The fact that your number of dispatched messages is an exact multiple of your prefetch size tells me that some of your threads (which sound like they're each acting as a JMS consumer, for a total of 8 consumers; you can confirm that via the brokers' web consoles) are probably slow or blocked, so I'd look at your message-consuming logic to see if you've got something that could go wrong there. If you can take a thread dump, that might provide more information.
Tim On Fri, Aug 12, 2016 at 12:21 AM, softwbc <soft...@163.com> wrote: > I have two consumers and Every consumer has 4 threads. > I'm use two broker to from cluster: > <networkConnectors> > <networkConnector > uri="masterslave:(tcp://192.168.83.167:61616,tcp://192.168.83.168:61617)? > maximumConnections=1000&wireFormat.maxFrameSize=104857600" > userName="admin" password="xinhua" duplex="true" > conduitSubscriptions="false" /> > </networkConnectors> > > There are 3000 message have dispatched on server168 and 2000 message on > server167 > > Im using the default prefetch size > > I adjusted the parameters being observed. > > Thanks > > > > -- > View this message in context: http://activemq.2283324.n4. > nabble.com/activemq-3000-messages-can-not-be-received- > tp4715396p4715440.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. >
