Backport #4554 to 10_2 gcc7

[mrodozov]

Backport requested: #4554 (comment)

[cmsbuild]

A new Pull Request was created by @mrodozov (Mircho Rodozov) for branch IB/CMSSW_10_2_X/gcc700.

@cmsbuild, @smuzaffar, @gudrutis, @mrodozov can you please review it and eventually sign? Thanks.
You can sign-off by replying to this message having '+1' in the first line of your reply.
You can reject by replying to this message having '-1' in the first line of your reply.

[mrodozov]

@fabiocos you start it from the corresponding cmssw PR because I don't know which one it is

[fabiocos]

@mrodozov in 10_4_X we have cms-sw/cmssw#25016 using #4554 . Now in order to let in 10_2_X cms-sw/cmssw#25386 working properly we need the backport of that data file for 10_2_X. The release tests should not crash as this code is not used in regular workflows, but still this is needed for use in analysis

[fabiocos]

@mbluj FYI

[mbluj]

Hello,
I do not see this in 10_2_9 which was released earlier today (14/12/2018): https://github.com/cms-sw/cmssw/releases/CMSSW_10_2_9. Is it expected?

[fabiocos]

@mbluj release-notes show #4579 in the external updates list, do you see missing files?

[mbluj]

Correct, I've overlooked it in release notes. But, I have tried to run DNN-tau-Ids with 10_2_9 at lxplus, but cmsRun failed due to missing files. And, indeed there is not RecoTauTag/TrainingFiles/data/ in /cvmfs/cms.cern.ch/slc6_amd64_gcc700/cms/cmssw/CMSSW_10_2_9/external/slc6_amd64_gcc700/data.

[fabiocos]

type bug-fix

[fabiocos]

sorry, not intended for this PR. Anyway it looks like there is something to be understood here, @mrodozov could you please cross check?

[mbluj]

I have also checked that the files are not present also neither in 10_4_0_pre4 nor in the most recent 10_2_X and 10_4_X IBs (dated 2018-12-14-1100).

[fabiocos]

I cross checked and noticed the same thing...

[smuzaffar]

@fabiocos , of course it will not be part of release as RecoTauTag-TrainingFiles dependency was not added in cmsswdata.

[smuzaffar]

@fabiocos , @mbluj sorry about this. Our fault, we missed the dependency in cmsswdata.

[smuzaffar]

#4590 should fix it.

[fabiocos]

@smuzaffar thanks, as this code is not running in official workflows (@mbluj may confirm) this should not be a big problem in the short term, but of course we need to fix it

[mbluj]

@smuzaffar thanks, as this code is not running in official workflows (@mbluj may confirm) this should not be a big problem in the short term, but of course we need to fix it

Yes, I confirm that the code is not run in official workflows and it is why the issue was not detected. But, obviously users will be unhappy that the files are missed in current releases.