• Icon: Bug Bug
    • Resolution: Fixed
    • Icon: Critical Critical
    • remoting
    • None
    • Platform: All, OS: All

      Seems like there is a regression between 1.136 and 1.143 where when build
      aborted on a slave the build process is not destroyed.

      I tried with plain clean installation one slave. The whole job is "sleep 1000"
      so I manually check with ps to see which processes are running. After the abort
      the initial bash process continues to run.
      Seems like sometimes (rarely) the process does get destroyed (at least on master).

      Anyway process always got destroyed with 1.136.

          [JENKINS-904] abort build regression

          akostadinov added a comment -

          FYI the issue is still present in current snapshot.

          akostadinov added a comment - FYI the issue is still present in current snapshot.

          Fixed in 1.148.

          To the best of my knowledge, this has always been the behavior of Hudson.

          Kohsuke Kawaguchi added a comment - Fixed in 1.148. To the best of my knowledge, this has always been the behavior of Hudson.

          akostadinov added a comment -

          By 1.136 the bash process was actually process.destroy-ed. I was testing a
          script that relied on the behavior but it stopped working when we upgraded to 1.143

          As well I think on master it is working.

          dvrzalik is working on a plug-in that calls process.destroy and it is not
          working on slaves (on master it works). Where if he uses an exec "kill <PID>"
          that works on slave as well on master.

          I'll try do some more investigation on that.

          akostadinov added a comment - By 1.136 the bash process was actually process.destroy-ed. I was testing a script that relied on the behavior but it stopped working when we upgraded to 1.143 As well I think on master it is working. dvrzalik is working on a plug-in that calls process.destroy and it is not working on slaves (on master it works). Where if he uses an exec "kill <PID>" that works on slave as well on master. I'll try do some more investigation on that.

          akostadinov added a comment -

          Don't know how that worked in 1.136 but now it does work.

          Thanks!

          akostadinov added a comment - Don't know how that worked in 1.136 but now it does work. Thanks!

            Unassigned Unassigned
            akostadinov akostadinov
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: