On Mon, 21 Apr 2008 13:41:08 +0200
Maarten Litmaath wrote:
[...]
> > There must have been a good reason for that...
>
> Indeed, David Smith put it in because jobs submitted by lcgpbs should
> never end up in the 'W' state, unless there is a configuration error,
> typically with stage-in, typically when a WN cannot "scp" from the CE.
Mmm.... maybe I'm losing something, but I have a missconfgiured node
which can't scp from/to ce and it has executed many "failed" jobs.
Could you please be more specific on this point?
Cheers,
Arnau
|