Recipe for PAT layer 1 production

1. Setting up the environment
• put your name into SiteDB (follow the link)
• check out and compile the most recent tags
• create a sub-directory where you will run production
• setup your CRAB environment
2. Preparing the CMSSW configuration file
• copy the latest CMSSW configuration file (above) in your production directory
• check that the output file name is generic (e.g., patLayer1.root)
• make sure that this file "works" (if there is any error, please report) by running python <file_cfg.py>
3. Preparing the CRAB configuration file
• copy the attached crab.cfg file in your production directory and modify it:

Keyword	Description	Example
datasetpath	the sample you [[][subscribed]] to	/SUSY_LM0-sftsht/Summer08_IDEAL_V11_v1/GEN-SIM-RECO
pset	the CMSSW configuration file name (see previous step)	patLayer1_fromSummer08redigi_V#_cfg.py
storage_element	the site where you will store the data (see SiteDB)	T2_DE_RWTH
publish_data_name	the name of the dataset you are going to publish. Name convention: [Primary Data Set]+[CMSSW release]+"SUSYPAT"+[Tag number]+[version number]	SUSY_LM0_228_SUSYPAT_V5_v1
eMail	your e-mail address, to be informed of the progress of your jobs

4. Running and publishing into DBS
• once this is all setup, submit your jobs

   crab -create
   crab -submit

• you will be informed by e-mail when the jobs have finished
• once all jobs have completed, retrieve the results with crab -get
• if everything went fine, publish with crab -publish
5. Post-processing
• check the data has been inserted into DBS (you can find the full dataset path from UserProcessedDataset in any of the CMSSW .stdout files):

InspectDBS2.py --DBSURL=https://cmsdbsprod.cern.ch:8443/cms_dbs_ph_analysis_02_writer/servlet/DBSServlet --datasetPath=<full dataset path>

• run over it once to check that it was correctly registered
• put a link to the DBS entry of your published layer 1 file in the table below
6. Trouble-shooting
If something went wrong, you can revert ( with caution!):
• delete the samples produced from disk (using lcg-del and checking the location in the CRAB .xml files)
• remove them from DBS:

python $CRABDIR/python/DBSDeleteData.py --DBSURL=https://cmsdbsprod.cern.ch:8443/cms_dbs_ph_analysis_02_admin/servlet/DBSServlet --datasetPath=<full dataset path, as above>

• start over