Data can be redistributed evenly across all nodes in the cluster when a node is either added or removed.
For VMware vSAN data redistribution happens is two ways:
1. Automated: when physical disks are between 30% and 80% full and a node is added to the vSAN cluster, a health alert is generated that allows the end-user to execute an automated data rebalancing run. For this VMware uses the term "proactive".
2. Automatic: when physical disks are more than 80% full, vSAN executes an data rebalancing fully automatically, without requiring any user intervention. For this VMware uses the term "reactive".
As data is written, all nodes in the cluster service RF copies even when no VMs are running on the node which ensures data is being distributed evenly across all nodes in the cluster.