Thank you for your answer. I'll try to make my topology thanks your advise. Anyway I've another question. Is there a dynamic way to configure network connector ? The only configuration way I found was editing xml configuration files (so a static configuration since you need to reboot activemq). Do I make a mistake ?
bsnyder wrote: > > On Mon, Mar 9, 2009 at 9:40 AM, mffrench <mffre...@axway.com> wrote: >> >> Hello, >> >> I would like to know if it's possible to configure ActiveMQ to route >> messages from a broker A to a broker C through a broker B. In fact I aim >> to >> achieve this kind of topology : >> >> client X (network 1) send Message to Broker C via Broker A -> broker A >> (network 1) send message to broker C via broker B -> broker B (network 1 >> and >> network 2) send message to broker C-> broker C (network 2) <- client Y >> (network 2) receive the message from client X >> >> The broker B in fact is a kind of hub broker which links network 1 and >> network 2. Client X is connected only on Broker A and can not send >> message >> to network 2 without connecting to broker B. After reading the broker >> network documentation I do not find the way to achieve this kind of >> topology. Any idea ? > > I believe the topology you're describing is what I've outlined below: > > ClientX --> BrokerA <--> BrokerB <-->BrokerC <-- ClientY > > This is easily achievable via a network of brokers in ActiveMQ and > something I demonstrate on a single machine quite often. (If you're > running on a single machine, you just need to make sure to change all > the necessary ports in the three brokers.) To achieve such a broker > network, you need to define network connectors between the brokers: > > http://activemq.apache.org/networks-of-brokers.html > > Network connectors can be uni-directional or bi-directional based on > the value of the duplex attribute on the connector. Setting duplex to > true simplifies the network definition a bit because only one network > connector needs to be defined between two brokers (one with > duplex=true) instead of two (one in each direction, i.e., brokerA --> > brokerB and brokerB --> brokerA). > > Bear in mind that messages will not be broadcast around the network. > You may already know this but I've found it's worth mentioning. > Messages sent to BrokerA will stay there until there is demand for > them on another broker in the network. Which leads to the correct > configuration of the network connector to allow messages to make hops > between brokers. > > To allow messages to hop between brokers you need to be aware of is > the networkTTL attribute on the network connector. This is what allows > messages to make hops between brokers. By default, the networkTTL is > set to 1. This means that messages can make one hop between brokers. > With this default networkTTL value in your topology, messages sent to > BrokerA would wind up on BrokerB and stay there. To allow them to flow > from BrokerA to BrokerB to BrokerC, you will need to set the > networkTTL to 2. This will allow messages to hop from BrokerA to > BrokerB to BrokerC. > > Hopefully that helps. > > Bruce > -- > perl -e 'print > unpack("u30","D0G)u8...@4vyy9&5R\"F)R=6-E+G-N>61E<D\!G;6%I;\"YC;VT*" > );' > > Apache ActiveMQ - http://activemq.apache.org/ > Apache Camel - http://camel.apache.org/ > Apache ServiceMix - http://servicemix.apache.org/ > > Blog: http://bruceblog.org/ > > -- View this message in context: http://www.nabble.com/Question-about-Active-MQ-network-of-brokers-tp22413530p22430453.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
