Three problems maybe related maybe not - Durable Topic Subsc

[mcaughey]

I have a Durable Subscription to a topic. Every now and then I notice the JBoss Server throws:

[SocketManager] Failed to handle: org.jboss.mq.il.uil2.msgs.AcknowledgementRequest from the

Here's the full Trace:

04:32:32,097 ERROR [SocketManager] Failed to handle: org.jboss.mq.il.uil2.msgs.AcknowledgementRequestMsg8251436[msgType: m_acknowledge, msgID: 232, error: null]
java.lang.NullPointerException
 at org.jboss.mq.server.JMSTopic.acknowledge(JMSTopic.java:291)
 at org.jboss.mq.server.ClientConsumer.acknowledge(ClientConsumer.java:328)
 at org.jboss.mq.server.JMSDestinationManager.acknowledge(JMSDestinationManager.java:519)
 at org.jboss.mq.server.JMSDestinationManager.acknowledge(JMSDestinationManager.java:503)
 at org.jboss.mq.server.JMSServerInterceptorSupport.acknowledge(JMSServerInterceptorSupport.java:198)
 at org.jboss.mq.server.JMSServerInterceptorSupport.acknowledge(JMSServerInterceptorSupport.java:198)
 at org.jboss.mq.server.TracingInterceptor.acknowledge(TracingInterceptor.java:481)
 at org.jboss.mq.server.JMSServerInvoker.acknowledge(JMSServerInvoker.java:198)
 at org.jboss.mq.il.uil2.ServerSocketManagerHandler.handleMsg(ServerSocketManagerHandler.java:87)
 at org.jboss.mq.il.uil2.SocketManager$ReadTask.handleMsg(SocketManager.java:369)
 at org.jboss.mq.il.uil2.msgs.BaseMsg.run(BaseMsg.java:377)
 at EDU.oswego.cs.dl.util.concurrent.PooledExecutor$Worker.run(PooledExecutor.java:743)
 at java.lang.Thread.run(Thread.java:595)

I'm running JBoss 4.0.2. My topic is still presisting to the HyperSonic DB. I notice that the message associated with this problem never makes it to my client nor nor does the one after it. I can tell this becuase the time stamp of when I sent it and when I recieve it matches on all the ones recieved. I do a System out just before i sent it and just after I recv it. There is only one subscriber and its on the same computer as JBoss. The client is running in a seperate JVM.

04:32:26,706 INFO [GetInformationJDBCTask] Publishing - 1122625946706
04:32:32,050 INFO [GetInformationJDBCTask] Publishing - 1122625952050
ERROR Occurred Here
04:32:37,113 INFO [GetInformationJDBCTask] Publishing - 1122625957113
04:32:42,128 INFO [GetInformationJDBCTask] Publishing - 1122625962128
04:32:47,175 INFO [GetInformationJDBCTask] Publishing - 1122625967175
04:32:52,191 INFO [GetInformationJDBCTask] Publishing - 1122625972191
04:32:57,238 INFO [GetInformationJDBCTask] Publishing - 1122625977238

Here are the time stamps of the recv'd

 1122625946738 - 0 1,1,
 1122625962175 - 1 1,1,
 1122625967175 - 2 1,1,
 1122625972222 - 3 1,1,
 1122625977253 - 4 1,1,

You can see that two messages were never recieved. How can this happen with a durable topic?

I do have a quick time out of 50 ms. I need it to check the topic and bug out if nothing is there, I return almost instantly. I have to do this becuase of how a plugin (Moka) works with Director, so every second we check again. In my test application I tried two things one is to simply return imediately:

for (int i = 0; i < 5; i++) {
 msg = m.retrieveInfo();
 if(msg == null){
 i--;
 }else{
 System.out.println(" "+ System.currentTimeMillis()+" - " + i + " " + msg);
 }
}

That produces the condition above. I mention this becuase I don't know if the short timeout is the cause the problem or not. I did see it once with a 5000 ms time out. I also mention this becuase I'm seeing another weird thing. If I add a sleep each time the msg comes back null like this:

for (int i = 0; i < 5; i++) {
 msg = m.retrieveInfo();
 if(msg == null){
 i--;
 try {
 Thread.sleep(750);
 } catch (InterruptedException e) {
 }
 }else{
 System.out.println(" "+ System.currentTimeMillis()+" - " + i + " " + msg);
 }
}

I can go a minute with no response from the topic, but all the while the publisher on the server is publishing a new message every 5 seconds. Here's how i'm establishing my Topic,

 public void init() throws JMSException, NamingException {
 Properties props = new Properties();
 props.setProperty(Context.INITIAL_CONTEXT_FACTORY,
 "org.jnp.interfaces.NamingContextFactory");
 props.setProperty(Context.URL_PKG_PREFIXES,
 "org.jboss.naming:org.jnp.interfaces");
 props.setProperty(Context.PROVIDER_URL, serverUrl);
 InitialContext ctx = new InitialContext(props);
 Object o = ctx.lookup("ConnectionFactory");
 TopicConnectionFactory tcf = (TopicConnectionFactory) o;
 conn = tcf.createTopicConnection(username, password);
 // conn = tcf.createTopicConnection();
 topic = (Topic) ctx.lookup("topic/lobbyTopic");
 session = conn.createTopicSession(false, TopicSession.AUTO_ACKNOWLEDGE);
 conn.start();
 sub = session.createDurableSubscriber(topic, username, "Type = "
 + type, true);
 }

Here's my recv method:

 public String recv() throws JMSException, NamingException {
 TextMessage tm = null;
 tm = (TextMessage) sub.receive(50);
 return (tm==null)?null:tm.getText();
 }

So you can see i have the Durable Sbscription the whole time. This test by the way is being run from a standalone commandline console app. It is not yet being run in MOKA.

One last issue. I'm now seeing this error in the log file and it is occurring many times in a row:

04:48:03,265 WARN [Connection] Connection failure, use javax.jms.Connection.setExceptionListener() to handle this error and reconnect
org.jboss.mq.SpyJMSException: No pong received; - nested throwable: (java.io.IOException: ping timeout.)
 at org.jboss.mq.Connection$PingTask.run(Connection.java:1323)
 at EDU.oswego.cs.dl.util.concurrent.ClockDaemon$RunLoop.run(ClockDaemon.java:364)
 at java.lang.Thread.run(Thread.java:595)
Caused by: java.io.IOException: ping timeout.

What is this? I do have 5 durable connections that were created and never unsuscribed. Is the problem related to this? If so that easy to fix. My code has been modified to unscribe the the user. If it is related however then how do you configure JBossMQ to simply drop a client after a period of time, like and expiration time? I would think you could remove all associated messages with that expiration.


Thanks,
michael

[mcaughey]

Actually I mistated one fact, there are five other Subscribers to a seperate Topic. They are not durable and subscribe with no timeout. In retro spec the last error could be coming from them, if the underlying connection is timing out. If this is the case I can just create a durable subscription and time them out and check back if the method comes back null. this is a topic subscriber in an MBean ehich is running in the same server instance as the topic is created in.

-Michael

[adrian.brock]

OK, you asked for this..., just to demonstrate why I wouldn't normally answer it. :-)

Moderated: FAQ
http://wiki.jboss.org/wiki/Wiki.jsp?page=IGetIOExceptionPingTimeoutHowDoIFixIt

Moderated: Which part of the warning "use an exception listener" don't you understand.
Did you even read the javadoc for that method?

Moderated: Use search
Why are you even using the remote client connection factory for in memory work.

See the countless discussions on the in memory connection factory

java:/ConnectionFactory

or the pooled resource adapter

java:/JmsXA

or this feature request:
http://jira.jboss.com/jira/browse/JBAS-1923

etc.

[mcaughey]

Thank you for the response. What seems to you obvious to others may not. And yes I did ask for it, thank you.

I looked at the Wiki about the ping/pong issue.

I had ONE thread connecting to a server. and it's running on the same computer as the server, for now. the Messages are around 500 bytes to 1 K in size. they are placed on the topic every 5 seconds.

This early test ran i a seperate JVM becuase eventually i will move it to a seperate client computer, where subscription will not be able to made from the same JVM. Didn't do the test in process exactly to uncover issues like this. So I'm using two JVMs. Hence the need for UIL2.

In everything I was sorting through I did miss the setExceptionListener(). I knew i was missing something if this is it thank you. Obviously if i had everything configured correctly i would not be having difficulties.

I am still tracking down valuable resources to use to troubleshoot these things myself. trust me i don;t want to have to wait for you or someone else to reply to get my problem fixed. i'd rather fix it myself.

The only thing i see to do is add an ExceptionListner and close and reopen the connection when I get an exception. However if there was only a single connection and both JVMs were o nthe same computer why might the connection fail?


Again thank you for responding,
Michael