Skip to content
Snippets Groups Projects

Disable Trigger in RecExCommon unless really needed

Merged Rafal Bielski requested to merge rbielski/athena:recexcommon-no-trig into master

The whole "flags saga" triggered by !38449 (merged) (see Issue Links in ATR-22360) uncovered that some jobs were trying to run Trigger selection algorithms when they clearly shouldn't have. Not only they were trying to run something they don't need, but also something that is deprecated (legacy trigger) and on top of that was misconfigured in these jobs.

The two cases reported in ATR-22516 and ATR-22515 were jobs starting from RDO without any Trigger content and intending to run RAWtoESD without Trigger. Instead of what was intended, these jobs were actually configuring Trigger like if they were legacy RDOtoRDOTrigger jobs rather than RAWtoESD. Another similar case was reported in ATLASRECTS-5833.

Fix this in the following way:

  1. Set recAlgs.doTrigger to False by default. Add a "TODO" comment saying it should be removed once we retire legacy Trigger in r22.
  2. Disable rec.doTrigger if running on POOL file without any Trigger content (like RDO) while also recAlgs.doTrigger=False. This fixes the two tickets mentioned above.
  3. Move the ESDtoAOD Run-3 Trigger setup in RecExCommon_topOptions.py under if rec.doTrigger: and remove Run-3 Trigger special case from ESDtoAOD transform skeleton. (ATR-19982)
  4. Fix the fallout: clean up a bunch of cases in Trigger configuration which had either explicit or implicit assumption of recAlgs.doTrigger.

Multiple local tests succeeded, but running full-unit-tests on this one as the RecExCommon configuration is clearly very fragile and the number of workflows affected by !38449 (merged) was slightly beyond my expectations.

Edited by Rafal Bielski

Merge request reports

Loading
Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • :negative_squared_cross_mark: CI Result FAILURE (hash e2bc0459)

    Athena AthSimulation AthGeneration AnalysisBase
    externals :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:
    cmake :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:
    make :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:
    required tests :o: :white_check_mark: :white_check_mark: :white_check_mark:
    optional tests :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:

    Full details available on this CI monitor view
    :white_check_mark: Athena: number of compilation errors 0, warnings 0
    :white_check_mark: AthSimulation: number of compilation errors 0, warnings 0
    :white_check_mark: AthGeneration: number of compilation errors 0, warnings 0
    :white_check_mark: AnalysisBase: number of compilation errors 0, warnings 0
    :pencil: For experts only: Jenkins output [CI-MERGE-REQUEST-CC7 25068]

    • Resolved by Rafal Bielski

      Hi @rbielski,

      thanks for this MR. Just for me to understand the timescale on this: Should we go ahead updating the Run3 reco CI including doTrigger=False such that we can update the CI already now (without waiting for this MR) or should we wait with the CI update for this MR going in such that we can run with autoConfiguration only? I think it basically depends on when/if you have time to look into this MR...

      Thanks, Nico

  • Rafal Bielski marked as a Work In Progress

    marked as a Work In Progress

  • Rafal Bielski changed the description

    changed the description

  • Rafal Bielski changed the description

    changed the description

  • Rafal Bielski added 324 commits

    added 324 commits

    • e2bc0459...eb12b181 - 322 commits from branch atlas:master
    • 27e5535b - Disable Trigger in RecExCommon unless really needed
    • 7aef499d - Clean up Trigger configuration from incorrect assumptions of recAlgs.doTrigger

    Compare with previous version

  • Rafal Bielski resolved all threads

    resolved all threads

  • Jenkins please retry a build

  • This merge request affects 5 packages:

    • Reconstruction/RecExample/RecExCommon
    • Reconstruction/RecExample/RecExConfig
    • Trigger/TrigAlgorithms/TrigT2CaloCommon
    • Trigger/TrigMonitoring/TrigCaloMonitoring
    • Trigger/TriggerCommon/TriggerJobOpts

    Affected files list will not be printed in this case

    Adding @goetz ,@joheinri ,@damazio as watchers

  • ATLAS Robot added DQ label and removed review-user-action-required label

    added DQ label and removed review-user-action-required label

  • :negative_squared_cross_mark: CI Result FAILURE (hash 7aef499d)

    Athena AthSimulation AthGeneration AnalysisBase
    externals :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:
    cmake :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:
    make :warning: :white_check_mark: :white_check_mark: :white_check_mark:
    required tests :o: :white_check_mark: :white_check_mark: :white_check_mark:
    optional tests :white_check_mark: :white_check_mark: :white_check_mark: :white_check_mark:

    Full details available on this CI monitor view
    :warning: Athena: number of compilation errors 0, warnings 1
    :white_check_mark: AthSimulation: number of compilation errors 0, warnings 0
    :white_check_mark: AthGeneration: number of compilation errors 0, warnings 0
    :white_check_mark: AnalysisBase: number of compilation errors 0, warnings 0
    :pencil: For experts only: Jenkins output [CI-MERGE-REQUEST-CC7 25187]

  • Rafal Bielski added 51 commits

    added 51 commits

    • 7aef499d...198c1e90 - 49 commits from branch atlas:master
    • 12dac682 - Disable Trigger in RecExCommon unless really needed
    • 3694d649 - Clean up Trigger configuration from incorrect assumptions of recAlgs.doTrigger

    Compare with previous version

  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Loading
  • Please register or sign in to reply
    Loading