I'm just getting started with autoscaling of VM's. I've read a lot of information on autoscaling but most of it pertains to web/working roles and not so much VM's. Here is where I'm confused about how autoscaling works. Lets say I have 2 VM's setup with autoscaling (VM1, VM2). VM1 is up and I make changes on it from time to time, when VM2 comes on-line I don't see the changes when log directly into the machine. With autoscaling of VM's is it JUST increasing the resources of the "primary" VM (CPU/Mem)? or is it, once the 2nd VM comes up users are split between the machines like Network Load Balance? If it's treated more like NLB, how do I keep the 2nd server "synced" with the 1st, especially since it's offline most of the time? Whenever I make a change to VM1 do I have to manually boot up VM2 and make the same change? If it simply increases the resources of VM1 does that mean I can't shut VM1 down and have my application continue to run on VM2...say for when I'm doing Windows Updates or something?
Thank you for your help on this, I'm trying to figure out how to keep my cost down while still getting the performance I need when necessary.
James