POLE1 conference

Thursday 1 Oct 2009, 14:00 → 15:30 Europe/Zurich

1/1-025 (CERN)

Meeting Access Information: * Evo Client Title: Pole1, Community: UNIVERSE Meeting password:pole1 From: 14:00 CEST To: 15:30 CEST EVO gate click "Start" and browse the right link in the list ("Pole1"). * Phone call Phone room ID: 37587 Phone password: 3376 EVO Phone Bridge Telephone Numbers: Slovakia (UPJS, Kosice) +421 55 234 2420 Switzerland (CERN, Geneva) +41 22 76 71400 Italy (INFN, several cities) Phone numbers Enter '4000' to access the EVO bridge Germany (DESY, Hamburg) +49 40 8998 1340 USA (BNL, Upton, NY) +1 631 344 6100 USA (Caltech, Pasadena, CA) +1 626 395 2112

14:00 → 14:05 Action list review

14:05 → 14:25 C-COD - ROD interactions

proposal: C-COD should send e-mail to ROC manager (and CC to ROD) asking for *explanations* in case ROD failed to act. We think that sending an e-mail to ROD again does not make much sense as ROD *already* failed to deliver their service.

14:25 → 14:45 Requirements for new NGI to join EGI from Operational Support perspective

What are the requirements for new NGI to join EGI from Operational Support perspective? A kind of operations SLA? - providing NOD (NGI Operator on Duty) service with specific rules on handling alarms and tickets - access to knowledge sharing DB or use of central GOC WIKI? - what else?

14:45 → 15:05 Quality metrics (NON-OK closed alarms)

proposal: Eliminate possibility to close alarms in NON-OK state completely. Pass this functionality to C-COD. Make a procedure for ROD to "ask for closing". NON-OK cases: a) SD solution: freeze alarm. Problem: SD can be defined in past. -> Protest it! b) moving node out of production let dashboard check if the node is in production c) monitoring system problems - fake alarms due to failure of mon. infr. possibly this will not last longer than 72h. d) core service problems - add alarm for a core service and mask, - but what if the CS is not in our NGI? Create ticket in GGUS and assign an alarm to this ticket.

15:05 → 15:15 AOB