Week of 050808

Open Actions from last week:

• testing of smoke tests (James - R/GMA - FTS) DONE (R-GMA)
• new FTS release / deployment plans.
• Maarten: Look into GRIDVIEW numbers - Phool DONE
• GRIDVIEW DB to be upgraded (Monday) DONE
• Check for "lost" files ( Vlado) DONE
• FTS Sensors - Scripts to write (Gavin) DONE

On shift : James/Patricia

Monday:

Log: Nothing

New Actions:

Discussion:

• Want to test new FTS1.3 hard against a site DONE
• gridview update - stop producers for duration DONE
• Code freeze date for FTS with srmCopy 15th August
• Got a spreadsheet for resource requirements from CMS for SC3.

• FIO suggest to not use Phedex pools, but re-use some of the other pools instead. Need to discuss what longer term (service phase) plans for the pools are - 1 or 2 (or 4???) and are they in WAN area or LAN area?

Tuesday:

Log: No

New Actions:

Discussion:

• R-GMA team will come and give a tutorial when we want one
• boxes are ready for DPM deployment. VLADO
• early next week for deployment of 1.3 - working on docs. =

• Simone: 2.6.0 is out now. Probably no chance of it being installed by Sept 1st. Is this a problem? Probably not...

Wednesday

Log: Nothing to report for service. Tests ongoing with DESY/INFN. INFN seems to be better tuned - 5 streams/10 concurrent.

Actions:

Discussion:

• How to use the phedex pools for CMS. They're now clean (no replicas).

Thursday

Log: Load noticed on GRID4 - GRIDVIEW from last night. Need to investigate.

Discussion:

• Found problem with FTS not filling queues -was due to a query to the DB - worked around with a query hint. QF should be released today.
• Added UNOSAT as another VO for LFC
• Met FIO/FS yesterday about requirements for Grid machines going to be managed by FIO. Need to state the reliability and storage requirements.

Actions:

Friday

Log: Nothing to report. Rate overnight to INFN/NDGF dropped ~80% with no changes.

Actions:

• To look at FTS for preasons for dropping rate (Paolo/Gavin)
• Upgrade UI Node to 2.6

Discussion:

• Phool found problem with GRIDVIEW DB - an agent that should have been running against the dev DB was running against the production one.
• Resource requirements from all experiments are provided