Hello Eric, I have a similar problem with ActiveMQ 5.2.
http://old.nabble.com/Producers-blocking-td26448458.html What i did when I encountered the same "global freeze" was to disable the producer flow control of several topics from the JConsole. Then, the message publishing of the applications could restart. (without restarting the brokers instances) I have the feeling that a global failure of the network of brokers may be linked with advisory messages topics, but I am not sure. Now, I have disabled the producer flow control for all my topics in the configuration file and everything is under testing. Greg 2009/11/22 Eric-AWL <eric.vinc...@atosorigin.com> > > Hi > > In ActiveMQ 5.2. > > I have a large network of brokers with a lot of topics and queues > consumers. > All brokers are connected together (with duplex network connectors, or not > duplex connectors), but with a network TTL of 1 only. > > We don't use persistent mode at all. > > It is possible that some consumer processes are down for a long time. So, > some messages are produced and not consumed at all. > > After some days, we see that problems appear progressively all over our > brokers, and finally all the brokers seem affected, even brokers that are > not producer or never consumed messages. Some other consumers seem not to > receive any messages produced. > > I know that it is not the normal behaviour of our application. But I just > want to know if the "global freeze" situation can be the result of the fact > that some messages are not consumed on some brokers. > > So, I wonder if the QueueBrowser and TopicBrowser usage can be a solution > for delete old not consumed messages ? > And I wonder if they are considered as consumers as far as Advisory > Messages > are concerned ? > > > Eric-AWL > -- > View this message in context: > http://old.nabble.com/What-happens-where-a-consumer-is-not-present---tp26464162p26464162.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > >
