Possibility of runTheMatrix to start from scratch when there is no dataset input (from DAS)

srimanob commented 3 years ago

From the core sw meeting (2 March), there was a discussion to try to run from scratch if there is no dataset input. Not sure if I understand the full idea on how to deal with the situation.

In the runTheMatrix, we may start from https://github.com/cms-sw/cmssw/blob/master/Configuration/PyReleaseValidation/python/WorkFlowRunner.py#L86-L99

Line 86 gives the cmsDriver for all steps and if the command is to run from scratch, the step-0 is the driver of GS step, i.e. (0, 'cmsDriver.py TTbar_14TeV_TuneCP5_cfi --conditions auto:phase1_2021_realistic -n 10 --era Run3 --eventcontent FEVTDEBUG --relval 9000,100 -s GEN,SIM --datatier GEN-SIM --beamspot Run3RoundOptics25ns13TeVLowSigmaZ --geometry DB:Extended --io TTbar_14TeV_TuneCP5_2021_GenSim.io --python TTbar_14TeV_TuneCP5_2021_GenSim.py')

Line 99 then checks if the command is a string (start from scratch), or not (with INPUT dataset, using DAS).

As all the workflow will pass here, it is maybe a possibility to check here, and forcing to run-from-scratch if DAS fail to return dataset. However, if input is set since the beginning, we will not have cmsDriver-from-scratch.

In addition, we may need to consider also few additions including

if fail on data, we do nothing
Trying to force to run from scratch, only when an option is set. So that, it will not try to do by default (i.e. when PdmV tries to use input datasets, but DAS fail at that moment.

FYI @smuzaffar @kpedro88 @davidlange6

cmsbuild commented 3 years ago

A new Issue was created by @srimanob Phat Srimanobhas.

@Dr15Jones, @dpiparo, @silviodonato, @smuzaffar, @makortel, @qliphy can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

makortel commented 3 years ago

assign pdmv, upgrade

cmsbuild commented 3 years ago

New categories assigned: pdmv,upgrade

@chayanit,@wajidalikhan,@jordan-martins,@kpedro88,@srimanob you have been requested to review this Pull request/Issue and eventually sign? Thanks

srimanob commented 3 years ago

Hi @smuzaffar @kpedro88 Should I try to move on, or you already have ideas on how to deal with it? Thanks.

smuzaffar commented 3 years ago

@srimanob , @kpedro88 suggested something like the following ( which we are doing for upgrade wfs).

set to empty by default: https://github.com/cms-sw/cmssw/blob/a96f16e7dba2bb196cf410b9b40a824fbc32a220/Configuration/PyReleaseValidation/python/upgradeWorkflowComponents.py#L1246-L1251
filled in only when relvals are available: https://github.com/cms-sw/cmssw/blob/a96f16e7dba2bb196cf410b9b40a824fbc32a220/Configuration/PyReleaseValidation/python/relval_steps.py#L3223-L3232
creates an INPUT step if datasets are available: https://github.com/cms-sw/cmssw/blob/a96f16e7dba2bb196cf410b9b40a824fbc32a220/Configuration/PyReleaseValidation/python/relval_steps.py#L3466-L3468
that step is actually used in https://github.com/cms-sw/cmssw/blob/master/Configuration/PyReleaseValidation/python/MatrixReader.py (if i recall correctly)

smuzaffar commented 3 years ago

@srimanob I have not yet looked in to it, if you any ida then please go ahead. We just need to make sure that if data set is available but das failed then we should also fail instead of forcing a gen-sim

cms-sw / cmssw

Possibility of runTheMatrix to start from scratch when there is no dataset input (from DAS) #33042