Enable crab to get Histo names from JobReport about crab2 HOT 3 CLOSED

Comment by afanfani on Tue Apr 8 07:24:21 2008

This might turnout not to be an entirly trivial change
since the current logic in CRAB (cms_cmssw.py,SchedulerGrid.py,SchedulerCondor_g.py)
rely on knowing the output filename upfront.

I've checked that in CMSSW_2_0_0_pre9 the FWjobreport does
contain the info about the file written with TFileService:

<AnalysisFile>
<FileName>/tmp/afanfani/afanalyzer/histo.root</FileName>
<Source Value="TFileService" />
</AnalysisFile>

however the FwkJobRep code in PRODCOMMON to parse the xml
is not up to date.
For that the files needed on top of the PRODCOMMON HEAD are: /afs/cern.ch/user/a/afanfani/public/CRAB/AnalysisFile/

from crab2.

Comment by slacapra on Wed Jun 11 12:58:22 2008

Done.

If TFileServic is in pset, the output file is automatically added to the list of crab output files.
Case if user already explicitly list the file is dealt with.

Can be turned off by setting
[CMSSW]
skip_TFileService_output = 1

in crab.cfg

from crab2.

Closed by slacapra on Wed Jun 11 12:58:22 2008

from crab2.