Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-57888

Custom log format broken by 2.177

    Details

    • Similar Issues:
    • Released As:
      Jenkins 2.218

      Description

      Upon upgrading from 2.176 to 2.177, it would appear that support for providing a custom log format has been broken, due to what looks like an undocumented change (https://github.com/jenkinsci/winstone/pull/63) in the Winstone library. 

      Logging.properties file:

      .level=ALL
      handlers=java.util.logging.ConsoleHandler
      
      java.util.logging.SimpleFormatter.format=[%1$tF %1$tT.%1$tL][%4$s][%2$s] %5$s %6$s%n 
      java.util.logging.ConsoleHandler.level=INFO java.util.logging.ConsoleHandler.formatter=java.util.logging.SimpleFormatter

      I've also added the following JVM argument:

      -Djava.util.logging.config.file=/var/lib/jenkins/logging.properties
      

       

      Log format with 2.176:

      [2019-06-06 15:10:07.881][INFO][jenkins.InitReactorRunner$1 onAttained] Completed initialization
      [2019-06-06 15:10:08.052][INFO][hudson.WebAppMain$3 run] Jenkins is fully up and running
      

       

      Log format with 2.177:

      2019-06-06 15:11:40.956+0000 [id=29] INFO jenkins.InitReactorRunner$1#onAttained: Completed initialization 
      2019-06-06 15:11:41.136+0000 [id=20] INFO hudson.WebAppMain$3#run: Jenkins is fully up and running
      

        Attachments

          Issue Links

            Activity

            Hide
            l_r Louis-Rémi Paquet added a comment -

            Is there a workaround to be able to provide a ConsoleHandler formatter?

            If not, I'll be submitting a PR removing the overriding of it.

            It will still be possible to set the formatter from the command line or the logging.properties as usual.

            What do you think?

            Show
            l_r Louis-Rémi Paquet added a comment - Is there a workaround to be able to provide a ConsoleHandler formatter? If not, I'll be submitting a PR removing the overriding of it. It will still be possible to set the formatter from the command line or the logging.properties as usual. What do you think?
            Hide
            jglick Jesse Glick added a comment -

            The new formatter is deliberate. There could be a Winstone option to select an alternate formatter if someone cares to write it, or perhaps just by disabling the new formatter when java.util.logging.config.file is defined.

            Show
            jglick Jesse Glick added a comment - The new formatter is deliberate. There could be a Winstone option to select an alternate formatter if someone cares to write it, or perhaps just by disabling the new formatter when java.util.logging.config.file is defined.
            Hide
            l_r Louis-Rémi Paquet added a comment -

            Yes, that would be possible.

            I'm wondering, though, why the standard java behavior is overridden. It's not something one would expect.

            Show
            l_r Louis-Rémi Paquet added a comment - Yes, that would be possible. I'm wondering, though, why the standard java behavior is overridden. It's not something one would expect.
            Show
            oleg_nenashev Oleg Nenashev added a comment - https://github.com/jenkinsci/jenkins/pull/4452
            Hide
            oleg_nenashev Oleg Nenashev added a comment -

            It was released in 2.218, thanks to Louis-Rémi Paquet! I will mark it as LTS candidate, but it would require Winstone backporting due to the JENKINS-60857 regressions in 5.6. If Oliver Gondža decides it is important enough, I will do such backporting. 

            Show
            oleg_nenashev Oleg Nenashev added a comment - It was released in 2.218, thanks to Louis-Rémi Paquet ! I will mark it as LTS candidate, but it would require Winstone backporting due to the JENKINS-60857 regressions in 5.6. If Oliver Gondža decides it is important enough, I will do such backporting. 
            Hide
            olivergondza Oliver Gondža added a comment -

            This will be reverted in 2.204.5 as it has caused regressions much more impactfull.

            Show
            olivergondza Oliver Gondža added a comment - This will be reverted in 2.204.5 as it has caused regressions much more impactfull.
            Hide
            jglick Jesse Glick added a comment -

            To be clear, it was not this fix which caused the regressions, but unrelated changes in winstone.

            Show
            jglick Jesse Glick added a comment - To be clear, it was not this fix which caused the regressions, but unrelated changes in winstone .

              People

              • Assignee:
                l_r Louis-Rémi Paquet
                Reporter:
                fatmcgav Gavin Williams
              • Votes:
                3 Vote for this issue
                Watchers:
                8 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: