GPFS issue / last maintenance. Changes performed: SM changed to normal host, firmware upgrades on all Sun cards, general kernel/ofed upgrades, IB multicast flood fixed, two bad cables replaced.
Thor issue (soft lockup-cpu stuck).
Compute nodes ordered (6x4).
CVMFS deployed
PSI (reports Fabio):
We're planning to migrate to Chimera before end of July. I'm preparing an SL6 64bit PostgreSQL 8.4 VM about that.
For SGE Admins: queue configuration with nodes of unequal compute performances.
UNIBE (reports Gianfranco):
Xxx
UNIGE (reports Szymon):
Quiet operations, SG has two other projects now. Occasional problems:
job stout and stderr not delivered by torque, found in /var/spool/torque/undelivered/
one 3 TB disk dead out of 10*12=120, in six months since we have IBM machines in prod
Upgrade news and plans:
ATLAS skeptical on the HEP-SPEC 2006. Me too.
offers from: IBM, DELL, Fujitsu: IBM with AMD comes first, only worry is 64 cores/machine DELL with Intel comes second, hame HEP SPEC but fewer cores