Choose timezone
Your profile timezone:
GridPP Technical Meeting
→
Europe/London
Virtual Only
Virtual Only
,
Description
Fortnightly meeting for technical topics looking further ahead than the weekly ops meetings on Tuesdays. There are also dedicated storage group meeting on Wednesdays. Each topic can go beyond the nominal 5 minute slot whenever necessary.
Federated Storage (which turned into a further discussion on sites in general)
Sam has been following up on the distributed storage front. While DPM say that they will provide distributed DPM but it may not be a distorted DPM but may not have it as a support configuration because of effort constraints. While distributed DPM would make the storage look simpler for the experiments it would not necessarily make things simpler for the sites as the greatest problems are from the servers rather than the head node. Also if a head node is responsible for several sites the you might want the head node to be more resilient. There is also a distributed dcache and this is to be tested and this is more developed that DPM so perhaps this would be better to test dcache.
Other distributed storage solution is of course xroot and using Marcus is looking at this for ScotGrid. There was a discussion as to whether or not it would better to have only a handful of sites with storage.
Sam will report back next meeting.
Alessandra reported on the GDB discussion (and her presentation) on what ATLAS wanted. As well as the federated storage solutions she pointed out that there was caching on for example the ARC CE and this seems to be popular. However there are clearly further discussion to be had. ATLAS also say that they can be run at SRMless sites.
She also pointed out that ATLAS want a difference between core sites and sites that may or may not be reliable.
Andrew M. pointed out thatcaching reduces site effort far more than distributed storage. If we have a mixture of sites with and without storage then we need have a system of pledges perhaps per Tier 2.
There was a discussion about DPM caching and ARC CE caching.
ACTIONS: Steve and Sam to draft a document specifying UK requirements so that we can define UK direction
VAC
Nothing new.(although did talk about this at the GDB).
ATLAS
Peter went to an OpenStack summit in Austin Texas. It was an "awesome" meeting. They have a "big tent" open management style. Quite a few HPC people looking at deploying Open Stack including federated sites.
CMS
Activity on getting ucernvm for T0
LHCb
Nothing new
GridPP Dirac
Dirac is running as normal. Had a load of LZ jobs failing because of a us VO development.Simon and Daniela will present GridPP experience at the next DIRAC workshop on the 23rd May.
Storage
Alitair D is looking at testing storageless T2 in Oxford. Testing is probably finished and this will be reported at te next g
Networking
There is an IPv6 F2F at CERN Wednesday.
Security
Ian was able to go to the pre-GDB but David reported on the status of the tracability ongoing work. They plan to have a F2F in a pre-GDB in July.
HEP Software foundation
There was a workshop since the last GridPP Technical meeting and Andrew said that this was a very positive atmosphere and that things are genuinely happening.
There are minutes attached to this event.
Show them.