Mailing-List: contact activemq-users-help@geronimo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: activemq-users@geronimo.apache.org
Received-SPF: pass (asf.osuosl.org: domain of lists@nabble.com designates
 72.21.53.35 as permitted sender)
Message-ID: <3908376.post@talk.nabble.com>
Date: Thu, 13 Apr 2006 14:02:35 -0700 (PDT)
From: skarthik <karthik.sethuraman@necam.com>
To: activemq-users@geronimo.apache.org
Subject: publish() call to remote JMS server hangs
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit


Hi,

 This is rather long post as it took me a while to narrow down the problem.

 We are running activeMQ RC2 server on a windows XP machine and a simple JMS
client test-program on another remote XP machine within the same LAN subnet.
The test-program opens one JMS connection to the activeMQ server,
establishes one session (with auto acknowledge), and creates one publisher
on a topic on the session. Then in a loop, it creates an ObjectMessage and
publishes it with PERSISTENT DeliveryMode, peppered by Thread.sleep() calls
after each publish. The following snippet provides an example:

         TopicConnectionFactory tcf = (TopicConnectionFactory)
jndiCtx.lookup (tcfName);
         Topic topic = (Topic) jndiCtx.lookup (topicName);

         TopicConnection jmsCon = tcf.createTopicConnection();
         TopicSession session = jmsCon.createTopicSession (false,
Session.AUTO_ACKNOWLEDGE);
         TopicPublisher publisher = session.createPublisher (topic);

         while (true) {
            ObjectMessage om = session.createObjectMessage ();
            om.setObject ("TESTSTRING");
            publisher.publish (om, DeliveryMode.PERSISTENT, 9, 20000);

            Thread.sleep (20000);
         }


 It was found that when we disconnect the LAN cable from the client machine
(to simulate network outage), if the test client tries to publish an
ObjectMessage before activeMQ client stub detects the network disconnect,
the publish call hangs.

 After further analysis, we found that even setting simple string payload to
the ObjectMessage as shown in the following code snippet causes the hang.

         ObjectMessage om = session.createObjectMessage ();
         om.setObject ("TESTSTRING");
         publisher.publish (om, DeliveryMode.PERSISTENT, 9, 20000);

 But using a byte array (byte[]) as payload does not cause an hang.

        byte[] payload; // populated by reading a specified input ascii file

         ObjectMessage om = session.createObjectMessage ();
         om.setObject (payload);
         publisher.publish (om, DeliveryMode.PERSISTENT, 9, 20000);

 Also setting a complex java serializable class instance as payload causes a
hang, but an instance of almost similar class, but with an additional byte[]
array attribute does not cause the publish() call to hang.

 Wanted to clarify that this hang is observered only in the test case
scenario involving LAN disconnects. In normal cases the same messages flow
sucessfully between the multiple remote producers and consumers. Also when I
say the publish() call does not hang, I meant the publish() class throws the
following exception which is the normal expected behavior. 

javax.jms.JMSException: Connection reset by peer: socket write error	at

org.apache.activemq.util.JMSExceptionSupport.create(JMSExceptionSupport.java:57)
	at
org.apache.activemq.ActiveMQConnection.syncSendPacket(ActiveMQConnection.java:1118)
	at org.apache.activemq.ActiveMQSession.send(ActiveMQSession.java:1524)
	at
org.apache.activemq.ActiveMQMessageProducer.send(ActiveMQMessageProducer.java:462)
	at
org.apache.activemq.ActiveMQMessageProducer.send(ActiveMQMessageProducer.java:384)
	at
org.apache.activemq.ActiveMQTopicPublisher.publish(ActiveMQTopicPublisher.java:151)	
                at
com.nec.tdd.tools.platformX.util.PingJMSMsg$Publisher.publishMessage(PingJMSMsg.java:122)
	at
com.nec.tdd.tools.platformX.util.PingJMSMsg$Publisher.run(PingJMSMsg.java:102)
Caused by: java.net.SocketException: Connection reset by peer: socket write
error	at java.net.SocketOutputStream.socketWrite0(Native Method)
	at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:92)
	at java.net.SocketOutputStream.write(SocketOutputStream.java:136)
	at
org.apache.activemq.transport.tcp.TcpBufferedOutputStream.write(TcpBufferedOutputStream.java:95)
	at java.io.DataOutputStream.write(DataOutputStream.java:85)
	at
org.apache.activemq.openwire.v1.BaseDataStreamMarshaller.tightMarshalByteSequence2(BaseDataStreamMarshaller.java:403)
	at
org.apache.activemq.openwire.v1.MessageMarshaller.tightMarshal2(MessageMarshaller.java:160)
	at
org.apache.activemq.openwire.v1.ActiveMQMessageMarshaller.tightMarshal2(ActiveMQMessageMarshaller.java:88)
	at
org.apache.activemq.openwire.v1.ActiveMQObjectMessageMarshaller.tightMarshal2(ActiveMQObjectMessageMarshaller.java:88)
	at
org.apache.activemq.openwire.OpenWireFormat.marshal(OpenWireFormat.java:240)
	at
org.apache.activemq.transport.tcp.TcpTransport.oneway(TcpTransport.java:120)
	at
org.apache.activemq.transport.InactivityMonitor.oneway(InactivityMonitor.java:141)
	at
org.apache.activemq.transport.TransportFilter.oneway(TransportFilter.java:86)
	at
org.apache.activemq.transport.WireFormatNegotiator.oneway(WireFormatNegotiator.java:77)
	at
org.apache.activemq.transport.MutexTransport.oneway(MutexTransport.java:44)
	at
org.apache.activemq.transport.ResponseCorrelator.asyncRequest(ResponseCorrelator.java:63)
	at
org.apache.activemq.transport.ResponseCorrelator.request(ResponseCorrelator.java:68)
	at
org.apache.activemq.ActiveMQConnection.syncSendPacket(ActiveMQConnection.java:1108)
	... 6 more

  I hope this was helpful and would lead to a fix as this is a crital test
scenario for us.

Thanks in advance

regards
karthik
--
View this message in context: http://www.nabble.com/publish%28%29-call-to-remote-JMS-server-hangs-t1446830.html#a3908376
Sent from the ActiveMQ - User forum at Nabble.com.