Attending: Jens (chair+mins), Steve (chatlog), Sam, Matt, Teng, Dan, Brian, JohnH, Luke, RobC, Winnie


0. Op. b.p.

   Quite a few operational issues this week, mostly relating to timeouts apparently, where SRM has a five minute
   hardwired timeout; however, SRM is supposed to be asynchronous and things shouldn't time out - but clients tend
   to be clever and keep connections open for reuse which may then cause problems.

   At approximately this point, one of RAL's work-preventing measures kicked in and disconnected Jens and Brian.
   Sam, however, has tried to summarise the state of knowledge.  It seems to be related to an old gSOAP which
   ignored the timeout, and a newer one which honours it thus causing problems?

   Luke also reports a problem which seems to involve a redirect from IPv6 to IPv4, in GridFTP.  Whether it applies
   to xroot remains to be seen.

   On a related note, Brian reports US sites implementing 2hr time out for data transfers (even busy ones?! that
   seems daft, even by grid standards.)  So in practice this may limit the size of files that can be transferred.
   There is also an idle timeout where, if the client does not receive a response, it will, er, time out, but this
   seems more sensible.

1. Discussion of yesterweek's GDB

   Dynafed - needs information about how much information is available.  Obviously it cannot just ask the
   information system...!?

   Small sites evolution - to be revisited later; also outcome of the DPM workshop will be relevant.

   OSG - change of funding may affect us or our services to users who use both us and OSG.

2. I have volunteered Brian to give us an update on the CERN datalake.

   Brian reports more sites participating; other than SARA, RAL, Aussie sites, 3 Russian, and 1 Italian.
   Reinventing the grid? or AFS? or iRODS? or HDFS? or dynafed?  Still lots of open questions.

   VMs accesing the data lake to be offered to NorduGrid for their distributed storage.

3.As a reminder, HEPiX is today with the storage/data session this
afternoon (I;'ve got it starting 3pm UK time but I might be wrong...)

4. AOB

   Jens away next week



Brian Davies @RAL-LCG2: (16/05/2018 10:10:25)
Im back
Steve Jones: (10:10 AM)
hi... jens?
Brian Davies @RAL-LCG2: (10:11 AM)
jens will be back in a munte
Samuel Cadellin Skipsey: (10:11 AM)
Problems with RAL networking, Brian?
Lukasz Kreczko: (10:14 AM)
" i was looking at the CMS IPv6 storage status and reminded about
your SRM endpoint redirecting to an IPv4 address. Did you get any
help/usefull hints from teh DPM team? (I re-run our test and it
still shows the redirect to IPv4, 137.222.79.15.)"
jens: (10:39 AM)
https://indico.cern.ch/event/651353/
Daniel Peter Traynor: (10:48 AM)
Brian Davies @RAL-LCG2: (16/05/2018 10:10)
Im back
Steve Jones: (10:10 AM)
hi... jens?
Brian Davies @RAL-LCG2: (10:11 AM)
jens will be back in a munte
Samuel Cadellin Skipsey: (10:11 AM)
Problems with RAL networking, Brian?
Lukasz Kreczko: (10:14 AM)
" i was looking at the CMS IPv6 storage status and reminded about
your SRM endpoint redirecting to an IPv4 address. Did you get any
help/usefull hints from teh DPM team? (I re-run our test and it
still shows the redirect to IPv4, 137.222.79.15.)"
jens: (10:40 AM)
https://indico.cern.ch/event/651353/