convert mkfit output warnings to LogInfo

[slava77]

this is in a follow up to #35798

the rate of warnings about bad candidates in the MkFitOutputConverter is too large to handle in production logs.
This PR silences them for now.

136.874 log example

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-35802/26166

• This PR adds an extra 16KB to repository

[cmsbuild]

A new Pull Request was created by @slava77 (Slava Krutelyov) for master.

It involves the following packages:

• RecoTracker/MkFit (reconstruction)

@jpata, @cmsbuild, @slava77 can you please review it and eventually sign? Thanks.
@makortel, @felicepantaleo, @GiacomoSguazzoni, @JanFSchulte, @rovere, @VinInn, @ebrondol, @gpetruc, @mmusich, @mtosi, @dgulhan this is something you requested to watch as well.
@perrotta, @dpiparo, @qliphy you are the release manager for this.

cms-bot commands are listed here

[slava77]

@cmsbuild please test

[slava77]

@mmusich
please clarify if this is an acceptable solution to #35798 for now.
Thank you.

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-b45b31/19864/summary.html
COMMIT: 724a53f
CMSSW: CMSSW_12_1_X_2021-10-23-1100/slc7_amd64_gcc900
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/35802/19864/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 48 differences found in the comparisons
• DQMHistoTests: Total files compared: 40
• DQMHistoTests: Total histograms compared: 2751113
• DQMHistoTests: Total failures: 10
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 2751081
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 39 files compared)
• Checked 170 log files, 37 edm output root files, 40 DQM output files
• TriggerResults: no differences found

[slava77]

• 136.874 step3 log baseline is 1.7M
• 136.874 step3 log this PR is 456K

This is a quite a significant reduction, but there are still quite a few warnings coming indirectly via MkFitOutputConverter. I'll dig around a bit more.

[slava77]

This is a quite a significant reduction, but there are still quite a few warnings coming indirectly via MkFitOutputConverter. I'll dig around a bit more.

at least the first warning comes after a propagation of the mkfit track output state to the last hit.
IIUC, this means that the earlier test of the positive-definiteness was incomplete (recall that CurvilinearTrajectoryError::posDef is checking only the diagonal elements).

So, this PR is as good as it gets at this point on the CMSSW side.

@mmasciov @makortel @osschar

[mmusich]

@slava77

please clarify if this is an acceptable solution to #35798 for now.

Based on your comment #35802 (comment) I think so, though I'd like to understand if we should (could?) plan for an update of the mkFit external within the time of closure of 12_1_0 for further clean-up.
Thanks

[slava77]

Based on your comment #35802 (comment) I think so, though I'd like to understand if we should (could?) plan for an update of the mkFit external within the time of closure of 12_1_0 for further clean-up.

I opened trackreco/mkFit#373

[slava77]

+reconstruction

for #35802 724a53f

• jenkins tests pass and comparisons with the baseline show differences only in the messageLogger counts e.g. in 136.874
  

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @perrotta, @dpiparo, @qliphy (and backports should be raised in the release meeting by the corresponding L2)

[perrotta]

+1