Re: [Condor-users] Job evicted

[Jaime Frey <jfrey@xxxxxxxxxxx>]

On Sep 29, 2006, at 11:06 AM, Alan wrote:

> I have some jobs, supposed to not take more than 10 min, taking hours
> to carry out, and looking at their logs I read several lines like the
> one below:
>
> 001 (10945.000.000) 09/29 16:44:14 Job executing on host:  
> <172.24.89.68:46508>
> ...
> 004 (10945.000.000) 09/29 16:45:48 Job was evicted.
>         (0) Job was not checkpointed.
>                 Usr 0 00:00:00, Sys 0 00:00:00  -  Run Remote Usage
>                 Usr 0 00:00:00, Sys 0 00:00:00  -  Run Local Usage
>         2367357  -  Run Bytes Sent By Job
>         10400133  -  Run Bytes Received By Job
>
> [and so on]
>
> My job is always being evicted and resubmitted to another node, until
> it finally terminates.
>
> Is there a way to look deeper in such a behaviour? Is there something
> that I can do to avoid or at least minimise it?
>
> It gets weirder because I also submit several other jobs at the same
> time and almost all of them terminates in a couple of minutes, as
> expected.

The shadow log on your submit machine and the start and starter logs  
on the execute machines are the places to start looking for the cause  
of these evictions.

+--------------------------------+-----------------------------------+
|           Jaime Frey           | I used to be a heavy gambler.     |
|       jfrey@xxxxxxxxxxx        | But now I just make mental bets.  |
| http://www.cs.wisc.edu/~jfrey/ | That's how I lost my mind.        |
+--------------------------------+-----------------------------------+