20140627 - Samir - First AAA/SRM Stageout test in the T3

• Instructions followed
• Input - /BprimeBprimeToBZBZinc_M-950_TuneZ2star_8TeV-madgraph/Summer12_DR53X-PU_S10_START53_V7C-v1/AODSIM
• CRAB - /tmp/CRAB_2_10_6/crab.sh

• CRAB Configuration :

[CMSSW]
output_file=outfile.root
total_number_of_events=1000
number_of_jobs=5
pset=tutorial.py
allow_nonproductioncmssw=1
datasetpath=/BprimeBprimeToBZBZinc_M-950_TuneZ2star_8TeV-madgraph/Summer12_DR53X-PU_S10_START53_V7C-v1/AODSIM

[USER]
se_white_list=T2_US_Caltech
copy_data=1
storage_element=T2_US_Caltech
return_data=0
user_remote_dir=testcrab2
se_black_list=T2_US_Wisconsin

[CRAB]
jobtype=cmssw
scheduler=condor

Jobs ran smoothly, stage-out was not as fast as the local one, but all output files were found.

20140605 - Samir - Example3

AAA Testing

This page is to aggregate the experience of users as they test workflows with remote input. After this is populated, we can have a feeling on how safe is to rely in this mechanism, which will possibly allow us to manage our storage resources better.

Tests

• Date format - YYYYMMDD

20140609 - Samir - Example

• Input - /SingleMu/Run2012C-PromptReco-v1/AOD
• Job Manager system - Condor
• Cluster - T3-Higgs

Jobs ran smoothly, not a single failure observed, worked fine

20140605 - Samir - Example2

• Input - /store/user/samir/myDataset - T2_US_Caltech
• Job Manager system - CRAB @ T2_CH_CERN

Had 2 failures. Retried the jobs and it worked

20140605 - Samir - Example2

• Input - /store/user/samir/myDataset - T2_US_Caltech
• Batch system - Condor @ T3-Higgs

Massive job failure. Looks like dataset was not found. Submitted a ticket to T2 support : https://its.cern.ch/jira/browse/CALTECHCMS-102

-- Main.samir - 2014-06-09