Question

I am trying to simulate WebSphere MQ reason code 2009 to handle in the below JMS code but not able to get it. Instead I am getting 2059. All I am doing is disconnecting SVRCONN channel while making the connection call. How can I get 2009 in my sample code.
I have added a sleep time prior making connection again and using transacted sessions. What else can be done to handle reason code 2009 properly that eventually Queue manager won’t get thrashed by frequent unsuccessful connection attempts.
Please find the code.

private static void connectToQmgr(MQQueueConnectionFactory cf) {
    // TODO Auto-generated method stub

    MQQueueConnection connection = null;
    MQQueueSession session = null;
    MQQueue queue = null;
    MQQueueSender sender = null;

    //While Statement to make sure multiple connection tries are made until connection establishes

    while (connection == null){

    try {
        connection = (MQQueueConnection) cf.createConnection();
        session = (MQQueueSession) connection.createQueueSession(true, Session.CLIENT_ACKNOWLEDGE);
        queue = (MQQueue) session.createQueue("queue:///LQ");
        sender = (MQQueueSender) session.createSender(queue);
        //MQQueueReceiver receiver = (MQQueueReceiver) session.createReceiver(queue);
        long uniqueNumber = System.currentTimeMillis() % 1000;
        TextMessage message = session.createTextMessage("MQJMSTest "+ uniqueNumber);
        // Start the connection
        connection.start();
        sender.send(message);
        session.commit();
        System.out.println("Sent message:\\n" + message);

       // JMSMessage receivedMessage = (JMSMessage) receiver.receive(10000);
       // System.out.println("\\nReceived message:\\n" + receivedMessage);

        session.commit();

        sender.close();
       // receiver.close();
        session.close();
        connection.stop();
        connection.close();

        System.out.println("\\nSUCCESS\\n");

    } catch (JMSException je) {
        System.err.println("Caught JMSException");

        // Check for linked exceptions in JMSException to catch MQException and Reason Codes 
        Throwable t = je;
        while (t != null) {
            // Write out the message that is applicable to all exceptions
            System.err.println("Exception Msg: " + t.getMessage());
            // Write out the exception stack trace
            t.printStackTrace(System.err);

            // Add on specific information depending on the type of exception
            if (t instanceof JMSException) {
                JMSException je1 = (JMSException) t;
                System.err.println("JMS Error code: " + je1.getErrorCode());

                if (t instanceof JmsExceptionDetail){
                    JmsExceptionDetail jed = (JmsExceptionDetail)je1;
                    System.err.println("JMS Explanation: " + jed.getExplanation());
                    System.err.println("JMS Explanation: " + jed.getUserAction());
                }
            } else if (t instanceof MQException) {
                MQException mqe = (MQException) t;
                System.err.println("WMQ Completion code: " + mqe.getCompCode());
                System.err.println("WMQ Reason code: " + mqe.getReason());

                //###################################################################################
                //MQ Reason Code Error Handle here
                //Currently Handling MQ Reason Code 2059 since unable to simulate 2009
                //If connection handle exists make sure you close everything and add a wait interval and try a new connection again
                if (mqe.getReason()== 2059){
                    System.out.println("Inside MQ Reson Code Handle");

                    if( connection != null){
                         try {
                            sender.close();
                            // receiver.close();
                            session.close();
                            connection.stop();
                            connection.close();
                        } catch (JMSException e) {
                            // TODO Auto-generated catch block
                            e.printStackTrace();
                        }
                    }

                    // Add Wait Interval for 5 sec
                    try {
                        Thread.sleep(5000);
                        System.out.println("Inside Thread Sleep for 5 sec");
                        //Try New connecting Again
                        connectToQmgr(cf);
                    } catch (InterruptedException e) {
                        // TODO Auto-generated catch block
                        e.printStackTrace();

                    }   
                }
                 //##################################################################################           
            } else if (t instanceof JmqiException){
                JmqiException jmqie = (JmqiException)t;
                System.err.println("WMQ Log Message: " + jmqie.getWmqLogMessage());
                System.err.println("WMQ Explanation: " + jmqie.getWmqMsgExplanation());
                System.err.println("WMQ Msg Summary: " + jmqie.getWmqMsgSummary());
                System.err.println("WMQ Msg User Response: "
                                   + jmqie.getWmqMsgUserResponse());
                System.err.println("WMQ Msg Severity: " + jmqie.getWmqMsgSeverity());
            }

            // Get the next cause
            t = t.getCause();
       }
    }

    }

}

Answer 1

A 2009 is MQRC_CONNECTION_BROKEN. Have you considered pulling the network cable or shutting down the network interface over which the connection travels? This is easy enough when running the code on your laptop. For servers, you can try routing the connection through an SSH proxy or other virtual interface that can be shut down.

To answer the second part of your question, make sure the program sleeps a few seconds between reconnect attempts. Or, better yet, use a modern version of WMQ client and use the automatic reconnect option. This will make a couple of fast retries and then slow down a bit for subsequent retries.