[jboss-jira] [JBoss JIRA] Closed: (JBCLUSTER-32) Problem using HA-JNDI and Fail Over

[Brian Stansberry (JIRA)]

     [ http://jira.jboss.com/jira/browse/JBCLUSTER-32?page=all ]

Brian Stansberry closed JBCLUSTER-32.
-------------------------------------

    Fix Version/s: Q3Y5
                       (was: Bugs)
       Resolution: Done

> Problem using HA-JNDI and Fail Over
> -----------------------------------
>
>                 Key: JBCLUSTER-32
>                 URL: http://jira.jboss.com/jira/browse/JBCLUSTER-32
>             Project: JBoss Clustering
>          Issue Type: Bug
>      Security Level: Public(Everyone can see) 
>         Environment: JBoss 4.01sp1, Windows XP Professional, Sun JDK 1.4.2_07, Apache 2.0.53
>            Reporter: Roberto
>         Assigned To: Adrian Brock
>             Fix For: Q3Y5
>
>         Attachments: NamingException Error.jpg, testcasecluster.ear, Working Result.jpg
>
>
> I have 3 machines: 
> "HostA" : Load balance base on Apache 2.0.53 with mod_jk 1.2.x with sticky session 
> "ClusterB" : Jboss AS 4.0.1 sp1 with my ear with war and ejb modules 
> "ClusterC" : Jboss AS 4.0.1 sp1 with my ear the same than in ClusterB 
> in my ear i have also a jndi.properties to use HA-JNDI: 
> Code: 
> java.naming.factory.initial=org.jnp.interfaces.NamingContextFactory
> java.naming.provider.url=jnp://localhost:1100
> java.naming.factory.url.pkgs=org.jboss.naming:org.jnp.interfaces
> #jnp.partitionName=DefaultPartition
> Load balance is configure for sticky session. 
> My client is a standard browser (Internet Explorer / Firefox). 
> So a user using browser access to the HostA and so the request is forward for example to ClusterB. 
> For now all work fine: the request is captured from a Servlet/JSP the lookup on JNDI using for example: 
> try
> {
> ABSEQBAuthHome vHome = (ABSEQBAuthHome)  new InitialContext().lookup("java:comp/env/ejb/ABSEQBAuth");
> }
> catch (Exception vErr)
> {}
> Setting a jndi.properties in the ear, this force to use port 1100 (HA-JNDI) 
> Now i shutdown (CTRL-C) ClusterB. 
> The user in the same browser session (or in another) do another request and so the load-balance try to use again ClusterB, but now is died and so correctly forward the request to the alive clusterA. 
> The request is always capture from the servlet/jsp that done another lookup (always using HA-JNDI), but now a NameNotFoundException occurs. 
> I think that when the shutdown occurs all ejb are undeploy from the HA-JNDI.... 
> I have prepared an EAR to test it (testcasecluster.ear).
> Use this step to replicate the problem:
> When HostA, ClusterA and ClusterB are online, try to access to:
> http://hosta/testcasecluster/TestCase.jsp (to login user "user" with psw "user")
> Ok.. at this moment all work fine.
> Now showtdown a cluster (using shutdown command or CTRL-C)
> When cluster is down try to access again to the same url.
> At this moment nothing works!
> I need to use HA-JNDI without change my java code.
> Thank You 
> Roberto

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira