Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-32228

Event when timeout is reach should be customizable

XMLWordPrintable

      I would like to edit the event executed when timeout is reach.

      In some scenario, the event "Send interrupt signal to process" won't kill the running process. If I could catch some event, I could force the process to stop and for instance rollback environment.

      timeout(5, unit:'MINUTES') {
            // 10 minutes process
} catch {
            // specific timeout signal
}

          [JENKINS-32228] Event when timeout is reach should be customizable

          Jesse Glick added a comment -

          In some scenario, the event "Send interrupt signal to process" won't kill the running process.

          I would argue that this is more the point: if DurableTaskStep.Execution.stop does not actually result in termination after some reasonable amount of time, perhaps SIGKILL should be sent and the step exit.

          Additionally/alternately, perhaps timeout should give its body a grace period to exit cleanly, after which running steps and/or CPS-transformed Groovy code and/or native code should be terminated, à la WorkflowRun.doTerm. Cf. JENKINS-32986.

          Jesse Glick added a comment - In some scenario, the event "Send interrupt signal to process" won't kill the running process. I would argue that this is more the point: if DurableTaskStep.Execution.stop does not actually result in termination after some reasonable amount of time, perhaps SIGKILL should be sent and the step exit. Additionally/alternately, perhaps timeout should give its body a grace period to exit cleanly, after which running steps and/or CPS-transformed Groovy code and/or native code should be terminated, à la WorkflowRun.doTerm . Cf. JENKINS-32986 .

          Jesse Glick added a comment -

          if DurableTaskStep.Execution.stop does not actually result in termination after some reasonable amount of time, perhaps SIGKILL should be sent and the step exit.

          As of JENKINS-38769, the step will exit.

          perhaps timeout should give its body a grace period to exit cleanly, after which running steps and/or CPS-transformed Groovy code and/or native code should be terminated

          As of JENKINS-39072, running steps should be stopped. As of JENKINS-25623, CPS-transformed Groovy code will be too. Native code is another case, tracked in JENKINS-32986.

          Jesse Glick added a comment - if DurableTaskStep.Execution.stop does not actually result in termination after some reasonable amount of time, perhaps SIGKILL should be sent and the step exit. As of JENKINS-38769 , the step will exit. perhaps timeout should give its body a grace period to exit cleanly, after which running steps and/or CPS-transformed Groovy code and/or native code should be terminated As of JENKINS-39072 , running steps should be stopped. As of JENKINS-25623 , CPS-transformed Groovy code will be too. Native code is another case, tracked in JENKINS-32986 .

            jglick Jesse Glick
            gtheraud Guillaume Théraud
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated:
              Resolved: