Computing Meeting, Monday August 18 [message #7206] |
Fri, 15 August 2008 10:51 |
Jacek Otwinowski
Messages: 51 Registered: June 2005
|
continuous participant |
From: *gsi.de
|
|
Let's meet on Monday 18th (10:30) in cafeteria.
Agenda:
1. ALICE data reconstruction strategy
- Where will be data reconstructed (CERN / Karlsruhe / ...)?
2. Data transfer procedure CERN -> Tier 1 -> GSI (CERN -> GSI)
- Will be allowed to copy data directly from CERN from 10th of September?
- What part of data will be transfered from CERN to Tier 1 (Karlsruhe)?
- What is procedure to transfer data from Karlsruhe to GSI?
3. First physics analysis at GSI
- batch performance
- proof performance
Best Regards,
Jacek
Jacek Otwinowski
|
|
|
Minutes - Computing Meeting, Monday August 18 [message #7209 is a reply to message #7206] |
Mon, 18 August 2008 14:20 |
Jacek Otwinowski
Messages: 51 Registered: June 2005
|
continuous participant |
From: *gsi.de
|
|
Agenda:
1. ALICE data reconstruction strategy
- Where will be data reconstructed (CERN / Karlsruhe / ...)?
2. Data transfer procedure CERN -> Tier 1 -> GSI (CERN -> GSI)
- Will be allowed to copy data directly from CERN from 10th of September?
- What part of data will be transfered from CERN to Tier 1 (Karlsruhe)?
- What is procedure to transfer data from Karlsruhe to GSI?
3. First physics analysis at GSI
- batch performance
- proof performance
4. interactive desktops, move from 32 bits to 64 bits etch
5. lustre_alpha -> production lustre
////////////////////////////////////////
Minutes:
ad. 1, 2
- The strategy is not clear. It will be clarified with ALICE-off heads this week in Sibiu (Silvia, Kilian).
Requirements for first physics:
- 10^6 good RAW events from the first ALICE runs have to be transferred from CERN to GSI
ad. 3
- a new queue (with the highest priority) for the first physics productions will be created (Carsten)
- proof will be extended to 20 PCs (160 workers) in two steps (15 PCs this week, 20 PCs next week) (Victor). No data on local disks!
ad. 4
- move to 64bits etch will be postponed till a new etch client will be available (Thomas)
- tests with 32bits etch will be checked by Dariusz next week
ad. 5
- for first physics the lustre_alpha will be used!
- functionality of the lustre_alpha (well tested) must be kept till the production lustre will be stable and tested
- GSI transition strategy lustre_alpha -> production lustre must be known (Thomas)
Jacek Otwinowski
|
|
|