privNanoAOD

How to privately produce nanoAODv5 samples

Setting up the framework

mainly based on the instructions given in https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookNanoAOD#How_to_check_out_the_code_and_pr

mkdir submitTopNanoAOD/
cd submitTopNanoAOD/
export SCRAM_ARCH=slc6_amd64_gcc700
source /cvmfs/cms.cern.ch/cmsset_default.sh
cmsrel CMSSW_10_2_15
cd CMSSW_10_2_15/src/
cmsenv
git-cms-addpkg PhysicsTools/NanoAOD

Update the following files to get the desired info for PDF and PS weights (more info about the modifications done here: afs/cern.ch/work/m/mullerd/public/priv_nAOD/modifications.txt):

cp -p /afs/cern.ch/work/m/mullerd/public/priv_nAOD/GenWeightsTableProducer.cc ~/submitTopNanoAOD/CMSSW_10_2_15/src/PhysicsTools/NanoAOD/plugins/
cp -p /afs/cern.ch/work/m/mullerd/public/priv_nAOD/nano_cff.py ~/submitTopNanoAOD/CMSSW_10_2_15/src/PhysicsTools/NanoAOD/python/

Compile it:

scram b -j 8

For CRAB submission

cd ~/submitTopNanoAOD/CMSSW_10_2_15/src/PhysicsTools/NanoAOD/test
cp -p /afs/cern.ch/work/m/mullerd/public/priv_nAOD/nano*_cfg.py .
cp -p /afs/cern.ch/work/m/mullerd/public/priv_nAOD/crabSubmit.py .

Update the following parameters in crabSubmit.py:

• config.General.requestName: name for your local directory
• config.JobType.psetName and config.Data.outputDatasetTag (X=6, 7 or 8 but the same in both entries)
• config.Site.storageSite: your storage element
• config.Data.outLFNDirBase: path of your storage element
• config.Data.inputDataset: miniAOD to be preocessed

source /cvmfs/cms.cern.ch/crab3/crab.sh
crab submit

Troubleshooting

In case you accidentally published a dataset twice under the same publication name, one can invalidate specific files of the dataset: https://twiki.cern.ch/twiki/bin/view/CMSPublic/Crab3DataHandling#Changing_a_dataset_or_file_statu

python $DBS3_CLIENT_ROOT/examples/DBS3SetFileStatus.py --url=https://cmsweb.cern.ch/dbs/prod/phys03/DBSWriter --status=invalid --recursive=False  --files=<LFN>

To invalidate a complete dataset, use this command:

python $DBS3_CLIENT_ROOT/examples/DBS3SetDatasetStatus.py --dataset=<datasetname> --url=https://cmsweb.cern.ch/dbs/prod/phys03/DBSWriter --status=INVALID --recursive=False