Attending: Brian, Jens (C+M), John H, Jon, Matt, Gareth, Sam, Jeremy, Chris W,
Wahid, Steve, Raul, Ewan, Tom


0. Operational blog posts

   None at all is precisely how many blog posts we have for October.

   Snoplus related ticket to RAL, GFAL related: see chat log.  DPM is fairly
   "relaxed" about its TURLs, but CASTOR is more strict.  Nevertheless, one
   would expect the TURL to be alive after a srmPrepareToGet until either it
   times out or the client calls getDone.

   Sam wonders how it'd behave if redirects are turned on in DPM?  Default is
   off, partly because it needs to be 1.8.8 or later on everything, but the
   redirect is then generally considered a Good Thing(tm).

   We used to have automated interoperation tests, a set run by LBNL and one
   run by CERN (or was it us?) but that was mostly in the early days of
   interoperation, but even if they're not running we might still have the
   tests.

   It could be useful to involve the GFAL developers.

1. The long awaited chat with T2K.  T2K - should they use xroot?  Use of
FTS?  Any other questions?

   Special guest star: Jon Perkin.

   Currently about 8% of expected data acquired.  Got world's first electron
   neutrino event, but still need more data.  Looking for CP violation,
   expecting to look for neutrino/antineutrino; also to work on cross sections
   for netrino/nucleon interactions.

   MySQL database queried for metadata associated with each job: database is
   run by Triumf and queried remotely each time a job runs (but only once) to
   get the "beam summary data" (which'd otherwise be O(10G) data replicated
   with all jobs).  Grid also holds MC data.

   Data volumes are O(100TB) per cycle with 2-3 cycles/yr.  Using FST3 now,
   but not using replica registration (yet) - couldn't find documentation for
   how to do this, but following up.

   Using DIRAC, good example.  Triumf has SE exclusively for T2K, but may
   become part of T1.

   Needed manpower to develop integrate DIRAC, now using Ganga with DIRAC.
   Mark Slater also doing Ganga stuff with SNO+, so could usefully be
   contacted (or interested).

   WMS to be "retired", but not till there are alternatives also for non-
   WLCG VOs.  Services and support are needed but what's there may run out of
   support.

   Ack'ed GridPP in acks section.  As a general question, how should GridPP be
   acked?  Either as a free text, or citation of a GridPP (project) paper -
   but the authors on the latter are not necessarily the same people as the
   ones now running the project - most would be the same but there'd be churn.
   Check with PMB.

   In general it'd be worth taking this T2K forward as a success story and a
   case study, particularly now that results are coming in.

2. Ends, loose:

   - Towards a todo list for readiness for Run 2?
   - https://indico.cern.ch/event/336753/ - no one attending(?) but it's at
     CERN so maybe the usual remotery will work.
   -
That-ATLAS-Jamboree-in-December-for-which-I-don't-have-a-URL-readily-available-and-I-probably-wouldn't-have-permission-to-read-it-even-if-I-had
     Brian expected to attend, also Alistair Dewhurst from T1.

   - KeyDocs?! and other wiki related updates?
     Reminder!!

3. AOB

   NOB

jonperkin: (05/11/2014 10:01:18)
i can hear you
Christopher J. Walker: (10:07 AM)
https://ggus.eu/index.php?mode=ticket_info&ticket_id=109694
is presumably the ticket you are talking about. 
Jeremy Coles: (10:09 AM)
I think so.
Christopher J. Walker: (10:09 AM)
It is presumably worth getting the gfal2 developers involved
wahid: (10:10 AM)
will be next week at best - after 1.8.9 upgrades
1.8.9 is coming into epel this week 
DPM that is 
Samuel Cadellin Skipsey: (10:13 AM)
(whoops on the blogs - I do have another entry in my EC stuff waiting)
Christopher J. Walker: (10:14 AM)
http://londongrid.blogspot.com/2014/10/xrootd-and-argus-authentication.html
Not on the storage blog, but about storage
Samuel Cadellin Skipsey: (10:15 AM)
And useful, Chris :D
Jens Jensen: (10:15 AM)
Thanks. It only counts on *our* blog though ;-P
Christopher J. Walker: (10:19 AM)
Ganga has backends for both wms and dirac - so if you got that working, you wouldn't need to worry about backends. 
Ewan Mac Mahon: (10:23 AM)
Ganga is usable from scripts as well as interactively; I wonder if it might be possible/as easy to port the homebrew framework to use ganga rather than to use the dirac client tools directly.
wahid: (10:37 AM)
Sorrry I better go .. see you guys later
jonperkin: (10:42 AM)
Thanks everyone.
Ewan Mac Mahon: (10:50 AM)
As we know from Pete Clarke, other communities cannot possibly use X509, so we don't have anything they can use. We're a PP grid for PP people.