Swiss WLCG Operations Meeting on 2011-05-19

• Date and time: Thursday at 9:30
• Place: EVO, password: chipp
• External link / EVO: http://evo.caltech.edu/evoNext/koala.jnlp?meeting=M9MIMi2a2tD2Dl999DDD99

Agenda

Status

• CSCS:
  • Implemented punish system for find/du/ls. Job gets 30 seconds sleep after more than 10 bad ops / minute.
• PSI (reports Fabio):
  • PSI hosted the HPC Forum on Scheduling and Queue System, in that context we presented our Sun Grid Engine Accounting and Reporting Console.
• UNIBE (reports Gianfranco):
  • Experiencing random spontaneous reboots of X8440 blades. One module refuses to power back up. Network upgrade still pending (tentatively next Wednesday)
• UNIGE (reports Szymon):
  • In general rather quiet (I can do other things)
  • Occasional slow response of NFS, under investigation since early march:
    • It is not as simple as overloaded file servers.
    • Slow response happens on one client at a time, not on all the same time.
    • Have a look at slides attached: UniGE.NFS-monitoring-results.pdf.
    • It would be good to understand what is the limit. Thanks in advance for any ideas of things to check etc.

• Change of CSCS Maintenance day: "First working Wednesday of the month". Next one: Wednesday 8th of June.
• Agreed to talk about technical things (like Szymon's) on the lcg at cscs list

Attendants

• CSCS: Pablo
• CMS: Derek, Fabio
• ATLAS: Szymon, Marc, Gianfranco
• LHCb:
• EGI: Andres

Action items

• Item1

• FabioMartinelli-SGE_ARCo.pdf: SGE Accounting and Reporting Console installation at PSI CMS T3

• Not Trivial Example of SGE Accounting and Reporting Console usage: