Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-27371

Parent builds sometimes hang on successful child builds of same type

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed (View Workflow)
    • Priority: Major
    • Resolution: Postponed
    • Component/s: multijob-plugin
    • Labels:
      None
    • Environment:
      jenkins 1.593
      multijob plugin 1.16
    • Similar Issues:

      Description

      With a job configuration like:

      parent_job:
      child_job
      child_job

      occasionally we'll get output that looks like this:

      Starting build job child_job.
      Starting build job child_job.
      Finished Build : #123 - Job : child_job with status : SUCCESS
      <this hangs forever, so we abort the parent_job>
      Aborting all subjobs.
      Finished Build : #124 - Job : child_job with status : ABORTED

      However, job #124 will always have finished successfully. We see this relatively rarely (~5% of parent_job builds?). My guess is that there's a race condition causing this, since we usually see this during times of high load (ie many child_job instances are being started).

        Attachments

          Issue Links

            Activity

            simonmweber Simon Weber created issue -
            mcantin Mathieu Cantin made changes -
            Field Original Value New Value
            Link This issue duplicates JENKINS-26678 [ JENKINS-26678 ]
            mcantin Mathieu Cantin made changes -
            Link This issue is duplicated by JENKINS-28662 [ JENKINS-28662 ]
            mcantin Mathieu Cantin made changes -
            Link This issue is duplicated by JENKINS-29981 [ JENKINS-29981 ]
            rtyler R. Tyler Croy made changes -
            Workflow JNJira [ 161576 ] JNJira + In-Review [ 180750 ]
            dansirbu101 Dan Sirbu made changes -
            Comment [ I believe I have a similar issue.

            A log shows :

            Polling SCM changes on master
                >> Job status: [ECM DB] subjob has no changes since last build.
            FATAL: SCM polling aborted
            java.lang.InterruptedException
            at java.lang.Object.wait(Native Method)
            at java.lang.Object.wait(Unknown Source)
            at hudson.slaves.WorkspaceList.acquire(WorkspaceList.java:257)
            at hudson.slaves.WorkspaceList.acquire(WorkspaceList.java:236)
            at hudson.model.AbstractProject.pollWithWorkspace(AbstractProject.java:1475)
            at hudson.model.AbstractProject._poll(AbstractProject.java:1452)
            at hudson.model.AbstractProject.poll(AbstractProject.java:1363)
            at com.tikal.jenkins.plugins.multijob.MultiJobBuilder.getScmChange(MultiJobBuilder.java:190)
            at com.tikal.jenkins.plugins.multijob.MultiJobBuilder.perform(MultiJobBuilder.java:279)
            at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:20)
            at hudson.model.AbstractBuild$AbstractBuildExecution.perform(AbstractBuild.java:779)
            at hudson.model.Build$BuildExecution.build(Build.java:205)
            at hudson.model.Build$BuildExecution.doRun(Build.java:162)
            at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:534)
            at com.tikal.jenkins.plugins.multijob.MultiJobBuild$MultiJobRunnerImpl.run(MultiJobBuild.java:136)
            at hudson.model.Run.execute(Run.java:1729)
            at com.tikal.jenkins.plugins.multijob.MultiJobBuild.run(MultiJobBuild.java:73)
            at hudson.model.ResourceController.execute(ResourceController.java:98)
            at hudson.model.Executor.run(Executor.java:404)
                >> Job status: [ECM Setup] subjob has no changes since last build.
                >> Job status: [ECM JARs] subjob does not contain lastbuild.
            Starting build job ECM JARs.

            SCM polling is disabled for all subjobs including the multijob parent itself.

            What is interesting is that this happens just when switching from the multijob to subjob.

            In my case it is the multijob that check out the source code & the subjobs are inheriting the workspace.

            Maybe this info could help. It is 100% reproducible.
            ]
            cloudbees CloudBees Inc. made changes -
            Remote Link This issue links to "CloudBees Internal OSS-583 (Web Link)" [ 18867 ]
            owenmehegan Owen Mehegan made changes -
            Assignee Chen Cohen [ chenc ]
            yorammi Yoram Michaeli made changes -
            Status Open [ 1 ] Closed [ 6 ]
            Resolution Postponed [ 6 ]

              People

              • Assignee:
                chenc Chen Cohen
                Reporter:
                simonmweber Simon Weber
              • Votes:
                3 Vote for this issue
                Watchers:
                11 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: