We are using SwiftMQ routers version 2_1_3 for our SwiftMQ network. Network contains only two routers. Within local network all works fine. Let's call our routers A and B. After installation of these routers in different networks we have met some strange behaviour of SwiftMQ network. Router A uses real IP address [209.118.X.X] that can be accessed directly from Internet. Router B is installed behind firewall and has internal IP address [10.1.1.14]. On firewall we are using NAT (Network Address Translation) for [10.1.1.14]. So, from outside router B is accessed as [64.28.X.X]. RouterA has not any connectors for RouterB. But RouterB has connector to RouterA. I will try to describe this configuration:


[RouterA:209.118.X.X]<-{Internet}<-[64.28.X.X][NAT][RouterB:10.1.1.14]


When we start SwiftMQExplorer at RouterA it reported:


error creating socket connection to 10.1.1.14:4443, message: Host unreachable: no futher information


Question 1: Why RouterA try to connect directly to [10.1.1.14].


Question 2: How RouterA detected internal IP address of RouterB. There is no configuration files at RouterA that contains IP address [10.1.1.14].


From documentation that is available from http://www.swiftmq.com/ I see that for Firewall connection SwiftMQ should use HTTP tunneling. We have reconfigured our network.


[RouterA:209.118.X.X]<-{Internet}<-[HTTP PROXY*][RouterB:10.1.1.14]


*(HTTP PROXY with allowed CONNECT command)


But the same problem still present. There is other strange behavior that we have detected. Our application can communicate over described JMS Network. But after 1 hour of stable working application reported the same error that was described above.


Assumption: We made assumption that there is could be special mechanism for SwiftMQ network that uses "handshake" logic beetween routers. And within this logic RouterB reported own internal IP address [10.1.1.14] to RouterA. Am I right?




Dmitry Kosarev