Both the nodes sharing the same FME Flow System Share and FME Flow Database (postgres sql DB) which are on a shared network drive. The two express installs have independent URLs. So, in this scenario, there is no real fault tolerance from load balancer but if one node fails, the FME clients can change the URL to point to the second node's URL
Page 1 / 1
I think that should be fine, but you'd want reverse proxies to be routing to each machine
- proxy1 routes fme1.com to machine1
- proxy2 route fme2.com to machine2
At this point, i'd argue it would be much better to just setup one load balancer and check the status of each machine and if machine1 is down, then the request goes to machine2. Both fme1.com and fme2.com can be pointed there, you could even still set it up that fme1.com goes to machine1 and fme2.com goes to machine2, but if one of the machines goes down, both urls will point to the active machine
Having clients swap between URLs is going to cause headaches, because things will get hard coded, then process will fail, and need to be updated to switch over, etc.