> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of Emyr James
>
> I have seen some jobs fail, for example
> http://panda.cern.ch/server/pandamon/query?job=1945970655
>
> I haven't seen this failure reason before so am not sure what to do about
> it. Does anyone have any ideas ?
>
Sounds like this:
https://ggus.eu/ws/ticket_info.php?ticket=97230
I'm not sure how much RAM you've got in that node, but it's clearly
another one of these big Opteron boxes, so it makes sense.
AIUI, the short term fix is to have an ATLAS internal setting tweaked
in AGIS so that the pilot doesn't set its resource limit at a level that
Java then blunders straight into; that needs an ATLAS person to make the
change. I suspect Alessandra will pick it up from this, but if you want
to be sure, the best thing to do is probably to drop an email to:
[log in to unmask]
Ewan
|