Why do we need multiple replicas of nfs controller?

[akagr]

@andyzhangx The cluster I use it on can scale to multiple nodes. Does a controller pod need to run on each node? Or, ies it purely to distribute monitoring and provisioning load? In other words, any blockers to using just 1 controller pod across, say 20 nodes? I'm only tracking a single NFS share that a service deployment with multiple pods uses.