10 Replies Latest reply on Oct 27, 2013 7:26 AM by LANDave

    policy tasks: some machines stay active

    Apprentice

      Hi all,

       

      since i've updated our core to 8.8 sp4 i've some trouble with policytasks:

       

      Some machines (about 30%) stay active even if the task successfully finished on the client. The log says:

       

      Tue, 17 Aug 2010 08:05:04 [Start Execute]
      Tue, 17 Aug 2010 08:05:04 + begin SDClientTask.CORESERVER.4283
      Tue, 17 Aug 2010 08:05:04   Status 229392385: Starting task, if nothing goes wrong we will treat this as successful.
      Tue, 17 Aug 2010 08:05:04   Setting the active operation to SDClientTask.CORESERVER.4283.xml
      Tue, 17 Aug 2010 08:05:05   + begin Task.CORESERVER.4283
      Tue, 17 Aug 2010 08:05:05     + begin DeliveryMethod.CORESERVER.184
      Tue, 17 Aug 2010 08:05:05       + begin Policy.Required.PortalRemoval
      Tue, 17 Aug 2010 08:05:05       - end
      Tue, 17 Aug 2010 08:05:05     - end
      Tue, 17 Aug 2010 08:05:05     + begin SWD.ExecuteOnly.Proxy.461
      Tue, 17 Aug 2010 08:05:05       + begin SWD.MultipleTaskProtection.CORESERVER.4283
      Tue, 17 Aug 2010 08:05:05         Policy is configured to always run.
      Tue, 17 Aug 2010 08:05:05         + begin ActiveNotification.Start
      Tue, 17 Aug 2010 08:05:05         - end
      Tue, 17 Aug 2010 08:05:05         + begin SWD.Package.Files.CORESERVER.461
      Tue, 17 Aug 2010 08:05:18           + begin SWD.Package.CORESERVER.461
      Tue, 17 Aug 2010 08:05:22             Status 229392397: Starting package deployment.

       

      That's it. The [STOP] "tag" is missing.

       

      Other machines get status "done" as expected - the log looks like:


      Tue, 17 Aug 2010 06:48:52 [Start Execute]
      Tue, 17 Aug 2010 06:48:52 + begin SDClientTask.CORESERVER.4283
      Tue, 17 Aug 2010 06:48:52   Status 229392385: Starting task, if nothing goes wrong we will treat this as successful.
      Tue, 17 Aug 2010 06:48:52   Setting the active operation to SDClientTask.CORESERVER.4283.xml
      Tue, 17 Aug 2010 06:48:52   + begin Task.CORESERVER.4283
      Tue, 17 Aug 2010 06:48:52     + begin DeliveryMethod.CORESERVER.184
      Tue, 17 Aug 2010 06:48:52       + begin Policy.Required.PortalRemoval
      Tue, 17 Aug 2010 06:48:52       - end
      Tue, 17 Aug 2010 06:48:52     - end
      Tue, 17 Aug 2010 06:48:52     + begin SWD.ExecuteOnly.Proxy.461
      Tue, 17 Aug 2010 06:48:52       + begin SWD.MultipleTaskProtection.CORESERVER.4283
      Tue, 17 Aug 2010 06:48:52         Policy is configured to always run.
      Tue, 17 Aug 2010 06:48:52         + begin ActiveNotification.Start
      Tue, 17 Aug 2010 06:48:52         - end
      Tue, 17 Aug 2010 06:48:52         + begin SWD.Package.Files.CORESERVER.461
      Tue, 17 Aug 2010 06:49:03           + begin SWD.Package.CORESERVER.461
      Tue, 17 Aug 2010 06:49:04             Status 229392397: Starting package deployment.
      Tue, 17 Aug 2010 07:04:55             Status 0: Package deployment complete.
      Tue, 17 Aug 2010 07:04:55             + begin Reboot.RebootHandler.CORESERVER.461
      Tue, 17 Aug 2010 07:04:55             - end
      Tue, 17 Aug 2010 07:04:55           - end
      Tue, 17 Aug 2010 07:04:55         - end
      Tue, 17 Aug 2010 07:04:55         + begin Policy.Portal.CompletePortal.184
      Tue, 17 Aug 2010 07:04:55         - end
      Tue, 17 Aug 2010 07:04:55         + begin ActiveNotification.Stop
      Tue, 17 Aug 2010 07:04:55         - end
      Tue, 17 Aug 2010 07:04:55       - end
      Tue, 17 Aug 2010 07:04:55     - end
      Tue, 17 Aug 2010 07:04:55   - end
      Tue, 17 Aug 2010 07:04:55   Reporting complete status of 0 for task 4283.
      Tue, 17 Aug 2010 07:04:55   SDClient task complete
      Tue, 17 Aug 2010 07:04:55 - end
      Tue, 17 Aug 2010 07:04:55 [STOP]

       

       

       

      What i've done:

      - several reboots of the coreserver (after applying sp4)

      - restarting scheduler service

      - restarting the policytasks

      - deleting and rebuilding the tasks

       

      Thanks for your help!

      Dirk

        • 1. Re: policy tasks: some machines stay active
          Specialist

          I would like to add that with LDMS 9 SP1 I am seeing a similar thing. I do have a stop tag, but a good bit of my policies have machines that are stuck in Active and have some that are even in pending with this problem. The software installs on the client but the status is not getting updated. Makes it impossible to track any kind of success rate and report it up to managers. Really frustrating.

          • 2. Re: policy tasks: some machines stay active
            Specialist

            Seems my task logs do not complete after STOP. The cases where there is a machine stuck in active most logs have the next policy's task IDN logging that the machine is getting the XMLFilename, then processing, as the last few lines of the log.

             

            Tue, 10 Aug 2010 18:14:46 [STOP]
            Tue, 10 Aug 2010 18:14:46 ProcessInvokerStateRecords() received status complete.
            Tue, 10 Aug 2010 18:14:46 GetXMLFileName() returning SDClientTask.XXXXXXXXX.1794.xml (XXXXXXXXXXXXX)
            Tue, 10 Aug 2010 18:14:46 ProcessInvokerStateRecords() ready to process file: C:\Documents and Settings\All Users\Application Data\LANDesk\ManagementSuite\landesk\files\Active.Task.xml (operation 4)

            • 3. Re: policy tasks: some machines stay active
              Apprentice

              i guess it happens, because vulscan.exe on the client updates itself (new vulscan.exe on core). while/after selfupdating the task somehow is unable to send its status to the core.

               

              is this possible?

              • 4. Re: policy tasks: some machines stay active
                Specialist

                I have checked around on some of my machines that are stuck in active, and there has been no instances of Vulscan run around the times that the software, stuck in active though complete, has been installed.

                • 5. Re: policy tasks: some machines stay active
                  Apprentice

                  I'm also seeing this on my 9.0SP1 core. Worse than it was with 8.8.

                  • 6. Re: policy tasks: some machines stay active
                    Rookie

                    I am seeing similar problem with V9 SP2 on core server.

                    • 7. Re: policy tasks: some machines stay active
                      Apprentice

                      Hi all,

                       

                      we have the simular problem. It is/was caused, because we were distributing a repair task of an patch group via policy.

                      The Vulscan rebooted the computer. During this shutdown the policy.client.invoker try to send the task status back, BUT the mashine isn't able to get a network connect during shutdown. The main prob. is, that the policy.client.invoker can't write his local alert file to send the task staus after reboot. Thats why this policy stuck for ever on active

                       

                      We still wait for the developer feedback from the US.

                       

                      We can't solve this issue (this have to be don by LANDesk...) but I have a recommendation to correct it on the clients within seconds. We are using a tool from a LANDesk partner to deal and rerest local policies remotly. It even help us a lot doing trouble shooting - access log files with just one click... If you like, have a look at the program "Policy Client Viewer"

                      • 8. Re: policy tasks: some machines stay active
                        Apprentice

                        Hallo Butti,

                         

                        I'm doing this via a batchfile with two vulscan calls. The first one does all the patching without a reboot and the second one is fired with "start" at the end of the batch. Works great because the package has enough time to finish.

                         

                        looks like this:

                         

                        "%LD_CLIENT_DIR%\vulscan.exe" /agentbehavior=61 /nosync

                        start "reboot" "%LD_CLIENT_DIR%\vulscan.exe" /rebootIfNeeded /agentbehavior=241 /nosync

                         

                        agentbehavior 61 scans all vulns and does autofix but no reboot

                        agentbehavior 241 is just for the reboot (deferrable reboot, auto reboot if no user logged on, delayable, etcpp)

                         

                        Maybe this helps ...?

                         

                        Viele Grüße,

                        Dirk

                         

                        PS: Your name sounds familiar ... have you used SLD in the past?

                        • 9. Re: policy tasks: some machines stay active
                          Rookie

                          We are seeing the same issue when deploying a self contained client installation package via a policy.  The task is completing but the status never changes.  Hopefully their is a fix for this bug out there.  The current status isn't very helpful.

                          • 10. Re: policy tasks: some machines stay active
                            LANDave SupportEmployee

                            Make sure that the user that the Scheduler service is set to log on as is also a member of the Core Server's local administrators group.