• Icon: Bug Bug
    • Resolution: Fixed
    • Icon: Major Major
    • core
    • Ubuntu 7.10 running on 32bit dual core. Jenkins with Winstone. Java version "1.6.0_03", Java(TM) SE Runtime Environment (build 1.6.0_03-b05), Java HotSpot(TM) Server VM (build 1.6.0_03-b05, mixed mode)

      I am not sure I have filed this under the right component.

      We are seeing steady gradual increases in memory usage for our Jenkins (see attached graphs). It does not result in the OutOfMemoryError but usually results in Jenkins utilizing 100% of CPU (one core in our case).

      Attached is a memory histogram report and the link include (see URL above) is to a heap dump generated using the Monitoring Plugin.

      This happens consistently and our work around for now is to restart Jenkins each time it gets too bloated. If this doesn't get fixed, we will have to stop using Winstone and use Jetty - because there seems to be a connection between the servlet wrapper (winstone/jetty) and the leak - pretty sure we did not see this when using Jetty.

          [JENKINS-14473] Jenkins server memory leak

          Did some analysis using Eclipse's Memory Analyzer Tool with the heap dump attached. Report zip files and a PDF attached.

          Anoop Karollil added a comment - Did some analysis using Eclipse's Memory Analyzer Tool with the heap dump attached. Report zip files and a PDF attached.

          Liya Katz added a comment -

          Upgrading to 1.492 (from 1.473) has solved this problem for us.

          Liya Katz added a comment - Upgrading to 1.492 (from 1.473) has solved this problem for us.

          Thanks Liya. I did see a few weeks ago (memory usage crept up to 5gigs of RAM), but things have been okay the past week with 1.491 - upgraded to 1.492 anyway. Hopefully this has been resolved - I will close this bug once I am sure.

          Anoop Karollil added a comment - Thanks Liya. I did see a few weeks ago (memory usage crept up to 5gigs of RAM), but things have been okay the past week with 1.491 - upgraded to 1.492 anyway. Hopefully this has been resolved - I will close this bug once I am sure.

          evernat added a comment -

          @Anoop
          Any news?
          Is it reproduced recently?

          evernat added a comment - @Anoop Any news? Is it reproduced recently?

          No, I think this has been fixed.I do not see it in v1.538

          Anoop Karollil added a comment - No, I think this has been fixed.I do not see it in v1.538

            Unassigned Unassigned
            akarollil Anoop Karollil
            Votes:
            2 Vote for this issue
            Watchers:
            8 Start watching this issue

              Created:
              Updated:
              Resolved: