First, in order to use the grid and initialize the CRAB we need to do the following:

• At lxplus (for example using CMSSW_5_1_3):

source /afs/cern.ch/cms/LCG/LCG-2/UI/cms_ui_env.csh
cd CMSSW_5_1_3/src/
cmsenv
source /afs/cern.ch/cms/ccs/wm/scripts/Crab/crab.csh

• At access : copy the scripts (setupgrid.sh and setupcrab.sh) from Thiago's area /home/trtomei/bin/ and do:

./setupgrid.sh 
cd CMSSW_5_1_3/src
cmsenv
./setupcrab.sh 

We divide the FullSim in three steps:

• Step 0: Generating events and doing detector simulation using GEANT4

Using cmsDriver.py we create a config. file:

cmsDriver.py Configuration/Generator/python/Zprime_to_WW_8TeV_pythia8_cff.py --step GEN,SIM --beamspot Realistic8TeVCollision --conditions START50_V13::All --pileup NoPileUp --datamix NODATAMIXER --eventcontent RAWSIM --datatier GEN-SIM --no_exec

And we use the resulting configuration file (Zprime_M1500GeV_to_WW_8TeV_pythia8_cff_py_GEN_SIM.py.txt) to create and submit jobs with CRAB. For this we use a crab config. file, for example (crab02.cfg).

To create and submit the jobs we use:

crab -create -c task
crab -submit -c task

And when the jobs are terminated we use crab -getoutput -c task to retrieve the output of all jobs submitted. After this we can publish the results stored in the SE, using crab -publish -c task.

Obs.: user data will be published as

/<primarydataset>/<yourHyperNewsusername>-<publish_data_name>-<PSETHASH>/USER

where publish_data_name is the name you provide in crab.cfg and primarydataset is the primary dataset of the dataset you are reading (the first part of the datasetpath in crab.cfg). If you are running with datasetpath=None then the primarydataset is set equal to publish_data_name.

/<publish_data_name>/<yourHyperNewsusername>-<publish_data_name>-<PSETHASH>/USER

• Step 1: REDIGI

With the Step 0 results published we use these as input for the second step. This step is called REDIGI and we use a newer version of CMSSW to perform this (for example, 5_2_5). First, we create a config. file using cmsDriver.py:

cmsDriver.py REDIGI --step DIGI,L1,DIGI2RAW,HLT:5E33v4 --conditions START52_V9::All --pileup 2012_Startup_50ns_PoissonOOTPU --datamix NODATAMIXER --eventcontent RAWSIM --datatier GEN-SIM-RAW --no_exec

This will create a config. file, for example (Zprime_M1500GeV_to_WW_8TeV_pythia8_cff_py_GEN_SIM_REDIGI_DIGI_L1_DIGI2RAW_HLT_PU.py.txt). Using a crab config. file like (crab03.cfg) we create, submit and publish the results.

Obs.: This step uses some MinBias samples. So, if SPRACE doesn't have these samples, the jobs will fail trying to open!!!

Obs.: To publish in this step we need to use crab -publishNoInp instead of crab -publish, because CRAB doesn't know to migrate the MinBias dataset used for the pileup. See the HN: https://hypernews.cern.ch/HyperNews/CMS/get/crabFeedback/6000.html

• Step 2: RECO

Use as input the datasets published in Step 1. Create a config. file using cmsDriver.py:

cmsDriver.py STEP2 --step RAW2DIGI,L1Reco,RECO --conditions START52_V9::All --pileup 2012_Startup_50ns_PoissonOOTPU --datamix NODATAMIXER --eventcontent AODSIM --datatier AODSIM --no_exec

Create, submit and publish the results.

-- CesarBernardes - 12 Sep 2012