Minutes 8 February 07

• Phone:
  • TRIUMF: Jim Li
  • GridKA: Doris
  • FNAL: Eric
  • ATLAS: Shasha
  • BNL: Iris, Carlos
  • NDGF: Olli
  • CNAF: Barbara

• CERN:
  • 3D: Dirk, Luis, Eva
  • ATLAS: Gancho, Florbela
  • LHCb: Marco

• General announcements:
  • GDB meeting:
    • presentations by Maria (technical side) and Sasha (experiment side)
    • backup policy has been approved
    • requests from ATLAS: 3rd node for the RAC installations - discussion about the real need, the sites should be prepared and check what is needed to meet the requirements
  • PIC has hired a new dba - Luis Diaz, he will come to CERN for training
  • Workshop @SARA
    • preparation for the dba meeting ongoing
    • agenda still to be completed
    • Dirk will propose some topics
    • dbas should prepare the issues to be discussed
    • Jamie Shiers will join the workshop to talk about the LCG milestones and experience

• Sites status
  • OEM installation to be completed
  • The sites must contact their ROC
  • BNL (Iris):
    • confirmed that the network problems are caused by the firewall
    • the oracle cluster will be moved outside the firewall
    • oem agents are down due to this work
    • ROC already contacted
    • metalink account: done
  • CNAF (Barbara):
    • problems with the new storage: there will be an intervention by the storage experts next monday
    • in the meantime, LHCb and ATLAS databases are available using the old storage
    • ROC already contacted
    • metalink account: actually using previous CNAF account. They should get new account for the SR related to the 3d project
    • patches will be applied when the storage problem is fixed
  • GridKA (Doris):
    • question about backups: how to estimate the space?
    • patches have been applied successfuly
    • oem agents up and running
    • Doris does not know whether ROC was contacted or not, she will check
    • metalink account: done
  • NDGF (Olli):
    • db using single instance up and running
    • patches have been applied
    • it is necessary to determine if the single instance database will be part of the tests - check if streams + oem can maintain the throughput
    • ROC already contacted
    • metalink account: done
  • RAL: still problems with the email list. One instance of LHCb cluster is down (problems during the patches apply)
  • SARA: pathes have been applied, oem agents up and running
  • Triumf (Jim Li):
    • iperf tests show very low network throughput
    • Jim Li should check if there is a firewall between CERN and Triumf
    • more tests will be run ( suggested by CERN network experts)
    • oem agent up and running
    • ROC to be contacted

• Experiments status:
  • ATLAS:
    • downstream configuration is ready
    • first tests with RAL as destination were successful (Eva)
    • GriKA and SARA joined: test during 27 hours was ok
    • IN2P3 and CNAF will be joined soon
    • IN2P3 will also be joined to the direct setup for Richard's tests
    • Michigan setup is ready - we should get a plan about their tests
    • move the Streams configuration to production to be discussed
  • LHCb:
    • CNAF to be added: data can be discarded
    • downstream capture configuration will be done tomorrow
    • discussion about the LFC: inside the LHCb cluster or will be a separate instance
    • LFC sw is being upgraded to allow bulk operations
    • schema changes will be done at the master database
    • new replicas will be added after the schema changes
    • LHCb will run COOL 2.0 for production

• CMS Frontier tests:
  • working with the online people to run the tests
  • 2 squid nodes have been requested at Tier0 (FIO)
  • as Luis is leaving, it is necessary to agree between CMS and FIO how to proceed with notifications, reboots, ...