Log onto fcdflnx3 as cdfopr in a terminal window. You will
read and edit the file LIST_OF_PROCESSED RUNS, and you will
setup and run the root job to make the physics histograms.
(i.e. make ntuples for all the runs processed by the
Express Production (stream A) ). To do this run emacs in the background
> emacs &
If you are unsure how to do this, you could log on with TWO terminals and do emacs on one and root in the other
- Background Information. A cron job by user cdfopr on fcdfsgi2 automatically
runs the ~/scripts/update_list.sh
script each hour to update the file ~cdfopr/LIST_OF_PROCESSED_RUNS
with recent runs which have EXPRESS production output in
the directory fcdfsgi2:/cdf/data01/stream_a.
(Do crontab -l on fcdfsgi2 to see cronjob command.)
Further background information: A cron job by user cdfopr on fcdflnx3 automatically runs the ~/scripts/make_stntuple.csh script each hour to launch an stnmaker job to make the stntuples for the last run in LIST_OF_PROCESSED_RUNS that does not have an stntuple yet.
(Do crontab -l on fcdflnx3 to see cronjob command.)
The resulting ntuples are stored in the
fcdflnx3:/cdf/scratch/cdfopr/val/results directory
> ls /cdf/scratch/cdfopr/val/results
If this cron job is working properly, it will put the word "done"
after the run number in the file ~cdfopr/LIST_OF_PROCESSED_RUNS
after an stntuple file is created for that run.
The log file for this cron job is /cdf/scratch/cdfopr/val/results/stnmaker_cronjob.log
- FTP the completed ntuples and logfiles to fcdfdata033: ntuples to the /val directory, ntuples to /val/results/
- Add the new ntuples to the catalogue in fcdflnx3:~/cafdfc/stntuple/express/. Be sure to remove the emacs backup files (rm *~) and then check that the new ntuples are OK by running root from the val directory:
root[0] .L root_scripts/stra_mon.C
root[1] make_hist(166778)
- where in the example 166778 is the run you want to check. If you get unrecoverable error messages comment out the offending file.
- Make physics histograms (1 set of histograms and 1 output histogram file per
run) for each new run that has an stuntuple file:
Prepare root for each of the new runs you wish to process by doing the following:
> cd /cdf/scratch/cdfopr/val
> source_me
For each unprocessed file listed in the /LIST_OF_PROCESSED_RUNS file, do this:-
> root.exe -q 'root_scripts/make_histograms.C(166778)'
etc.
The Histograms are stored in the /cdf/scratch/cdfopr/val/hist subdirectory
- Next,make a 2-nd level (physics monitoring) ntuple
> cd /cdf/scratch/cdfopr/val
> source_me
> root.exe -q root_scripts/make_physmon_ntuple.C
- Run root and plot the run-dependent histograms: these will appear automatically in windows
First start root. Must run root in the same directory where the macro is situated.
>cd /cdf/scratch/cdfopr/val
> root.exe
Then cut and paste the following arguments into the root[n] command that appears in your terminal window (include the first .)
root[0] .x root_scripts/plot_nwenu.C("physmon.root")
root[1] .x root_scripts/plot_nzee.C("physmon.root")
root[2] .x root_scripts/plot_zee_mass.C("physmon.root")
root[3] .x root_scripts/plot_nJpsi.C("physmon.root")
root[4] .x root_scripts/plot_jpsi_mass.C("physmon.root")
root[5] .x root_scripts/plot_eoverp.C("physmon.root")
Look carefully at the histograms:
Number wenu - Is the result within about +-0.75 of the average result? Is there any change in the current trend of values?
Number zee - Results should be less than 0.1 and not 0. Is there any change in the current trend of values?
Zee mass - Result should be flat within the errors across all the runs. The latest runs should align with earlier data.
nJpsi - Are the results within +-1.0 of the current average. Is there any change in the current trend of values?
Jpsi mass - Result should be flat within the errors across all the runs. The latest runs should align with earlier data.
E/p - .....
Report problems to the Scico (shift leader) immediately.
Finally, type .q to quit root.
Old instructions (these are now done by cron jobs described in steps 1 and 2
above):
- see FCDFSGI2:/cdf/data01/stream_a for the list of runs processed by Express
Production
- see /cdf/scratch/cdfopr/val/results/LIST_OF_PROCESSED_RUNS for the list of
already processed runs. Edit this as you process runs.
- make ntuples for the runs which don't have such:
> cd /cdf/scratch/cdfopr/val
> source_me
> run_stnmaker.sh run_number &
where run_number is the run # , corresponding to the run processed by
express production on FCDFSGI2 (find express production output in
/cdf/data01/stream_a directory there,
check the date, take a while for the files being written)
example: run_stnmaker.sh 25a61 &
note: ignore HLOJet error
To convert run number from HEX to decimal:
> printf "%i\n" 0x25a61
154209
and from decimal to HEX:
> printf "0x%x\n" 154209
0x25a61
Last updated Jul 12, 2003 by Pasha Murat