Only 3 tickets still open. No updates this week.
RAL is still a bit below the pledge line, despite increasing the ATLAS fairshare.
QMUL HammerCloud issues marked the site TEST over the weekend. Fixed on Tuesday.
Patrick is reinstalling the nodes without anything except CVMFS.
2.1 TB still on Sheffield disk. Elena has contacted Dimitrios, who said it would be cleaned by the end of the week.
The protocols for accessing Ceph are in AGIS. Oddly the setting for writing from jobs uses the "write_wan", not "write_lan" function.
Writing from jobs to Ceph still not working. The Rucio copytool does not upload the file correctly. We saw similar issues with RAL Echo. Tim asked Tomas Javurek to apply the same fix used for RAL Echo, but no reply.
QMUL is only GU UK site so far.
Manchester have unified their local clusters in preparation for GU, so could be next.
Alessandra: Testing GPU stuff for b-tagging scan. Not quite working yet.
Dan: New storage has arrived, so setting that up is taking most of his time.
Matt: ATLAS not keeping Lancaster queues full. Peter said there were problems with submissions; also there are no new single core jobs, and tessellation issues keep the multi-core jobs from starting quickly.
Vip: Oxford, like Lancaster, not getting many single core jobs from ATLAS, so just 86% CPU running.
Alessandra and Tim are going to the ATLAS Software and Computing Week.
New ATLAS-Tier1 Liaison, James Walder, will be there so he can meet ADC people.
No ATLAS UK Cloud Support meeting next week, unless requested.