Randall-Sundrum Analysis Package

Introduction

G* -> ZZ -> Jet + MET

Set up Environment - Understand the code

Set up a release of CMSSW (we are here using 3_9_9_patch1):

cmsrel CMSSW_3_9_9_patch1

cd CMSSW_3_9_9_patch1/src

cmsenv

Have to setup kerberos to be able to check out from cvs.

setupkerberos

Checkout package:

cvs co -d RSGraviton/RSAnalyzer UserCode/ThiagoTomei/RSGraviton/RSAnalyzer

Build package:

scram b

Initialize proxy:

voms-proxy-init

Create an archive named condor.submit with the following content:

executable  = /bin/ls
transfer_executable  = false
universe  = grid
grid_resource  = gt2osg-ce.sprace.org.br/jobmanager-condor
log  = teste.log
output  = teste.out
queue

If it doesn't work, you can try to use /bin/pwd instead of /bin/ls

Now, to submit to the grid:

condor_submit condor.submit

To check the status of the job on the grid:

condor_q YOURUSER

The output (in file teste.out) will be a path to where is your user related directory in OSG:

/home/OSG/uscmsYOURUSERNUMBER/gram_scratch_XmShfMl15l=

Now, go to the directory ZZ2q2nu, and open the file setupCMSSW.submit. The field arguments is which version of CMSSW you want to set up on OSG directory. The transfer_input_files is the .tar.gz package with the code.

Now, open the file setupCMSSW.sh. This is script which is run. Change de ID for yours, which is output in the previos step, i.e.

/home/OSG/uscms062 --> /home/OSG/YOURUSERNUMBER

Now, get the tarball:

wget http://cmssw.cvs.cern.ch/cgi-bin/cmssw.cgi/UserCode/ThiagoTomei/RSGraviton.tar.gz?view=tar 

mv RSGraviton.tar.gz?view=tar RSGraviton.tar.gz

Now, submit it to condor:

condor_submit setupCMSSW.submit

And check the output. If it's built, it's good.

The archive script.sh runs rsanalyzer_JetMET_PATsignalAndControlTogether_cfg.py, with the four arguments you give as an input (open the archive and check it!).

Now, open the python configure file it runs. In the beginning, you'll see all the analysis cuts used. Then, all the logic of the analysis path.

This is the file one should run to get all the histograms. The instructions on how to normalize them to the correct luminosity are on the file README.txt.

Running the Script

You can test if the script is properly running:

cmsRun rsanalyzer_JetMET_PATsignalAndControlTogether_cfg.py input W0j

The list of valid inputs are on the README.txt file, and corresponds to paths to the datasets in SPRACE.

If properly running, now one can use the condor to submit all the scripts from a given process.

As an example, let's run on condor the Z+jets invisible. Firts, one have to make the directory where the correspondent histograms will be saved.This should be the same as listed in the condor submit script condor_ZinvisibleJets.submit. You also have to make a copy of the big config file (rsanalyzer_etc) to the new directory.

mkdir ZinvisibleJets

scp rsanalyzer_JetMET_PATsignalAndControlTogether_cfg.py ZinvisibleJets/

Then you have to submit the condor job:

condor_submit condor_Zinvisible.submit

The output archives one will have in the folder are:

output.root  
resultado.error  
resultado.log  
resultado.out

-- FlaviaDias - 15 Apr 2011