Config 2 isn't recommended, but configs 1 and 3 are both valid and should behave equivalently. The fact that they don't is concerning.
Is your 2000-message test the same thing as your stress test, or are those different tests? If they're different, can you please give more details about how the 2000-message test fails? Can you give more details about the network bridge disconnections? Logs, frequency, etc.? Do you have any slow consumer handling in place? And are you using durable or non-durable topic subscriptions? Tim On Mon, Apr 8, 2019, 5:32 AM denisbrondy <denis.bro...@thalesaleniaspace.com> wrote: > Hi guys, > > I'm facing a very strange problem. I tested multiple versions of ACTIVEMQ > and I'm still seeing this issue on 5.15.9. > > The setup is the following. > > I have component which exposes a service interface with one basic method > which only perform a println. The exporter reads either a queue or a topic > to perform the service call and answers back to consumer to a topic / or > queue depending on the consumer strategy (we use JMSReply concept) > > I have a component consuming the service method using a proxy that connects > the topic or queue for the acknowledgement of the service call and writes > the RPC on the queue or topic to the service. Service call is synchronous. > > I'm using a network of broker with two brokers. > > I'm voluntary connecting the service provider on Broker 1 and the service > consumer on Broker 2 to test the message forwarding in between the two. I'm > performing 2000 service calls. > > Setup 1: If I ONLY configure a full duplex bridge from B1 to B2, the test > passes. > > Setup 2: If I configure two full duplex bridges from B1 to B2 and from B2 > to > B1 (which I don't known if it's recommended), the test passes but I'm > observing network bridges disconnecting and reconnecting during the stress > test. > > Setup 3: If I only configure two half duplex bridge from B1 to B2 and from > B2 to B1, the test fails after some calls. I see also my network bridge > disconnecting and reconnecting during the stress test > > Do you have an idea of what is going wrong ? > > Tks for your help ! > denis > > > > > -- > Sent from: > http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html >
