John, Mark, Stephen, Matt, Gareth, Pete, Brian, Wahid, Chris, Ewan 1. Accounting records - what happened to them? The UR-WG are preparing something and projects seem to be using them. May pick this 2. Advice to small VOs - what happened to it? Related to that, do we have the KeyDocs we think we need? Perhaps that's a question for the new guys...? Is there a small VO we can play with? If there is a single range of VO data models, then we can help them. Snoplus? Jens will get in touch. Should we invite them to the storage? We did this with T2K, and it was useful - there is overlap between snoplus and T2K. Liverpool, QMUL, Oxford, RAL. 3. Next milestones...? What shall we be focusing on over the coming quarter? * Snoplus - updated documentation * xrootd/ATLAS (maybe CMS) - federation test sites. It is using certificates. * 4. Storage research projects - any interest? Lots of projects develop innovative stuff but not production ready - however I have this idea that we can sometimes use some of this stuff to improve services. Are we not using new stuff, innovation, enough? Should we invite them to tell us about their stuff? Should they invite us to talk about the real world? Experiments must worry about preservations - we "just" provide bitlevel storage. We could offer the VOs more, but they only wanted the bitlevel storage. Cloudy stuff - could also be a way to make life easier for the small VOs. Checksumming data, syncing across site ("federation"). Also a language issue - we tend to use words differently. Tier model, analysis model (job to data). EPSRC-project at QMUL: federated access model. Durability - what can the clouds provide? Is the trust problem with commercial providers or with "outside our organisation" Maybe worth following up on some of the preservation stuff, and the cloud stuff. GridPP - should speak to cluster projects in the UK who seem to be confused about things we have already solved. E.g. David Wallom? There is a SIG mailing list and meetings every six months. "All the grid stuff is hot air" Maybe plugging into HPC directly. Neasan could make contacts. 5. AOB [09:59:05] John Hill joined [10:00:06] Mark Norman joined [10:00:15] Stephen Jones joined [10:00:28] Matthew Doidge joined [10:01:49] Gareth Roy joined [10:03:26] Pete Gronbech joined [10:05:39] Brian Davies joined [10:05:45] Wahid Bhimji joined [10:05:49] Christopher Walker joined [10:06:00] Wahid Bhimji hi - sorry I'm late had to find some headphones [10:06:25] Govind Songara joined [10:06:55] Ewan Mac Mahon joined [10:06:58] Brian Davies sorry I am late - Fire alaram ( test not actual) [10:07:12] Christopher Walker snoplus.snolab.ca? [10:07:15] Jens Jensen https://www.gridpp.ac.uk/php/KeyDocs.php [10:07:42] Stephen Jones Matt Mottram is the guy [10:08:38] Stephen Jones We do [10:08:54] Stephen Jones Chris has a page with that data on it [10:09:17] Christopher Walker http://pprc.qmul.ac.uk/~walker/votable.html [10:09:41] Christopher Walker RAL, QMUL, Liv, Oxford [10:13:45] Ewan Mac Mahon Is this really 'test' or is it more 'implement'? [10:15:03] Ewan Mac Mahon Is there anything we want to consider related to the Tier 1 EOFS/Lustre/Whatever project? [10:15:14] Ewan Mac Mahon Can we claim that for the storage group? [10:15:14] Wahid Bhimji (though with Brian last week I remember we did say that we would track the xrootd stuff in the atlas thu meeting so maybe having it as a milestone here is wrong. [10:15:46] Wahid Bhimji Yes - well on the T1 Disk solution , I was surprised to see from Alastair in Atlas s/w week [10:15:51] Wahid Bhimji some criteria [10:16:08] Wahid Bhimji and options . I don't think they have been shared with us [10:16:34] Stephen Jones Re keydocs: these are the section in https://www.gridpp.ac.uk/wiki/Main_Page : [10:16:52] Stephen Jones 1) Getting up and running on the grid - new VOs [10:17:03] Stephen Jones 2) Getting up and running on the grid - users [10:17:38] Wahid Bhimji There is no point telling us about the preservation - they have to tell the experiments [10:20:20] Ewan Mac Mahon The trouble with the LHC VOs is that they all want more than dumb services, but they can't agree on all wanting the same thing above dumb storage. [10:24:19] Ewan Mac Mahon I think there are different ones depending on how much you pay them. [10:27:31] Christopher Walker http://www.janetbrokerage.ac.uk/ [10:28:02] Wahid Bhimji Sorry , I have to go in a min for mtg with DPM guys. I can tell you next week about the discussion on storage accounting at GDB [10:28:35] Ewan Mac Mahon Amazon (at least) do take some account of this sort of thing with their multiple regions setup, so as far as hardware failures go you can certainly have mulitple copies across separate places. [10:28:55] Ewan Mac Mahon That doesn't protect you from Amazon as a whole going boom, of course. [10:30:14] Wahid Bhimji left [10:30:50] Brian Davies ah , but i invented the two hemispsherical wheel, and have the patent.... [10:31:51] Stephen Jones What is DRI Money? Can I get some? [10:32:43] Ewan Mac Mahon There is no point talking to Dave Wallom [10:32:48] Ewan Mac Mahon Ever. [10:36:11] Ewan Mac Mahon There's a nice poster tag line there: GridPP - we did this already. [10:37:29] Ewan Mac Mahon Or we could have a special meeting.