Week of 060703

Open Actions from last week:

Chair: H.Renshall

SMOD: T.Kleinwort

Monday:

New Actions: Thorsten to try gLite 3 environment on 10-20 worker nodes with a view to converting the complete batch farm tomorrow. This will allow CERN to publish its gLite CEs.

Discussion: ATLAS ran exporting data at about 400 MB/s until 5pm Saturday when IN2P3 dropped out losing 100 MB/s. P.Badino reported he had seen 'transfer protocol not supported' error reports from their dcache srm scheduler.

Tuesday:

New Actions: H.Renshall to check low gridftp rate with Atlas.

Discussion: FIO converted 60 WNs to gLite with no problems so will do the whole farm today. Many grid file transfer failures of CMS were due to double entries of files (reported to the Castor developers as a bug). FIO has cleaned up the double entries. FIO is adding disk servers for CMS and ALICE in preparation for their forthcoming challenges. lxplus 06 to 10 have a new castor client. Full roll-out scheduled for next Monday.

Wednesday

Actions:

Discussion: All worker nodes upgraded to gLite 3.0.1. The gLite CEs (ce103 and 104) will now be published in the site bdii. Three new LCG CEs will be added today.

Thursday

Actions:

Discussion: ATLAS exporting data at 4-500 MB/s and CMS have already reached their target of 150 MB/s. The 3 new LCG CEs are in production but not attracting jobs which needs to be understood.

Friday

Actions:

Discussion: