FYI - anyone know what's happening at Sheffield?
-----Original Message-----
From: Jiri Chudoba [mailto:[log in to unmask]]
Sent: 03 February 2005 10:40
To: atlas-lcg (ATLAS-LCG integration task force); [log in to unmask]
Subject: se01.gridpp.shef.ac.uk
Hi,
I have 5 atlas jobs on skurut farm, all are blocked in the lcg-cp
operation.
These jobs are trying to get the same input file
gsiftp://se01.gridpp.shef.ac.uk/data/atlas/generated/2004-10-13/file8826
3e93-029c-4f6a-9f35-91871da3c2c3
from se01.gridpp.shef.ac.uk, which is probably down (not accessible by
ping or edg-gridftp-ls).
This SE was down already on Monday, when I reported the same problem
on golias farm. Is it just a temporary failure or was it removed from
LCG?
(Another SE in the same domain listed in GOC pages is
lcgce0.shef.ac.uk.)
There are many replicas of this input file, why the script chooses just
the one site which is down? Bad luck? Shall we deploy the timeout
mechanism
in lexor?
Obviously, such obsolete entries should be deleted from the replica
catalog, if
the SE is permanently down. Next week I can help with it if needed.
Cheers,
Jiri.
--
E-mail: [log in to unmask]
Phone: +420 266 052 145
Fax: +420 286 585 443
|