[ http://issues.apache.org/jira/browse/GERONIMO-1422?page=all ]
Kevan Miller closed GERONIMO-1422:
----------------------------------
Resolution: Fixed
As Jules suggested, ActiveMQ's shutdown handler was being invoked. This meant
that the broker was stopping before Geronimo could stop any internal clients.
Log entries and exposure of several ActiveMQ bugs resulted. I've set the
activemq.broker.disable-clean-shutdown property to true using the
SystemProperties GBean in the ActiveMQ broker config. Shutdown looks much
cleaner.
ActiveMQ warnings are still possible during shutdown, but should be greatly
reduced. The ActiveMQ shutdown bugs are not fixed by this change. I'll create
new jira's to describe them. The bugs I've seen are:
1) infinite recursion in RA reconnect processing
2) deadlock in Connection close processing
> Geronimo shutdown does not complete due to ActiveMQ attempting to reconnect
> endpoints to broker every 30 seconds
> ----------------------------------------------------------------------------------------------------------------
>
> Key: GERONIMO-1422
> URL: http://issues.apache.org/jira/browse/GERONIMO-1422
> Project: Geronimo
> Type: Bug
> Components: ActiveMQ
> Versions: 1.0
> Environment: Solaris 10 x86 under VMWare player 1.01
> Java 1.4.2_10
> tomcat build of geronimo
> Reporter: John Sisson
> Assignee: Kevan Miller
> Fix For: 1.0.1
> Attachments: geronimo_shutdown_stdout.txt, shutdown.txt, shutdown_rhel3.txt
>
> Can anyone reproduce this problem on other platforms?
> If I start the tomcat build of the release candidate and then shut it down
> once the startup has completed it shuts down almost cleanly:
> Server shutdown begun
> 11:25:47,951 INFO [Http11Protocol] Stopping Coyote HTTP/1.1 on
> http-0.0.0.0-8443
> 11:25:48,986 INFO [Http11Protocol] Stopping Coyote HTTP/1.1 on
> http-0.0.0.0-8080
> 11:25:49,001 INFO [StandardContext] Container
> org.apache.catalina.core.ContainerBase.[Geronimo].[localhost].[/] has not
> been started
> Server shutdown completed
> I have shutdown issues If I do the following:
> * start the tomcat build of release candidate using geronimo.sh run --long
> * connect to the daytrader web app
> * populate the daytrader database via the daytrader configuration page
> * log into daytrader and view account, portfolio etc.
> * press ctrl-C in the window that geronimo was started in to shut it down.
> * You will see ActiveMQAsfEndpointWorker messages every 30 seconds.
> 10:46:56,356 WARN [BrokerContainerImpl] Got duplicate deregisterConnection
> for client: ID:unknown-34799-1136504488185-10:0
> 10:46:56,358 WARN [TransportChannelSupport] Caught exception dispatching
> message and no ExceptionListener registered: javax.jms.JMSException: Error
> reading socket: java.io.EOFException
> javax.jms.JMSException: Error reading socket: java.io.EOFException
> at
> org.activemq.util.JMSExceptionHelper.newJMSException(JMSExceptionHelper.java:49)
> at
> org.activemq.transport.tcp.TcpTransportChannel.doClose(TcpTransportChannel.java:509)
> at
> org.activemq.transport.tcp.TcpTransportChannel.run(TcpTransportChannel.java:330)
> at java.lang.Thread.run(Thread.java:534)
> Caused by: java.io.EOFException
> at java.io.DataInputStream.readByte(DataInputStream.java:333)
> at
> org.activemq.io.AbstractWireFormat.readPacket(AbstractWireFormat.java:230)
> at
> org.activemq.transport.tcp.TcpTransportChannel.run(TcpTransportChannel.java:313)
> ... 1 more
> 10:47:27,401 INFO [ActiveMQAsfEndpointWorker] Endpoint connection to JMS
> broker failed: Initialization of TcpTransportChannel failed. URI was:
> tcp://localhost:61616 Reason: java.net.ConnectException: Connection refused
> 10:47:27,402 INFO [ActiveMQAsfEndpointWorker] Endpoint will try to reconnect
> to the JMS broker in 30 seconds
> 10:47:27,403 INFO [ActiveMQAsfEndpointWorker] Endpoint connection to JMS
> broker failed: Initialization of TcpTransportChannel failed. URI was:
> tcp://localhost:61616 Reason: java.net.ConnectException: Connection refused
> 10:47:27,403 INFO [ActiveMQAsfEndpointWorker] Endpoint will try to reconnect
> to the JMS broker in 30 seconds
> I will have attached a capture of stdout also including a thread dump to this
> issue.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
