Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-10623

Mismatching dataId in logger output

    Details

    • Type: Bug
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: log, pipe_base, pipe_tasks
    • Labels:
      None

      Description

      When running processCcd.py on multiple visits, self-contradictory outputs are generated by the logger. For example:

      INFO  2017-05-25T00:07:04.004 processCcd (DataId(initialdata={'visit': 242618039}, tag=set([])))(processCcd.py:181)- Processing DataId(initialdata={'visit': 242619360}, tag=set([]))
      

      and

      FATAL 2017-05-25T00:07:04.004 processCcd (DataId(initialdata={'visit': 242713680}, tag=set([])))(cmdLineTask.py:351)- Failed on dataId=DataId(initialdata={'visit': 242715702}, tag=set([])): Unable to match sources
      

      The problem above is that the visit numbers don't match, but there should be no way for this to ever be the case, as far as I can see. FWIW, the second number is the true visit number, at least in my case.

      This is only noticeable when --longlog is used, to allow disentangling of parallel processed data, but I don't know whether that means the logger is also wrong when sequentially running multiple visits on a single core (I suspect it's OK in that instance, but I'm not certain).

      This bug is bad for tracking down which visits are failing and why.

        Attachments

          Issue Links

            Activity

            Hide
            mfisherlevine Merlin Fisher-Levine added a comment -

            Other than the one comment I made this looks fine. I haven't tested the fix as that requires rebuilding a whole stack, but assuming it fixes the problem then this is good to merge

            Show
            mfisherlevine Merlin Fisher-Levine added a comment - Other than the one comment I made this looks fine. I haven't tested the fix as that requires rebuilding a whole stack, but assuming it fixes the problem then this is good to merge
            Hide
            fritzm Fritz Mueller added a comment -

            You could have had five points of vacation out of that one, Andy! (Thanks, Andy)

            Show
            fritzm Fritz Mueller added a comment - You could have had five points of vacation out of that one, Andy! (Thanks, Andy)
            Hide
            salnikov Andy Salnikov added a comment -

            Thanks for review! Merged and done with a minimum amount of points while also helping poor HEP guys at CERN

            Show
            salnikov Andy Salnikov added a comment - Thanks for review! Merged and done with a minimum amount of points while also helping poor HEP guys at CERN
            Hide
            salnikov Andy Salnikov added a comment - - edited

            One note for Merlin Fisher-Levine - testing does not require rebuilding whole stack, patches are trivial and binary compatible (in case of log). You can just clone log and pipe_base (latter is optional), build them locally, do setup -k -r ... and it will work. I tested it that way to make sure that it actually fixes output.

            Show
            salnikov Andy Salnikov added a comment - - edited One note for Merlin Fisher-Levine - testing does not require rebuilding whole stack, patches are trivial and binary compatible (in case of log ). You can just clone log and pipe_base (latter is optional), build them locally, do setup -k -r ... and it will work. I tested it that way to make sure that it actually fixes output.
            Hide
            mfisherlevine Merlin Fisher-Levine added a comment -

            Oh, that's good to know, thanks. I never touch the C++ layer, so thought that anything like that automatically required a rebuild from there on up (down?). I will do that soon then, just so that I can have the bug-fix (though presumably this will go into the next weekly, so I might not bother given that it's already merged...)

            Show
            mfisherlevine Merlin Fisher-Levine added a comment - Oh, that's good to know, thanks. I never touch the C++ layer, so thought that anything like that automatically required a rebuild from there on up (down?). I will do that soon then, just so that I can have the bug-fix (though presumably this will go into the next weekly, so I might not bother given that it's already merged...)

              People

              • Assignee:
                salnikov Andy Salnikov
                Reporter:
                mfisherlevine Merlin Fisher-Levine
                Reviewers:
                Merlin Fisher-Levine
                Watchers:
                Andy Salnikov, Fritz Mueller, Hsin-Fang Chiang, John Swinbank, Kian-Tat Lim, Merlin Fisher-Levine, Paul Price, Robert Lupton
              • Votes:
                0 Vote for this issue
                Watchers:
                8 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved:

                  Summary Panel