In Product Plan
Score0
When I have a machine in a cluster which is stopped but another is started and I start or publish a service I have to wait for the stopped machine to start. I do not want that machine started, that's why I explicitly stopped it. So I am forced to watch a machine start and then stop it again. They share a config store so it shouldn't need to be started in order to read the new service definition when its started back up later.
I believe ESRI is moving to deprecate the idea of clustering machines in AGS. What version of AGS are you currently working with?