GDB February 2013
* Welcome (Michel Jouvin)
- No April GDB/register for March in Karlsruhe
He noted that it wasn't anymore difficult to get Karlsruhe than it was to get to CERN. Comment from the audience (I cannot identify all speakers) wrt pre-GDB: "We don't send technical personnel to meetings outside CERN."
From a T2 point of view: That attitude while not particularly surprising is not helpful.
Matt's tarball was mentioned (also in other talks). The view on this is on the verge of moving from "we are so grateful someone is doing this" to "this is not good enough". I don't think anyone is aware that it's not Matt's job, nor that we haven't signed any kind of MOU.
Apparently gLexec is still on the wishlist, though implementation is slow (I haven't had an actual request for it in months). Apparently now implemented in 107 CEs (that has to be a tiny fraction of all CEs around, not sure this number makes sense).
* future support of EGI services (Tiziana Ferrari)
Grand plans, no people or money. UMD (apparently it's not seen as a dead duck by the upper echelons which makes me wonder if they ever check what people actually use) got a lot of mention, Marcus Schulz got rather sarcastic when being confronted with the standard EGI advertising blurb.
* IPV6 testbed progress and planning update (Dave Kelsey)
Start with dual stack approach. Please don't ignore it until the last minute. Talk to Dave Kelsey wrt tests.
See also: [IPv6]
* Final status of upgrade to EMI MW and future plans (Peter Solagna)
Most sites have upgraded.
Can we stick the tarball in cvmfs (honestly, I don't think the distribution is the biggest issue with the tarball...)
From March 1st alarms for EMI1 will appear in the operations dashboard.
* Report from Operations Coordination/Discussion on SL6 migration issues and schedule (Andrea Sciaba)
Operations Coordination News:
More Tier 2 participation (UK has Alessandra) from different regions.
perfSONAR - can this be treated as production infrastructure (though no clear reason was given why and a lack of manpower was noted, but UK participation was noticed as good)
squid monitoring (register squid servers in GOCDB - not sure this is technically feasible yet)
SL6:
If a site runs SL5 and SL6, Atlas needs different CEs for Sl5 and SL6 WNs as not all their releases work on both.
T1s are asked not to moved until the final validation, others can (no push from experiment to upgrade)
* AAI (Authentication and Authorization) on Storage Systems (pre-GDB summary/Maarten Litmaath)
Summary by Maarten
No immediate action by anyone required.
* Summary of the discussions in the cloud e-group during last month (Michel Jouvin)
There's an eGroup (which has to be the most inconvenient way to discuss things - ever). Of course stating the name of the eGroup in the talk would have been far to easy, but project-lcg-gdb-clouds-wg is a good bet (archive).
Topics so far:
Image contextualization
VM instantiation and scheduling
VM duration
Agreement that contextualisation is needed (startup script?).
Accounting
F2F/Vidyo meeting in Karlsruhe
* Storage Federation WG Final Report (Fabrizio Furano)
The notes in the minutes are more detailed than anything I could provide, so please check those:
[Storage Federations]
From a T2 point of view: Some lively discussion over arguments that clearly everyone had heard before, probably multiple times.
Conclusion: xroot/webdav are our newest tools/toys/reason to justify our existence and they will be implemented to solve problems we might or might not have(*) and then we see how it goes.
(*) Fallover for unavailable files/diskless Tier3.
