@CERN: David Front, Marco(LHCb), Dirk(3D), Eva(3D), Ioannis(POOL), Radovan(POOL), Giacomo, Flavia
General points:
Testbed interventions during last week
replica database upgraded to 10gR2, CPU January patches applied
Kernel upgrade on the OEM and Frontier Servers machine, CPU January patches applied
allocated one disk server for the 3D - it will be the downstream capture database
Dirk's asked for 3 caches servers
Frontier with Tomcat and Squid
CMS will be taken over from the support
No meeting next week. See you on Monday 6 February Database Readiness Workshop
Roundtable Tier1 sites status
BNL (Iris): sporadic node reboot problems solved, now the cluster for production is stable
CNAF (Barbara): RAC is not stable: upgrading OS and to OCFS2. Kernel panic often - still under investigation
RAL (Gordon): problems using OCFS2 and Red Hat 4, Oracle has been already contacted
Gordon will contact Barbara, similar configuration and similar problems
IN2P3 (update by Dirk): 2 cluster installation ready
Roundtable testbed test progress
Birger: tests re-done, they run fine. Birger will update the table. Time for stress tests requested
Birger will show some slides about the results during the readiness workshop (requested by Dirk)
Saima (update by Eva): replication of some operations which open a big number of cursors (up to 500) cause the the apply process to abort - number of open cursor excedeed error. To be clarified together with Saima
Gancho (online-offline ATLAS tests): got appropieate accounts, running first tests (tables creation and data population), preparing job to introduce big load on the server
Stefan: exercising the load of COOL database from different locations. Dirk suggest to start the replication tests asap (contact Gordon or Eva for help)
Jerome and Eva are in contact for testing AMI database replication
Marco: no more tests so far
Database Services and Grid User Support
presentation by Flavia Donno
slides available on the agenda
Summary
provides one interfarce for the user and the appropiate channels to forward the problems to specific experts
how to contact the user whose jobs are misconfigured? people behind support lists will look at the user causing the problem,...
applications are not really interface: the request is forwarded to the Generic Grid experts and, in case of the applications, they forwarded the request to the special unit of the application