1 2 Previous Next 26 Replies Latest reply on Jan 23, 2013 5:05 AM by silver06 Go to original post
      • 15. Re: 7.1.3.Final feedbacks
        silver06

        Hi Emanuel

         

        I run all my host controllers (about 15) on virtual machines

        (VMWare) on SLES 11 (Suse Linux Enterprise Server) SP2 (Kernel 3.x)

        My DC is on a dedicated Vm based on the same OS

        Everything is on a LAN (same subnet) without any firewall nor IDS

        In Jboss 7.1.1 and EAP 6.0.0 everything was

        was working fine by increasing the timeout manually in the source code (moving from 30000ms to 90000ms) in 7.1.1 or by using system properties in eap 6.0.0.

        The same setup and infra doesnt work in eap 6.0.1 with xnio timeout issues. It's thus impossible to migrate our infra to it for the time being whereas we were looking forward to having the Ping/pong mechanism to fix registration issues that we have when restarting the DC

         

        Do you need other information?

        I have 10 simple servers in my host controller with associated jvm of 512m

        As Xmx.

         

        Let me know if you need network info as well.

        • 16. Re: 7.1.3.Final feedbacks
          silver06

          I forgot to mention that all my servers within my host.xml are linked to different server groups.

          So I've got 10 servers linked to 10 different server groups

          Don't know if it could come from there though

          • 17. Re: 7.1.3.Final feedbacks
            silver06

            Finally, with the Redhat support we have been able to reproduce the issue by having less powerful VM.

            The timeout issue appears when you have a machine (physical or virtual) not powerful enough to communicate with the domain controller in a specific timeframe.

            If I upgrade my virtual machine with more virtual CPU I don't have the timeout issue anymore.

             

            The expected behaviour for me now is to be able to configure the XNIO timeout in order to work with VM with less CPU.

            This is in the Redhat support's hands now

             

            Regards

            1 of 1 people found this helpful
            • 18. Re: 7.1.3.Final feedbacks
              emuckenhuber

              Ok, thanks. That is what i suspected originally. I will continue with my inital approach then and follow up with support to make sure that this really fixes this issue.

              • 19. Re: 7.1.3.Final feedbacks
                silver06

                OK thanks Emanuel

                When do you thing you will be able to provide this fix to the support ?

                • 20. Re: 7.1.3.Final feedbacks
                  emuckenhuber

                  I hope we can validate the patch today and then port it to the EAP branches.

                  • 21. Re: 7.1.3.Final feedbacks
                    silver06

                    any feedbacks Emanuel ? Did you succeed ?

                    • 22. Re: 7.1.3.Final feedbacks
                      emuckenhuber

                      We did include a patch in upstream which basically does 2 things. First we wait for the remote connection is established before starting the next server by default now. Then we also introduced a property "-Dorg.jboss.as.host.start.servers.sequential=true", to start all servers sequential. I wasn't able to reproduce the issue, so i am not entirely sure that this will fix your issue for good. Probably only if you start the servers sequential. I will check if this patch can be applied directly to EAP 6.0.1 though.

                      • 23. Re: 7.1.3.Final feedbacks
                        silver06

                        Yes I would like to know if a patch can be delivered for EAP 6.0.1 as I'm waiting for this fix and don't know what to do today to gather it

                        • 24. Re: 7.1.3.Final feedbacks
                          emuckenhuber

                          AFAIK you should get a one-off patch containing this fix through the support portal. The patch for upstream applies cleanly to the 7.1 branch. I can think of one issue in that branch which might appear, but only if the patch actually does not help.

                          • 25. Re: 7.1.3.Final feedbacks
                            silver06

                            OK I will wait for a patch from the support for the timeout

                            Regarding the 8 server groups limit, as the support is quite long for such a simple bug resolved since 20th of December I've gathered the source code and build from the Maven root file of

                            https://github.com/jbossas/console/tree/1.4.x in which a fix has been performed since 20 days

                            Command: mvn -Peap clean install (for eap 6.0.1)

                             

                            I then copied the file gui\target\as7-console-resources.jar to EAP_HOME\modules\org\jboss\as\console\eap modifying of course the module.xml file and it works perfectly fine

                             

                            Regards

                            • 26. Re: 7.1.3.Final feedbacks
                              silver06

                              For the timeout issue, it has been fixed thanks to https://issues.jboss.org/browse/AS7-6230

                              Will be available in 7.2.0 CR1 or through a patch for EAP 6.0.1 (under construction by the support for the time being)

                              1 2 Previous Next