I'm guessing you still have client-side code problems, which is very common. I suggest using the out-of-the-box test harnesses available from ActiveMQ to help eliminate the broker configuration. Try running 100, 1,000 and then ramp up to 100,000 messages through your Master broker. If your configuration is correct, you won't have any problems. If this test succeeds with a couple hundred thousand messages going through it, than you know you need to focus on the client-side code.

http://activemq.apache.org/examples.html

Hope this helps!
Matt Pavlovich

On 2/9/12 12:12 PM, Abimael wrote:
Hi Matt
First, thank you.
Second , unfortunely it does not work.
I did both changes in the configuration you had suggested, but it does not
work.
In the first 2 attempts, I was really excited, because it had worked . But
when I did another attempt, it did not.
To give you a scenario : I 'm running 2 Java IDEs (IntelliJ ), one with the
producer ; other with consumer .
THe url configured on Producer and Consumer apps :
"failover:(tcp://sim1648:61616,tcp://sim1660:61616)?randomize=false";

For the brokers, I 'm using 2 machines , one as master, other as slave.
To do the test, I do : start master; start slave . Start consumer , start
producer .
I am able to see the messages sent by producer to consumer .
Then, I stop the master .

This is the message in the console of the Consumer app,  for the failure :
9193 [ActiveMQ Transport: tcp://sim1648/127.0.1.1:61616] WARN
org.apache.activemq.transport.failover.FailoverTransport - Transport
(sim1648/127.0.1.1:61616) failed to tcp://sim1648:61616 , attempting to
automatically reconnect due to: java.io.EOFException

Process finished with exit code 0
As you can see, the consumer just ends without any other message.

I really do not know and have no ideas. I guess that I'm not alone, as
another person started this thread.

thanks in advance

--
View this message in context: 
http://activemq.2283324.n4.nabble.com/Activemq-Clustering-and-failover-tp4326849p4373786.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.

Reply via email to