Encounter ManagedConnection already destroyd when creating JMS session

[niefeng]

Hi all,

I encountered the following exception when trying to create a JMS Session object.

javax.resource.spi.IllegalStateException: ManagedConnection already destroyd
    at org.jboss.resource.adapter.jms.JmsManagedConnection.getConnection(JmsManagedConnection.java:245)
    at org.jboss.resource.connectionmanager.BaseConnectionManager2.allocateConnection(BaseConnectionManager2.java:411)
    at org.jboss.resource.connectionmanager.BaseConnectionManager2$ConnectionManagerProxy.allocateConnection(BaseConnectionManager2.java:849)
    at org.jboss.resource.adapter.jms.JmsSessionFactoryImpl.allocateConnection(JmsSessionFactoryImpl.java:390)
    at org.jboss.resource.adapter.jms.JmsSessionFactoryImpl.createSession(JmsSessionFactoryImpl.java:370)

My code is as follow

Connection conn = (Connection)getConnection(jmsConfigProps, destName);
Session session = conn.createSession(false,Session.AUTO_ACKNOWLEDGE); //<-- throw the above exception

There is total two nodes (node 1 (main node), node 2) in the cluster. In order to see how my application behave if the cluster failed, i simulated a node crash by killing the node 1(main node) in the cluster. In the node 2, i am observing the above mentioned exception.

The clustered is deployed in the Fedora box. The jboss i am using is Jboss AS 4.2.3, jre 150_20, DB is postgres 8.4.4. The JMS setup is a HA JMS. The following is my JMS XA configuration that is found in the ha-jndi-jms-ds.xml

<!-- JMS XA Resource adapter, use this to get transacted JMS in beans -->
  <tx-connection-factory>
    <jndi-name>JmsXA</jndi-name>
    <xa-transaction/>
    <rar-name>jms-ra.rar</rar-name>

    <track-connection-by-tx/>
 
    <connection-definition>org.jboss.resource.adapter.jms.JmsConnectionFactory</connection-definition>
    <config-property name="SessionDefaultType" type="java.lang.String">javax.jms.Topic</config-property>
    <config-property name="JmsProviderAdapterJNDI" type="java.lang.String">java:/DefaultJMSProvider</config-property>
    
    <max-pool-size>100</max-pool-size>
    <blocking-timeout-millis>400000</blocking-timeout-millis>
    <idle-timeout-minutes>2</idle-timeout-minutes>
    <no-tx-separate-pools/>
    
    <security-domain-and-application>JmsXARealm</security-domain-and-application>
  </tx-connection-factory>

My program flow is as follow

MD Bean receive the jms message

MD Bean process some logic

MD Bean delegate to the next handler via JMS (<-- this is the point where the above exception happen)

What could be the cause of the above exception? Is it possible to recover from such error?

Many thanks

Niefeng

[b.eckenfels]

You did not show the code to close the connection (finally?) and you did not show the content of "getConnection" method. I suspect your getConnection method caches the managed connection in a static member, and this is your problem.  For the JCA factory the best thing you can do is to not cache it at all, but request it every time new from the factory (and close it in a finally after each request).