Choose timezone
Your profile timezone:
LT2 Technical Support Group
→
Europe/Zurich
532 Blackett Lab (Imperial College London)
532 Blackett Lab
Imperial College London
Description
The meeting will be reachable by phone conference at
0845 862 0333 pin: 963852
Present:
* IC: D. Colling, O. van der Aa, M. Aggarwal
* IC-LESC: K. Sephton
* UCL: G. Sciacca, B. Waugh
* Brunel: P. Hobson
* RHUL/Brunel: D. Rand
* QMUL: G. Mazza
Experiment software installation status
It is still the case that site admins have to chase VOs to install and test software, although this is likely to change when the demand for resources starts to meet or exceed the resources available. In some cases the VO software installation is carried out by only one person, who is therefore too overloaded to investigate problems at each site.
It is hard to find and interpret information about the status of experiment software at a given site. CMS maintains a list of CEs to monitor that is distinct from the information system and requires sites to be added by hand.
A "champion" is needed to represent the interests of each experiment (ATLAS, CMS and LHCb) within the London Tier-2. The champion would tell us what is needed by their VO and would chase up problems within site and help us to prioritise them. They would be invited to LT2 management board meetings.
The CMS champion will be Stuart Wakefield.
DC will ask Ulrich to nominate an LHCb champion.
BW will find an ATLAS champion.
Experiment software installation contacts:
CMS: Peter Elmer
LHCb: Joel Closier, Roger Nandakumar
ATLAS: Kondo Gnanvo
LT2 interim coordination
Today is OvdA's last day in the post of LT2 technical coordinator. The post cannot be advertised until the advertisement is approved by Imperial HR, which should happen on Tuesday. Until the post is filled, a different person will attend the DTeam meeting on Tuesday each week, in addition to DC. There will be a short weekly phone meeting on Mondays at 2:30pm for sites to report their status and any problems.
We can continue to use the LT2 blog set up by OvdA at londongrid.blogspot.com and will use this to report any significant site problems in advance of the Monday meeting.
Upgrade to SL4
The experiments would like all sites on SL4/RHEL4 by 1st September, but not all at once, although CMS would like significant SL4 resources immediately.
The WN/UI middleware is available for 32-bit SL4 (and will run on x86_64 with the 32-bit compatibility libraries). A release for the CE is due soon if not already available. Releases for the other service nodes will follow. It is not clear what versions are available in the tarball format.
Brunel will shortly install a 32-bit SL4 cluster, and QMUL will install a 64-bit SL4 cluster with SGE. These can be used for testing of the VO software before the other London sites upgrade.
Links with the experiments
Already discussed under experiment software etc.
Glexec
The UK position is that sites will not be recommended to use Glexec for pilot jobs until it meets requirements including a code review to ensure that it is sufficiently secure to use.
Actions
All:
* read documents on Glexec linked from agenda, send comments to LT2 mailing list
* check srm-cp working (picking up Java correctly): source grid environment on WN and run srm-cp by hand with valid proxy
DC:
* tell Stuart he is the London CMS champion
* ask Klaus Rabbertz to add London CEs to CMS monitoring list
* ask Ulrich to find London LHCb champion
* contact Phillipa to get London helpdesk synchronised with GGUS
* arrange phone conference for Monday meetings
* tell Jeremy/Roger that LHCb should run SAM tests more than once per month, as results only available for 5 days
* combine comments on Glexec into a "London opinion"
DR:
* with DC and Stuart, find out from Peter Elmer how to get CMS software installed and know whether it is working
* attend next Tuesday's DTeam meeting
* set up LFC for LT2 VO, with help from Mona
BW:
* find out ATLAS production status and plans
* find London ATLAS champion
* find out what tarballs are available for SL4
OvdA:
* give admin access to blog to Mona
There are minutes attached to this event.
Show them.