I'm working on a web application that will use Rhino Service Bus to send messages that are then consumed by a windows service on the app server. I've been able to test this on my machine (hosting the web app and the windows service) and it works fine. I was also able to test this in our dev environment, which has one web server and one app server, without any problems. However, our staging environment has two web servers and two app servers, so I'm not sure how to configure the endpoint to which the messages are sent.
I know I can edit the config section for each web server to point to one of the app servers. I can also put the windows service on only one machine and send everything to a queue on that machine. Neither of these sounds like a good option. What's the best practice in a scenario like this?
Any help would be appreciated.
It depends on which transport you're using. If you're using Rhino.Queues you can leverage hardware based load balancing + DNS. If you're using MSMQ, then you would need to use the MSMQ load balancer in RSB. You can find tests in the source that demonstrate this. Your workarounds that you mentioned would also work.