A phone meeting via CERN took place Monday 16 Jan 2006 at 16:00 UK time
Present: RJ(Chair), AF , DT , MAH
Topics for discussion:
(0) Upgrades to Tier-2 computing hardware:
An email from PPARC and the e-science office (Janet Seed) to grant holders was sent around
last year that said there would be a second call for bids "at the end of Jan 2006 or in the Spring".
This was for Tier-2 operations (technical and support staff) and upgrades to hardware.
RJ has asked that a precise description of what can be bid for is attached to the call when it comes.
Note that GRIDPP2 finishes in 2007(?) and there is no GRIDPP3 to run UK GRID. People will have to be co- opted from the
running experiments. Buying new kit is not to be done by depreciation of existing kit under FEC.
RJ has heared rumours that there may be a new SRIF call....
(1) wLCG Collaboration Board and Chair:
The world-wide LCG collaboration is about to have its first meeting on 3rd Feb at CERN. This body should operate like the
collaboration board of an experiment, covering the World GRID.
RJ is a member of this board because of his NorthGRID Chairmanship. At the meeting, a new Chair person will be elected: RJ suggests
NorthGRID votes for the chair of the Prague Tier-2 , Milos Lokajicek.
NorthGRID can raise major GRID-related issues with this body.
(2) VO access:
In principle, NorthGRID sites should be as open as possible to allow all registered GRIDPP VOs to access the facilities. However
to give sites control over its facilities, it is proposed that AF sets up a Fair-Shares template for the NorthGRID sites to control
access to facilities.
There are advantages in setting up a specific NorthGRID VO that could be controlled via the Fair-Shares template. AF will do this.
(3) Technical Co-ordinator's report: AF gave a brief report of the status of each NorthGRID site.
(4) Service Challenge 4 (SC4).
The next major Service Challenge will start in the first quarter of 2006.
The Milestones for SC4 are available on the GRIDPP Wiki (AF to supply the URL). They include a sustained data transfer exercise
and stress testing network connections.
AF has chosen dCache as the SE to be used by NorthGRID.
AOB: There was some discussion of the requirement in the LCG MoU that sites should have an up-time of 95%.
It is not clear what this means in terms of the fraction of nodes that should be available. Access to data stored on SE was seen as
the chief requirement, as in general it will not possible to have copies of data stored at different sites.
To get experiments to use Tier-2 sites, some sites will be designated storage locations for ATLAS data.
Summary of Actions
Report on wLCG meeting: RJ
Make template for Fair-Shares VO access-to-sites control mechanism: AF
Set up NorthGRID VO: AF
Check Service Challenge 4 Milestones and work towards them: ALL
Date of next meeting: No specific date was decided: Before the Service Challenge 4 starts??
|