Attending: Jens (chair+mins), JohnH, Dan, David, Sam, RobC, Steve, Winnie, JohnB, Matt

0. OPerational blog posts

   OK on blog posts for past quarter, thanks to all who wrote stuff, but particularly Brian who wrote most of them.

1. Continuing the follow-up from the WLCG workshop

   - Interest in storage accounting - generally the DPM plugin is doing the right thing; there is accounting, but
     the reporting and publishing (a la GLUE).  Jens had been looking more closely at SwordFish and the amount of
     data available inside - which is also REST-JSON and more standardised, but containing much more stuff than what
     we need to make available (such as physical location of disk servers).  GLUE is of course available also in
     XML and JSON flavours; both are quite verbose but XML has the advantage you can transform/stylesheet it.
     And both XML and LDAP can be queried (XML with XQuery/XPath).

     Difficulty getting agreement between experiments, particularly with ALICE who are doing things differently, and
     experiments are not keen on changing their models.  ALICE streaming data with xroot, also require information via
     xroot.  Sam points out it should have been sorted last year...

     On this topic, we either wait and see, or we prototype some stuff, or we work with what we've got.  And ALICE non-SRM
     is essentially just Birmingham.  We don't want to have to redo things for each experiment, and ideally we should be
     using open standards and/or existing work, rather than reimplementing information systems from scratch.

     Need recent versions of dCache and DPM to have the non-SRM accounting.  How is Bristol managing it, running DPM
     daemons but without SRM - check with Luke.

     Perhaps there is scope for a writeup which could turn into a publication (and not just for HEP)?  Sam and Winnie are
     also interested.

   - Site evolution - Sam's todo list still includes task to get in touch with Kashif about Oxford/RALPP cross site testing
     for CMS.

     This would also be useful (along with all the other stuff) to present/report at the September data pre-GDB.

   - Also the dual stack SE; Matt still aiming to try it out but needs to sort out other stuff, first, and obviously get
     site networking involved; iperf went down the wrong path, the main campus link.  Brian assisting.  We'd eventually need
     to figure out whether to volunteer Matt or volunteer someone else.

   - EOS in the UK: Sabah told us he'd be trying it out; also Mark Slater.

2. End of quarter; if you have anything particularly interesting, or known risks/issues, please send to Jens who will put it
   into the report for Pete.  Been a good quarter.  In this coming/current (ie Q3) quarter, we could make some of the above
   tasks our reporting targets.

3. Swordfish; as mentioned above Jens had been looking into it more closely again; will send more information around (or alternatively
   we'd devote a bit of time to introducing it in one of the coming meetings.)  Idea is, like Redfish, to have information about
   the systems, but Swordfish is obviously about storage-related extensions.

4. AOB

   NOB


Paige Winslowe Lacesso: (05/07/2017 10:03:13)
OF COURSE!!!!
Jens Jensen: (10:15 AM)
Depending on your definition of "standard"...
Daniel Peter Traynor: (10:25 AM)
EOS is just xrootd isn't it 
Robert Andrew Currie: (10:29 AM)
g2g have to another meeting!