P14 Reprocessing: Meeting of 19-Nov-2003: 10:15-11:00 VRVS Forest.

Agenda

1. Certification of sites with p14.05.02
2. When to start processing?
3. Technical Issues (Grid Certificates, python 2.2.2)
4. Status of Sites.

Minutes

Participants

Topics

1. Certification of sites with p14.05.02
   p14.05.02 certification sets copied to d0mino by GridKa (TMB and DST) and In2p3 (TMB and DST) and Imperial (TMB). They agree. Clued0 is used to produce a comparison set (expected to finish today). We should look at RAW vs. DST reprocessing (Mike, Harry and Suyong). Decision for those sites expected tomorrow.
2. When to start processing? Individual sites may start processing before certification is finalised, but should be prepared to through away results in case of certification problems.
3. Technical Issues (Grid Certificates, python 2.2.2, copy and storing scripts)
   • Grid Certificate: gridftp certificate expires on Dec. 3rd. make sure to update
   • Mc_runjob might show problems with python 2.2.2 which needs to be resolved as upcoming versions of SAM do need that version. The problem might be not coming from runjob.
   • The existing copy scripts can be used for TMB transport to FNAL when using
     export RERECO_TAGFILE_PATTERN="recoT*.tagfile"
     in the calling configuration script (Example is gridka_copytod0mino.sh). A new set of scripts is currently developed to allow tagged copying as well as tagged storing into DST. Before copying of storing, the number of events in each file will be checked. DSTs will be stored centrally until we notice problems.
   Improvements in routing can be partly due to reduced pick event activity. It might therefor reappear.
4. Site Reports:
   • Farm: is running 14.05.00 (except for certification dataset).
   • GridKa: p14.05.02 certification dataset processed and copied. Awaiting final certification.
   • Lyon: Using special station to push data to Lyon. Currently copying their input data. No delays in transport observed.
   • Nikhef: Non EDG: EDG: Using p14.05.01 for testing purpose. Around half of the set ran. The other half failed because it ended up at nodes with not enough space. As the testbed will be down for 2-3 weeks, investigating other possibilities. Move from EDG2.0 testbed to LCG-1 testbed.
   • NPACI: Working on TMBfix installation. Direct reprocessing is at lower priority.
   • SAR: (Jae by mail): MC production is first priority.
   • UK: Imperials certification plots are available; RAL is running certification set, about to finish.
   • WestGrid: All datasets downloaded. Certification sample running (3 files missing).

Action Items:

Next Meeting