Also, the network connector you have listed (see below) will force the embedded broker to open two connections, one to the broker on ytsvmx01s and the other to the broker on ytsvmx02s. So if you have this is in each of your activemq xml files, it would result in each broker trying to connect to itself.
<networkConnector name="pool" uri="static://(tcp://ytsvmx01s:61616,tcp://ytsvmx02s:61616)" conduitSubscriptions="false" dynamicOnly="true" /> Joe Eric-AWL wrote: > > It's because I don't know ActiveMQ very well :-) . I will try your > configuration which seems far easier. > > I think I find the reason why my programs hangs. > > On my process client, I create the temporary queue and just behind, I send > a message with the JMSSetReply set to this temporary queue. I think that > the second ActiveMQ instance has not had the time to create the temporary > queue, when the server process tries to send the answer on it. > > If I put some "Thread.sleep" after I created the temporary queue and > before I delete them, the things seems to go in a better way. But I have > to practice a little more to be sure. > > It seems to be a problem of synchronization between the ActiveMQ intances > in such a configuration. > > I continue my tests > > Eric > -- View this message in context: http://www.nabble.com/%22Embedded-broker%22%2C-%22network-of-brokers%22-and-%22temporary-queues%22-tp15023996s2354p15051352.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
