| Meeting on Thursday, 26th of July 2007 [message #4749] |
Thu, 26 July 2007 15:21 |
Kilian Schwarz
Messages: 91
Registered: June 2004
Location: GSI, Darmstadt
|
continuous participant |
From: *gsi.de
|
|
ALICE Computing Meeting
Protocol: KS
- xrdcp
each node believes it is a redirector. According to recent investigations this is a known problem and does not hurt, though. The problem has been fixed a short time ago in CVS.
The experienced copy problems are most probably due to permission mismatch with directories created manually.
Silvia experienced xrd copy problems to a certain directory and created a corresponding ticket.
A procedure for deleting files in the xrd cluster should be found.
- transfer of data from CERN to GSI:
the transfer jobs are currently pending in status "transferring". Pablo will look into the problem. The acchieved transfer speed so far is 69 Mbps with 8 parallel streams.
A script method is needed to automatically create a list of LFN/PFN copied to GSI. This should be done during the mirroring process preferably.
- visit of Federico:
this will be a topic of the analysis meeting on 21st of August.
A dedicated meeting will be on 27th of August (11:00-14:00). KS will order Pizza.
Federico will come on August 28th.
Talks: - Analysis Scheme (Marian)
- plans for 2008 (Silvia ?)
- status and visions of ALICE T2 (Kilian)
discussion about role of T2 centre started:
According to ALICE computing model a T2 centre should deliver simulation events and send it to CERN or T1. This results in too much disk/tape space needed at T1. To save disk space GSI could serve as analysis centre only. Especially since analysis took place where Physicists are sitting ==> at T2s rather than at T1s. A T1 centre is a dedicated production site for simulation. At GSI we have the detectors TRD and TPC and provide the heart of calibration and analysis, therefore GSI deserves a special treatment.
What amount of data of end of the year run should go to GSI in 2007 ? 1 event of non zero supressed data: 250 MB. Per day 400/500 GB can be copied to GSI.
- Christmas Break
machines will be left running. Maintenance on best effort basis during working days by call.
- PROOF tutorial
Anna will give a general presentation, then simple PROOF exercises based on GridKa School 05 will take place. Afterwards Marian will show real life ALICE examples. To make sure that only a limited number of participants will start as test case the tutorial shall start with:
- Ana Marin
- Stefan Gaertner
- Raphaelle Bailhache
- Lars Bozyk
- Alexander Kalweit
- Wilfried Monange
- Alex Bercuci
- Sylwester Radomski
everybody else will follow in a second tutorial, based on the eperience with the first one.
|
|
|
|
GSI Forum
Search
Help
Members
Login
Home
