P14 Reprocessing: Meeting of 7-Jan-2004: 10:15-11:00 VRVS Forest.

Agenda

1. Status of Sites.
2. What remains to be done?
3. What needs to be improved in the next round?
4. Ideas about writing down our experiences.
5. AOB.

Minutes

Participants

Topics

1. Site Status:
   • Farm:
   • GridKa:
     5-Jan-2004: Finalised all assigned datasets (~21M events). All TMBs copied to d0mino, all except a few DSTs available in SAM (located at GridKa).
     Assigned files: 10662. Failures: 61. Of these 50 not located in SAM, 9 bad on tape, 1 genuine reco crash, 1 with non-reconstructable event.
   • Lyon:
     6-Jan-2004: Finalised assigned datasets (~36M events). All TMBs transported to d0mino (by bbftp). DSTs available in SAM (located at ccin2p3).
     Assigned files: 18896. Failures: 10. Of these 4 not available in SAM, 5 with crash of infinite loop, 1 with DSPACK error.
   • Nikhef: 2998 processed, TMBs at FNAL. DST stored at Sara (not in SAM).
     Assigned files: 3399.
     EDG: Files shall be certified as a test for.
   • UK: 7-Jan-2004: All assigned datasets processed (19.5M events). Around 1000 TMBs (of 12700) to be copied to FNAL. Most DSTs in SAM at FNAL.
     Assigned files: 12333.
     Mike: Some files need rerunning. Sabah will do so.
   • WestGrid:
     5-Jan-2003: Finished allocated datasets (~12M events). Many TMBs and DSTs to be copied to d0mino.
     Assigned files: 6268.
2. What remains to be done?
   Copy TMBs.
   Mike suggests to leave DSTs at their location if they are in SAM. Other sites should store the DSTs at FNAL.
   (Nikhef plans to add SAM acces to their Sara.)
   Events missing in the Web-page bar are due to failures in the summer.
3. What needs to be improved in the next round?
   Some ideas in radnom order: File delivery wasn´t good. Next time from RAW. Creating datasets is clumbsy need automatic request system.
   More integrety checks in transport back. Move out d0mino-tasks to remote sites, especially merging. Thus DB access is needed. Collection job is needed. Alternative: Store each TMB into sam before merging. (Maybe transient SAM-files?) EDG: has problems with a node needs SAM ansatz. SAM client in RTE would be needed.
4. Ideas about writing down our experiences.
   Please send your ideas to Mike and Daniel.
5. AOB.
   • Announcement in D0 News. Mike as soon as merge of remaining files is
   • FermiNews article.
     Amber: FermiNews people are looking for ideas of photos (e.g. Postcards from each participating site); So if you have suggestions for nice graphics for that article please mail us.

Action Items:

• Mail problems you had during the reprocessing to Mike and Daniel.
• Mail ideas on graphics for an FermiNews article to Mike and Daniel.
• Estimate the person power that went into reprocessing at the individual sites: Separated by setup work and operational work.
• Send ideas of what to improve the next time to Mike and Daniel by mail.

Next Meeting