I'm applying a very simple test deployment to my new AKS test cluster. The deployment has a replica set with two replicas of my pod. My AKS cluster has two empty nodes, I expected that each replica on a different node by default for improved failure tolerance, however, I always see them both scheduled on the same node. Do I really have to manually work with pod anti affinities to get this to work properly? Am I doing something wrong, is there a simple catch I've overlooked?
Asked
Active
Viewed 18 times
