that you see like that?  Look 
in StartLog on the worker node to see the specific reason for the refusal.

--Dan

David A. Kotz wrote:

> I'm getting a *lot* of evictions like the one below, where the reason is
> listed as unknown.  How might I diagnose the cause of the problem?
>
>
>
> 2/14 16:47:49 ******************************************************
> 2/14 16:47:49 ** condor_shadow (CONDOR_SHADOW) STARTING UP
> 2/14 16:47:49 ** /lusr/opt/condor-6.6.6/sbin/condor_shadow
> 2/14 16:47:49 ** $CondorVersion: 6.6.6 Jul 26 2004 $
> 2/14 16:47:49 ** $CondorPlatform: I386-LINUX_RH72 $
> 2/14 16:47:49 ** PID = 27611
> 2/14 16:47:49 ******************************************************
> 2/14 16:47:49 Using config file: /lusr/condor/etc/condor_config
> 2/14 16:47:49 Using local config
> files: /lusr/condor/etc/LINUX-INTEL/local/mast
> 2/14 16:47:49 DaemonCore: Command Socket at <128.x.x.x:58663>
> 2/14 16:47:50 Initializing a VANILLA shadow
> 2/14 16:47:50 (15.46) (27611): Request to run on <128.x.x.y:32772> was
> REFUSED
> 2/14 16:47:50 (15.46) (27611): Job 15.46 is being evicted
> 2/14 16:47:50 (15.46) (27611): logEvictEvent with unknown reason (108),
> aborting
> 2/14 16:47:50 (15.46) (27611): **** condor_shadow (condor_SHADOW)
> EXITING WITH STATUS 108
>
>