1 2 3 4 Previous Next 50 Replies Latest reply on Nov 18, 2008 12:43 PM by vblagojevic Go to original post
      • 45. Re: Changes on the JBM stack on JBoss5
        clebert.suconic

        So... what's the solution for us then? Comment out the test for now and aways make the tests are using multiple VMs?

        • 46. Re: Changes on the JBM stack on JBoss5
          timfox

           

          "clebert.suconic@jboss.com" wrote:
          So... what's the solution for us then? Comment out the test for now and aways make the tests are using multiple VMs?



          The tests are valid so shouldn't be commented out.

          If Jgroups used to be able to run multiple stacks in the same VM and now can't seems like a bug to me. Or at least it's an incompatible change which shouldn't occur in a minor version.

          On the other hand, if there's some config we have to change to get them to work, fair enough, but we need to know what that is...

          • 47. Re: Changes on the JBM stack on JBoss5
            brian.stansberry

             

            "clebert.suconic@jboss.com" wrote:
            So... what's the solution for us then? Comment out the test for now and aways make the tests are using multiple VMs?


            The solution is to do what I asked and provide the details so we can debug the issue.

            You can't run multiple channels with the same singleton name and the same channel name in the same VM. That's a hassle in testsuites, which is where people typically try to do that. This is not a regression; JBM didn't have a singleton name in its config before. You and I discussed the two workarounds to this, which are basically testsuite config changes. You say the workarounds didn't work. We need to understand why not, so we can determine if that means there is an actual JGroups regression.

            • 48. Re: Changes on the JBM stack on JBoss5
              clebert.suconic

               

              "bstansberry@jboss.com" wrote:

              You can't run multiple channels with the same singleton name and the same channel name in the same VM. That's a hassle in testsuites, which is where people typically try to do that. This is not a regression; JBM didn't have a singleton name in its config before. You and I discussed the two workarounds to this, which are basically testsuite config changes. You say the workarounds didn't work. We need to understand why not, so we can determine if that means there is an actual JGroups regression.


              I had tried your suggestions and I couldn't get each JChannel to talk to each other.

              After Vlad's post I understood that this wouldn't work any way.


              But today I was playing with JChannel alone in a test, and I figured out I needed to make sure I had java.net.preferIPv4Stack=true on my VM. After setting that it worked.

              • 49. Re: Changes on the JBM stack on JBoss5
                brian.stansberry

                Good news. Thanks. :-)

                • 50. Re: Changes on the JBM stack on JBoss5
                  vblagojevic

                  Great stuff Clebert! Thanks for scrutinizing and battle testing FLUSH. You found an important bug and I am glad we caught it in time! Cheers.

                  1 2 3 4 Previous Next