The 3.x implementation does not use spring anymore. So we do not have
the spring implementations of MessageListenerContainer that worked with
separate consumers and polling.
If you look at
org.apache.cxf.transport.jms.JMSDestination.createTargetDestinationListener()
you see how the MessageListenerContainer we use now is created.
If there is a transaction manager then we use the
PollingMessageListenerContainer which uses one thread. If there is no
transaction manager then we use a MessageListenerContainer based on a
MessageListener.
In my performance tests
(http://liquid-reality.de/display/liquid/2014/03/28/Revisiting+JMS+performance.+Improvements+in+CXF+3.0.0)
this worked at least as fast as the spring variant with several consumers.
I only tested with ActiveMQ though so if you use a different provider
you may see a different performance.
So as a short answer you currently can not set the consumer threads. We
could add this if there is need but currently I did not find a good
reason. So I stayed with the simplest solution.
Which provider do you use?
Did your jms performance with default settings decrease after the switch?
Christian
On 07.10.2014 14:43, Eernie wrote:
Hi,
I'm currently migrating my code from cxf 2.6.1 to 3.0.1.
I'm having some problems with the jms protocol. We're currently using the
maxConcurrentConsumers from the jmsConfiguration. I see this property has
been removed. Can someone tell me how I should configure this now?
The Documentation isn't explaining how this should be configured. Atleast, I
didn't find it.
I did find a StackOverflow post, but now comments were post on that.
See:
http://stackoverflow.com/questions/24785219/apache-cxf-jms-3-0-and-spring-configuration#comment41150668_24785219
Thanks in advance,
Erwin Oldenkamp
--
View this message in context:
http://cxf.547215.n5.nabble.com/CXF-3-0-1-doesn-t-have-concurrentConsumers-tp5749534.html
Sent from the cxf-user mailing list archive at Nabble.com.
--
Christian Schneider
http://www.liquid-reality.de
Open Source Architect
http://www.talend.com