Attending: Govind, John H, Matt, Winnie, Raul, RobC, Sam, Brian, Jens (chair+mins) 0. Operational blog posts Brian is still way ahead as our best blogger! And now we have the purdah; but hopefully people can see it is very unpolitical.... 1. Reminder: KeyDocs and other docs in the wiki. Currently there are still some very red storage key docs, and the storage wiki pages have not been updated in a looong time except for Bruno adding (what is currently) a stub on CEPH. 2. Apropos CEPH, we haven't had a technical talk in a while - suggestions for what and who? And where - could be a storage meeting, could be hepsysman Suggestion to have a talk on adding a new VO (see also links in chat). Setting up storage for a new VO is non-trivial, particularly if they are not clear about their data model. On which topic, we also need someone to represent CERN@School and otherwise fill the Whyntie-shaped gap in our portfolio. Not so much a storage thing, but Jens suggested giving a talk at hepsysman on multiple subject alternative names in host certificates (provided the support situation has improved a bit by then...) Also on the topic of technical talks, HEPiX is going on as we speak; ... https://indico.cern.ch/event/595396/ Sam points out the distinction between rapid onboarding of a new VO vs correct onboarding - takes time to do things properly. Is it worth trying to document the Wisdom of supporting a VOs data infrastructure - Sam thinks so, Jens thinks it might be better to have VOs join the meeting and discuss their requirements. However, the two need not be mutually exclusive. 3. Amongst what I call the loose ends, the more obtrusive ones are probably the caching testing progress (incl any plans for CVMFS) and T2 evolutions including support and new/old data transfer protocols, accounting (incl webdav deletions), and of course the wiki. CVMFS seems to be a reasonable approach at least for VOs with cacheable data, and we will do a new round of testing with LIGO which seems to have a well understood data model; apparently the planned work is at RAL only but it would be interesting to have a cache elsewhere like Glasgow. The xrootd redirection stuff unchanged since we last talked about it. ARC cache - Durham OK, need another site...? Raul has a set of new users bringing gifts (ie resources); discussions have only just started - they are doing some kind of bioinformatics - in particular whether they should use go the ARC path or xroot one. Dual stack: Imperial and QMUL have dual stack (ie IPvs 4 and 6) and Raul has used both and not seen a noticeable difference. 4. AOB Paige Winslowe Lacesso: (26/04/2017 10:11:20) Just checke again - right, they said they don't need storage. Well quick topic, DONE then. Jens: (10:11 AM) https://www.gridpp.ac.uk/wiki/New_VO_deployment (!) https://www.gridpp.ac.uk/wiki/Instruction_for_VO_administrators