Attending: Brian (or "Brain"), David, Duncan, Elena, Gang, Gareth, Jens
(chair+mins), John B, Sam, Steve, Wahid, Jeremy, Ewan, John H

-1. Was there a meeting last week?  Did I miss some notes being written up?

0. Operational issues?

DPM 1.8.8 issues - Wahid had a problem with config, but it could be
resolved (package dependency got a wrong package, this was before
the metapackage); also the public documentation could be improved
("installation of pool node is like a head node").  Ewan had 1.8.8 on
pool nodes, using Yaim, not puppet.

DMLite should be OK with puppet (1.9.8).

1. How is the WebFTS evaluation coming along?  Related to that, did we
ever find out what ATLAS connect is?

WebFTS needs RAL's FTS for the testing to be done, and RAL hasn't deployed
WebFTS yet - Andrew L is checking with the developers whether there is
a stable enough release.  There are links to the one at CERN, see chat.

ATLAS connect is a USATLAS only "federated" "cloud" - MTcondor flocking
with SSO and FAX behind the scenes.  There is also now an OSG Connect,
which is similar.

2. Summary of the GDB last week?  Mostly the configuration management
may be of interest, and IPv6, and EMI MW, and maybe the upgrade status.
http://indico.cern.ch/event/272621/

Wahid and Jeremy attended.  Also the pre-GDB on data access.  See chatlog
for details.

3. Data storage at hepsysman
https://indico.cern.ch/event/318354/

There are still slots available - Jeremy suggests cloud storage might
be useful but hasn't had any feedback.  RAL doing cloud storage with
EGI fedcloud Really Soon Now(tm).

4. AOB

   NOB

Gareth Roy: (21/05/2014 10:01:59)
purple rhubarb?
John Bland: (10:08 AM)
we have 1.8.8 and aren't aware of any issues (this doesn't mean there are no issues)
Jens Jensen: (10:08 AM)
But if the problem is with FTS, other sites and other SRMs should have the same problem.
Samuel Cadellin Skipsey: (10:16 AM)
(Ah, no, we have 2 CMS-oriented sites having problems with PheDEX/FTS on DPM 1.8.8 sites now. So, I'm calling this as a CMS/1.8.8 issue at the moment.)
Jens Jensen: (10:17 AM)
Yes, it'd be good to discuss them on the list - that's what it's for...
Also the wiki
wahid: (10:20 AM)
you can ask me (but I don't know castor) 
Brain@RAL: (10:21 AM)
which were the cms sites with issues?
Jens Jensen: (10:21 AM)
http://indico.cern.ch/event/272620/contribution/11/material/slides/1.pdf
Samuel Cadellin Skipsey: (10:21 AM)
Brian: T2_RU_INR and T2_IN_TIFR
Re (the next item, ATLAS Connect): http://connect.usatlas.org/ It's basically US-ATLAS trying to get in on the "Federated Cloud" thing.
wahid: (10:23 AM)
There is also now OSG connect
http://osgconnect.net/assets/files/OSG%20Connect%20Handout%20-%20AHM%202014.pdf
Jeremy Coles: (10:24 AM)
The FTS3 link was https://fts3oradevel01.cern.ch:8446/. You are probably correct.
Ewan Mac Mahon: (10:25 AM)
"It looks like you're transfering a dataset, would you like some help with that?"
wahid: (10:26 AM)
"Efficiently connect your science to cycles and data"
Brain@RAL: (10:30 AM)
regarding dodgy cm tramsfers: http://dashb-fts-transfers.cern.ch/ui/#date.interval=2880&dst.site=%28t2_rU_inr,tifr%29&grouping.dst=%28site,host,country%29&m.content=%28errors,successes%29&p.grouping=server&tab=transfer_plots&vo=%28cms%29
Jens Jensen: (10:30 AM)
http://indico.cern.ch/event/272621/
wahid: (10:31 AM)
Jeremy and I were there
Jens Jensen: (10:31 AM)
https://indico.cern.ch/event/272621/contribution/5/material/slides/0.pdf
wahid: (10:31 AM)
http://indico.cern.ch/event/272621/contribution/5/material/slides/0.pdf
Ewan Mac Mahon: (10:33 AM)
I've seen it falling over a lot, haven't noticed problems with restarting it though.
It's odd though - it's not like apache httpd is generally fragile.
Samuel Cadellin Skipsey: (10:35 AM)
Indeed. I think it's something whereby the dpm plugin wierdly times otu.
...out
(Update on the CMS/FTS/DPM1.8.8 issue: it might be a problem with FTS *2*)
wahid: (10:38 AM)
if we can spot what the cause of http flakeyness it would be interesting to kno
Samuel Cadellin Skipsey: (10:38 AM)
(Apparently we have perfectly happy FTS3 transfer tests against 1.8.8)
Jens Jensen: (10:38 AM)
https://indico.cern.ch/event/318354/
Samuel Cadellin Skipsey: (10:38 AM)
Wahid: well, we should take an action to prod this, and throw logs at Fabrizio. ;)
Ewan Mac Mahon: (10:40 AM)
We should find out more about this too:
https://www.ja.net/products-services/janet-cloud-services/amazon
Looks like the upshot of the janet peering with amazon thing from ages ago; might allow the use of AWS without the ruinous bandwidth charges.
wahid: (10:41 AM)
I won't physically be at hepsysman btw - but if there is vidyo I'll try and be on it 
Jens Jensen: (10:44 AM)
How many FTSes do CMS have?
wahid: (10:44 AM)
we are not aware of any issues is the summary