GDB 11th July 2012
*** Welcome (Michel Jouvin)
Future topics (no comments from audience)
Computing mentioned as part of Higgs discovery, but WLCG never mentioned, why ? Branding ? Paper ? Some disagreement with that statement - apparently there's a book (but nobody know's about it ?!)
*** CVMFS Speaker(Ian Collier)
Deployment status: Atlas: 78 of 104 T2 sites deployed (all of the UK)
LHCb: 26 out of 86 - officially requesting sites now all needs conditions database, so far only at Tier 1 , T2s should install as well preferring sites with CVMFS
CMS now starts as well.
Next steps for CVMFS: NFS exported client (very problematic) - for sites with not enough diskspace, problem are stale handles - end of August ?
Question: Who apart form the experiments scan help to push the sites to install it? Alessandra Forti relies her experiences: follow up, simplified instructions
*** CE extensions (Davide Solomoni)
concentrating on multicore
Atlas worried about extra complexity it says it doesn't need. 'Pilot' problem exacerbated - if you reserve multiple cores and then no job arrives, it's worse than just a single core pilot job.
***post-EMI discussion (Michel Jouvin)
Jamie Shears disagree with everything.
*** topbdii (Maria Alandes Pradillo)
LCG_GFAL_INFOSYS should be configured with 3 bdiis.
Andreas: Where are these recommendation published ? (Attach to baseline services webpage ?)
*** glexec (Maarten Lithmath)
The gLexec deployment page has been revamped as of 10th July.
Sites still dragging their heels.
*** WG on storage fedarations (Fabrizio Furano)
This is still very much work in progress. So far xrootd is the natural way to go.
*** DPM: Towards a Community Model - Oliver Keeble
We need money, please.
Will be discussed in a NGI France meeting.
UK contribution ?
***RFC/SHA-2 Proxy Update (Maarten Lithmath)
Nothing decided yet, but we are all doomed.
*** Argus/Central banning Valery Tschopp
Some discussion on how to use central banning without Argus (possible in principle, work with OSG)
*** Maria Girone WLCG Operations Coordination Team
We solve problems through yet another meeting, apparently, but can't find volunteers.
Goal: "Computing as a service at the Tier2"
***Jobs with High Memory Profiles -- Stephen Gowdy (CMS), Maarten (Alice), ?? LHCb:
Apparently some sites have no swap space, why killing the job,
not the process (what else is a site to do ?)
Atlas: swap should be 2x RAM size
