Reprocessing: Meeting of 21-Mar-2005: 9:30-10:30 ESNet video conference

Agenda

1. News
2. Last tweaks before startup
3. Status of remote site Certification
4. JIM Deployment and Remote Setup of remaining sites.
5. AOB.

Minutes

Participants

Topics

1. News
   
   • The start of p17 reprocessing is approaching quickly. p17.03.02 is frozen since Thursday night. Iain promised tar-balls for Monday evening.

     Our plan is to do a quick test on the d0farm by rerunning the production certification with vs without JIM, before we give green light to the remote sites.

   • p17.03.02 runs beautyfull on the farm. Seems much more stable: only 1 crash in four days of running. This is better than 14.06.02.
   • Efficiency plots on the web.

2. Last tweaks before startup (DW)
   In order to prepare the final tweaking of the site that are ready to go I've upgraded the p17 reprocessing page. The description of the software releases was moved to this page and some words on operation were added. The March release definition contains a small addition and specifies samgrid_util v2_0_1. This is needed to clean up the durable location i.e. remove files which are no longer needed.

   Please install samgrid_util v2_0_1 and d0repro v0_7_4 (reread the install notes) and reread the instructions on web pages. These upgrades shouldn't alter any of the products involved in the actual production.

3. Status of remote site Certification
   • Status of production certification.
     We have plots for Lyon and Westgrid on our certification page.

   • Status of Merge Certification of Sites
     
     • DØFarm: done.
     • WestGrid: done.
     • Lyon: done
     • SAR: UTA: done
     • GridKa: Not merged yet (see below).
     • Wisconsin: Plots available. (Awaiting preemption to be turned off again, to submit production certification)
     • Prague: Ready to go (should run their merging datasets)
     • UK:
       • Manchester: Ready to go (should run their merging datasets)

4. JIM Deployment and Remote Setup of remaining sites.
   • GridKa: Last week jobs failed due to SAM station restarted; GridKa NFS performance slow. When choosing Fermilab SAM naming service, jobs run OK. With the local naming service turned on error messages occur in the d0reco task-log file. Without a bug fixed mc_runjob the local naming service can't be used.

     Test jobs are queued since 3 day; pbs fairshare for dzerojim needs to be increased. When they go through GridKa will proceed with merging. Thereafter load tests should be performed and the transport mechnism of raw files should be optimised.

   • UK:
     • IC: Success rate at 90%. Problem storage notification never comes back. Merging under test.
     • Manchester: Ready to go (should run their merging datasets)
     • RAL: Configuration options under investigation.
   • CMS Farm: Joel reported Condor problems, will come back to us. Check with Ruth Bordes
     

5. AOB:
   

Action Items:

Next Meeting