Attending: Jens (chair+mins), Winnie, Brian, Sam, Pete, Raja

Jens had not finished his intended stuff for the pre-GDB, at least not yet - something on accounting and publishing storage information
but there's quite a lot of stuff involving quite a lot of things (from GLUE2.1 to SwordFish, from StAR to experiment(al) JSON formats)
and it usually doesn't get a high priority so ...

RobA will be attending and has been asked to report back; also, being a pre-GDB, it can of course be attended remotely (if, necessarily,
mostly in passive mode.)  Also, helpfully, the following day's GDB will have a summary.


Regarding "our" talk at GridPP, Sam has started putting together slides.  Topics suggested today were:
  - why ZFS?
  - history of ZFS
  - neat features that we are using - or not yet using
  - Raja suggested ALICE may have had problems with a NDGF site in Finland, using ZFS behind dCache.  Sam montioned the performance
    of checksums have improved.  Also, dCache will do its own ADLER checksums and the filesystem checksum may not be necessary.
 
If there is time, we could also cover any updates on the T2-evolution (as discussed/proposed at the previous GridPP)
  - expecting "fate" of each site to be driven by experiments, but WLCG experiments need to be consistent in their intentions for
    each site, and as Brian points out, of course not clobber any work by non-WLCG VOs.  The last thing we want is for sites to
    need even more effort due to running N different storage "solutions"
  - any technical experiences (cf. Bristol's sole non-SRM HDFS based SE)

There will be no storage call next week.  See you at GridPP!


Jens Jensen: (06/09/2017 10:00:16)
Can you hear me Winnie?
Brian.Davies@RAL-LCG2: (10:03 AM)
winnie can you here me
:)
Raja Nandakumar: (10:25 AM)
http://trac.dcache.org/wiki/developers-meeting-20160112
http://trac.dcache.org/wiki/developers-meeting-20160322