Re: Unhandled exception when using High Availabilty in Qpid Java Broker 6.0.0

Thu, 10 Mar 2016 15:28:30 -0800 

On 03/10/2016 05:44 PM, Adel Boutros wrote:

If I compile while skipping tests and deploy the jar then the broker no
longer crashes.
Nevertheless, the client doesn't try the 2nd URL, instead it fails with the
below stack. If I put back the MASTER URL before the Replicate URL, then
everything works fine.

It seems to me that the order in which the URLs are provided in the
failover shouldn't matter. Am I correct?
What is happening at the time this error is received? The InvalidDestinationException error is treated differently as I recal because it indicates you have done something invalid not that the client has lost it's connection to the broker. Since an invalid destination error is something that can be recovered from the client wouldn't try and reconnect because of it. Now if you are losing your connection following this then I'd expect a failover but I don't see enough info here to infer that the connection dropped say I can't say. 




javax.jms.InvalidDestinationException: Received error from remote peer
without description [condition = amqp:not-found]
at
org.apache.qpid.jms.provider.amqp.AmqpSupport.convertToException(AmqpSupport.java:109)
at
org.apache.qpid.jms.provider.amqp.AmqpAbstractResource.remotelyClosed(AmqpAbstractResource.java:115)
at
org.apache.qpid.jms.provider.amqp.AmqpAbstractResource.processRemoteClose(AmqpAbstractResource.java:210)
at
org.apache.qpid.jms.provider.amqp.AmqpProvider.processUpdates(AmqpProvider.java:772)
at
org.apache.qpid.jms.provider.amqp.AmqpProvider.access$1900(AmqpProvider.java:92)
at
org.apache.qpid.jms.provider.amqp.AmqpProvider$17.run(AmqpProvider.java:701)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
at java.util.concurrent.FutureTask.run(FutureTask.java:262)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:178)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:292)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:744)

Regards,
Adel

On Thu, Mar 10, 2016 at 11:36 PM, Adel Boutros <sta191...@gmail.com> wrote:


Hello Rob,

I have 3 tests failing under *ProtocolEngine_1_0_0Test* when compiling 
*broker-plugins
6.0.2-SNAPSHOT *in Intellij or maven

    - ProtocolEngine_1_0_0Test.testProtocolEngineWithNoSaslNonTLSandAnon
    - ProtocolEngine_1_0_0Test.testProtocolEngineWithNoSaslTLSandExternal
    - ProtocolEngine_1_0_0Test.testProtocolEngineWithSaslNonTLSandAnon


All 3 tests are failing with the same issue:

Wanted but not invoked:
virtualHostImpl.registerConnection(<any>);
-> at
org.apache.qpid.server.protocol.v1_0.ProtocolEngine_1_0_0Test.testProtocolEngineWithNoSaslNonTLSandAnon(ProtocolEngine_1_0_0Test.java:176)

However, there were other interactions with this mock:
-> at
org.apache.qpid.server.protocol.v1_0.Connection_1_0.openReceived(Connection_1_0.java:182)
-> at
org.apache.qpid.server.protocol.v1_0.Connection_1_0.openReceived(Connection_1_0.java:191)

Wanted but not invoked:
virtualHostImpl.registerConnection(<any>);
-> at
org.apache.qpid.server.protocol.v1_0.ProtocolEngine_1_0_0Test.testProtocolEngineWithNoSaslNonTLSandAnon(ProtocolEngine_1_0_0Test.java:176)

However, there were other interactions with this mock:
-> at
org.apache.qpid.server.protocol.v1_0.Connection_1_0.openReceived(Connection_1_0.java:182)
-> at
org.apache.qpid.server.protocol.v1_0.Connection_1_0.openReceived(Connection_1_0.java:191)

Can you please advise?

Regards,
Adel Boutros

On Thu, Mar 10, 2016 at 9:20 PM, Rob Godfrey <rob.j.godf...@gmail.com>
wrote:


Apologies for this - you have indeed run into bug :-(

I've raised a JIRA - https://issues.apache.org/jira/browse/QPID-7136 .

I've committed a fix both to trunk, and the 6.0.x branch which I hope will
fix this issue...  If you want to test it out you can use the 6.0.x branch
which can be found here:

https://svn.apache.org/repos/asf/qpid/java/branches/6.0.x


Hope this helps,
-- Rob

On 10 March 2016 at 20:10, Adel Boutros <sta191...@gmail.com> wrote:


Hello,

When we provide a failover URL where the first URL is that of a

replicate

broker, the connection will trigger an IllegalStateException which is

not

caught anywhere. Thus making the broker crash and the session creation

fail

on the client-side.

Shouldn't the "close" method log the exception silently instead of

crashing

the broker?

*Code*

String brokerUrl =
"failover:(amqp://localhost:5672,amqp://localhost:10352)";
ConnectionFactory connectionFactory = new

JmsConnectionFactory(brokerUrl);

Connection connection = connectionFactory.createConnection();
Session session = connection.createSession(false,
Session.AUTO_ACKNOWLEDGE);

*Test case*
Assuming that the broker on *localhost:5672 *is alive but as a

*REPLICATE*

and the one on* localhost:10352* as the *MASTER*

*Stack trace*
java.lang.IllegalStateException: The virtual host state of UNAVAILABLE

does

not permit this operation.
         at



org.apache.qpid.server.virtualhost.berkeleydb.BDBHAReplicaVirtualHostImpl.throwUnsupportedForReplica(BDBHAReplicaVirtualHostImpl.java:567)

         at



org.apache.qpid.server.virtualhost.berkeleydb.BDBHAReplicaVirtualHostImpl.deregisterConnection(BDBHAReplicaVirtualHostImpl.java:562)

         at



org.apache.qpid.server.protocol.v1_0.Connection_1_0.closed(Connection_1_0.java:282)

         at



org.apache.qpid.server.protocol.v1_0.AMQPConnection_1_0.closed(AMQPConnection_1_0.java:470)

         at



org.apache.qpid.server.transport.MultiVersionProtocolEngine.closed(MultiVersionProtocolEngine.java:112)

         at



org.apache.qpid.server.transport.NonBlockingConnection.shutdown(NonBlockingConnection.java:359)

         at



org.apache.qpid.server.transport.NonBlockingConnection.doWork(NonBlockingConnection.java:299)

         at



org.apache.qpid.server.transport.NetworkConnectionScheduler.processConnection(NetworkConnectionScheduler.java:108)

         at



org.apache.qpid.server.transport.SelectorThread$ConnectionProcessor.processConnection(SelectorThread.java:499)

         at



org.apache.qpid.server.transport.SelectorThread$SelectionTask.performSelect(SelectorThread.java:337)

         at



org.apache.qpid.server.transport.SelectorThread$SelectionTask.run(SelectorThread.java:86)

         at



org.apache.qpid.server.transport.SelectorThread.run(SelectorThread.java:457)

         at



java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)

         at



java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)

         at java.lang.Thread.run(Thread.java:744)

If you invert the order of the URLs to put the master first or simply

kill

the REPLICATE broker, then there will be no exception triggered.

Regards,
Adel Boutros






--
Tim Bish
twitter: @tabish121
blog: http://timbish.blogspot.com/


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org
For additional commands, e-mail: users-h...@qpid.apache.org

Reply via email to