Reprocessing: Meeting of 11-July-2005: 9:30-10:30 ESNet video conference

Agenda

1. News
2. Features for new JIM release cut (Andrew Baranovski)
3. Status of remote site Certification
4. Status of production
5. JIM Deployment and Remote Setup of remaining sites.
6. AOB.

Minutes

Participants

Topics

1. News
   
   • p17.05.xx won't be deployed for reprocessing. We think about using p17.05.xx for recovery of crashed or corrupted files.
   • Marco Verzocchi found 7 files (so far, after check 90% of what is done) which crash when being read with dsdump.
     
   • General policy of logging and how long they should be kept was discussed. Log files should ideally be kept 6 months. To have enough space we will investigate whether to remove core files.

2. Features for new JIM release cut (Andrew Baranovski)
   
   • Job throttling:
     Control / reduce rates jobs are submited to a site based on the current snapshot of all jobs in the submitted state. The goal is to reduce machine load during submission and improve operations.
   • Ability to configure sam batch adapter batch queues on per application basis.
   • Ability to configure data queues (fcp) and storage locations on per application basis (binary download, reco data, merge data, reco output , merged output ) to improve CPU utlization.
   • Multi site submissions. Random , "most data" criteria brokering when selecting target site per submission attempts.
   • Web monitoring improvements.
     -- w3c compliance changes.
     -- GMT to local timezone conversions.
     -- monitoring filters per user and per site.
     
   This has to be cut to a formal release, which then will be tested on the FNAL farm. Timescale: a bit more than one week.

3. Status of production
   • WestGrid: --
   • Lyon: Eagerly awaiting new cut.
   • SAR-UTA: Limited to 5 merge jobs, to be investigated.
   • SAR-Oscer:
   • CMS-Farm:
   • Wisconsin:
   • Prague: Stuck with expired host certificate.
   • Manchester: --
   • Imperial College: Jobs are going to idle status; durable location seems at the limit.

4. Status of remote site Certification
   • Status of production certification.
     
     • GridKa: produced opteron and xeon results separately. Joe will provide plots asap.
     • SAR-SPRACE:

   • Status of Merge Certification of Sites
     
     • DØFarm: done.
     • WestGrid: done.
     • Lyon: done.
     • SAR: UTA: done
     • SAR: Oscer: done
     • GridKa: done.
     • Wisconsin: done.
     • Prague: done.
     • GridKa:
     • SAR-SPRACE: dataset will be assigned.

5. JIM Deployment and Remote Setup of remaining sites.
   • RAL: Admins are insisting on using an existing gatekeeper.
   • Lancs: --

6. AOB:
   
   • Gavin Davies: Resources will have to be balanced, i.e redirected towards MC production.

Action Items:

Next Meeting