Feb-25-2004

        Prescription for removing bad runs/LBN from calo/jet/met 


1) Bad CAL runs from the Quality-DB (available on full statistics)
They have been classified bad if a clear hardware problem was detected,
or if they fail the off-line analysis with DQ_CALO (S. Shary).

Note that the bad dq-calo runs were not in the DB until today, so as of
today there are 35 additional bad-CAL runs in the DB (less than 0.5%
of the lumi). If you were not already fetching this info directly from the
in2p3-calo-quality page, then you need to update your bad-CAL run list
from the DB for the next round.

2) Bad JETMET lumi blocks (available on full statistics)
Lists availables with and without T42 from:
http://www-d0.fnal.gov/computing/algorithms/calgo/jet_met/runsel.html

The corresponding JETMET run-selections are also available but
NOT RECOMMENDED since they entail larger data loss (about 10% of the
lumi).

3) Bad Ring-of-Fire lumi blocks (available after run 175030)
List available at:
http://d0-france.in2p3.fr/WORKING_GROUPS/DQ/RF/lum_blocks_mht30_bad.dat

The bad luminosity blocks list was produced using the MHT30 skim.
This list will be updated in the next weeks for the whole statistics, but
it is  usefull to use the partial one already now.

In conclusion, 1) 2) and 3) should be applied  and the overall lumi loss
(after removing the summer 2002 period) is below 5%.