Let’s see if I understand your environment:
- You have many different IS instances
- Each instance hosts a different set of integrations
- Apache is being used as a proxy to direct connections to specific IS instances
Question here: what URL is being used to connect to Apache? Is it the same for all “applications?” How does Apache differentiate the requests so that it knows the proper instance to pass the request to?
Correct me if I’m off base, but you’re looking for 1) virtual host names; and 2) URL translation.
What’s the plan for invoking services on each of these instances? The standard URL is http://hostname:port/invoke/folder.folder/service, where “invoke” is a required literal. How will Apache deal with that?
This seems to me that you may be trying to apply “application server” approaches (e.g. WebLogic has a notion of an explicit app accessed using an URL of the form http://host:port/myapp/somefile.jsp where “myapp” is a required part of the name so that WLS knows what set of “stuff” to run) to something that is not an application server. Integration Server is not an app server like WLS or WebSphere. It doesn’t do things the same way those environments do, though it looks similar.
Masking or “front-ending” WmRoot is going to give you trouble.
Given the information we’ve shared thus far, if you need to segregate traffic using logical application names, then I’d suggest using a scheme of logical hostnames. That would let you have load-balanced clusters per “application” and would permit accessing WmRoot properly in each of those environments.
Please let me know if I’ve misunderstood and am completely off-base here.