Attending: Brian, Ewan, Gareth, Jens (chair+mins), John B, John H, Matt D, Winnie, Raja, Sam, David, Raul, Elena 0. Operational blog posts - two weeks into the quarter and we have zero blog posts! Still no blog posts... Operational issues? Raja is currently happy, at least with the UK. A Russian chap had got Cyrillic or something into the VOMS extensions which caused some unhappiness at least on SL5 (maybe it used UTF-8 and UTF-8 is better supported on SL6?) FTS backlog from Glasgow? Oddly slow transfers from Glasgow? Backlog has (been) cleared. Still looking into packet loss; Brian will run the meeting after the meeting and explain how where the packet loss is thought to happen. Also the GFAL issue is thought to have been resolved in the latest version. New UMD is expected to have new GFAL. (Brian is to sketch the relevant parts of the T1 networks for us - DONE) (Brian - was the GFAL issue resolved in the later versions? - DONE) 1. Quick updates on VOs with open issues - such as 1.1. UKQCD and LIGO? Did Paul respond? (Sam was going to contact Paul about "New DIRAC", Jens had a note about getting feedback from them - ONGOING) 1.2. DiRAC - update, current stats The current main issue with DiRAC is the low volume of data transferred up till now, only slightly more than 80TB. This needs resolving so it can speed up a bit. 1.3. CMS - response to feedback on syncat proposal? None - Jens may need to prod someone. 1.4. ATLAS - any storage or data model related things from ATLAS software week (hereinafter ASW)? "Forward progress on future storage" - it is thought that some T2s will be running more T1 style workloads, at least T2Ds which will have primary data. Leave less resource for user analysis? But there will still be room at T2s for user analysis. The current number of T1 related jobs at T2s is between 1/30 and 1/2 compared to T1s, so >0. Alastair Dewhurst (ATLAS T1) will join next week to give an update on CEPH at RAL, but would presumably also be able to answer questions on ASW. Also Matthew Viljoen (who recently gave a talk on DPHEP) attended ASW and could be persuaded to give an update (but note he is leaving STFC soon to work for EGI). 2. Updates on Ewan's (and Sam's) proposal. No other feedback since last week, so not progressing further, but would appreciate more comments and contributions. Costing numbers - generic, drives - reports that archival drives took 57 hrs to rebuild, so not good in RAID. Could be OK in object stores (but then that is what we propose to investigate) Contributions - such as costing, risks to list, or to Ewan if it's secret(ish). There is a PMB on the 27th which could usefully be used as a deadline. 3. Stress testing Bristol - Luke? We didn't send the RAL code around but it's like a script which writes (or reads) a file which is a one line, more or less. 4. Next week's meeting Jens will take two weeks leave; last time was just short of eight years ago. The meeting will be in good hands with Brian and Sam. 5. AOB Brian gave a talkthrough of the network diagram after the meeting. Matt Doidge: (15/07/2015 10:06:28) I'm being somewhat reminded by Star Trek IV... *of Star Trek IV Ewan Mac Mahon: (10:07 AM) It's a really odd thing to happen though. Not so much that the server side wouldn't be UTF8 clean, but that the client side tools wouldn't already be written to avoid doing anything funny. :-) Raja Nandakumar: (10:29 AM) Have a good holiday - Jens! And my apologies I need to go ... Paige Winslowe Lacesso: (10:34 AM) Have a restful ... rest, Jens!