Minutes 8 December 05

• VRVS/phone: Iris(BNL), Julie(FNAL), Lee(FNAL), Barbara(CNAF), Sasha(ATLAS)
• @CERN: Luis(3D), Radovan(POOL), Birger(ARDA), Marco Clemencic(LHCb), Dirk(3D), Eva(3D)
• Apologies: Gordon(RAL), Saima(CMS), Stefan(ATLAS)

• General points:
  • The milestones presented last week in the meeting have been accepted
  • CHEP abstract has been accepted. Dirk will send around the list of contributers)
  • Readiness workshop: better no have it during CHEP, choose some place close to CERN

• Setup status:
  • There is a teable with the status of the available resources to be completed in the 3D wiki main page
  • BNL (Iris): 2 nodes, o.s. RHES3, storage: 500GB, setting up the cluster
  • Requested dba presentation on how to setup the cluster for next 3D meeting
  • CNAF (Barbara): RAC environment setup ongoing, results report during next 3D meeting
  • FNAL (Lee): temporary Squid cache setup at several T1/T2 sites for tests, GB ethernet is important for SQUID perfromance
  • Easy to have one Squid cache per experiment but this depends on the sites
  • more detail for the configuration when the tests will be finished (next week)
  • Lee will make a proposal according to sample results

• Testbed tests progress:
  • AMGA replication tests: Replication test finished, Birger showed looking at the timing which is still affected by large back-log as IN2P3 replica is still unavailable
  • LHCb replication tests (Marco): database was replicated, functionality is correct. Marco will send a mail to Gordon to get the connection string to RAL
  • ATLAS online tests: waiting to be connected to one of the replicas
  • Proposal for the new Streams configuration for Tier0 - to be discussed with Oracle experts (Eva)

• POOL / Frontier plugin tests:
  • CMS asked for POOL release at the Architecture Forum
  • POOL based on RAL against ROOT5 should be ready before Xmas.
  • Not able to finish Persistency release this year - new schedule for the first 2 weeks of January
  • Persistency release will address some of the requests from CMS/FroNtier
  • Lee agreed with the 2 releases split
  • Frontier doesn't use the concept of sessions - technical discussion to be followed offline
  • Application Server: CMS will investigate if DES J2EE services are suitable for FroNtier production next year