Hi,
I had a look at the data files I produced (those in
/nfs/farm/babar/AWG12/ISL/tcls_sum02/data/2001 ) and encoutered the
following:
A "run -d -c" does not list all the runs "with a problem" and which might
have to be rsubmitted or the .tcl files split or..
.
When running over the .hbook files and trying to open ntuple 1
(CompBRecoNtuple) I got a much longer list of files where either
1) the ntuple contained 0 events --> 0
2) the ntuple contained -1 events --> -1
3) the ntuple wasn't booked at all --> NA
see
http://www.slac.stanford.edu/~henning/run-c-data2001.txt
http://www.slac.stanford.edu/~henning/run-c-newdata2001.txt
for a comparison.
The main reasons for jobs being recognised as crashed by the -c option are
1) time limit exceeded
2) tcl file contains no input collection
3) IslBrecoilUserApp: /u1/steinke/bfdist/packages/TagDataP/V00-01-34/TagAttributeBase.cc:112:
void TagAttributeBase::bind(opiHandle<BdbEventTag> &, opiHandle<HepExplorableDescr> &):
Assertion `! (( theDesc )==0)' failed.
Any ideas on how to proceed with this?
Henning
------------------------------------------------------------------
| Henning Flaecher |
| Physics Department [log in to unmask] |
| Royal Holloway Phone: +44 (0)1784 443479 |
| University of London Fax: +44 (0)1784 472794 |
| Egham, Surrey TW20 0EX, UK http://www.pp.rhul.ac.uk/~henning |
| |
| SLAC M/S 35, |
| 2575 Sand Hill Road, [log in to unmask] |
| Menlo Park, CA 94025, USA Phone: (+1) (650) 926 2630 |
------------------------------------------------------------------
|