Daniel Peter Traynor: (01/05/2018 11:08)
ucl use qm as well
Brian Davies @RAL-lCG": (11:08 AM)
UCL is
Daniel Peter Traynor: (11:22 AM)
east london
reduced cpu due to cooling
for may qmul has wall time atlas 75%, lhcb 13%, cms 5%, cepc 3%, ilc 1%, pheno 1%.
I can up other VOs on request
for specific runs 
Daniela Bauer: (11:27 AM)
@Pete: https://etf-cms-prod.cern.ch/etf/nagios/cgi-bin/extinfo.cgi?type=2&host=t2arc01.physics.ox.ac.uk&service=org.cms.WN-isolation-%2Fcms%2FRole%3Dpilot
(You can ignore the warning in the test.)
This means that CMS is using Singularity successfully at OXford and there's nothing you need to do.
@Dan -- I'll talk to you and Elena offline
Lukasz Kreczko: (11:28 AM)
check_webdav: error: no path specified. Using --path is mandatory with --hostname. Use either --uri or (--hostname, --port, and --path) to specify the target.
Steve Jones: (11:41 AM)
Only 5 million emails? 
Samuel Cadellin Skipsey: (11:43 AM)
So, re GridFTP session reuse timeouts: the ATLAS documentation was updated here: https://twiki.cern.ch/twiki/bin/view/AtlasComputing/Tier3gGridftpSetup

The key point is that you need to set:

control_idle_timeout 7200
(in your relevant gridftp configuration in /etc/ )
Chris Brew: (11:44 AM)
Puppet
Lukasz Kreczko: (11:45 AM)
May 1 10:12:23 io-37-02 puppet-agent[73679]: (/Stage[main]/Gridftp::Config/File[/etc/gridftp.conf]/content) -control_idle_timeout 600
May 1 10:12:23 io-37-02 puppet-agent[73679]: (/Stage[main]/Gridftp::Config/File[/etc/gridftp.conf]/content) +control_idle_timeout 7200

Paige Winslowe Lacesso: (12:01 PM)
Sorry I gotta go!
Daniel Peter Traynor: (12:05 PM)
i think i've seen a core reson why atlas are causing so many problems and maybe a route cause ofthe other vos tickets. https://ggus.eu/index.php?mode=ticket_info&ticket_id=134827
I got nhc scripts running at qmul now. they can drain and resume a node based in the checks. https://github.com/mej/nhc
John Hill: (12:13 PM)
I did it a year ago - with a few hacks it worked