The deployment of the vSphere Supervisor in VMware Cloud Foundation (VCF) 9.0 requires a highly available and automated underlying vSphere cluster configuration. Two mandatory prerequisites for enabling the Supervisor are the use of a vSphere Distributed Switch (VDS) and the activation of vSphere HA and Fully Automated DRS.

First, the networking requirement mandates that all ESXi hosts in the cluster are connected to a vSphere Distributed Switch (A). This is necessary because the Supervisor utilizes advanced networking features for its primary components, including the Supervisor Control Plane VMs and the integration with either NSX or the vSphere Distributed Switch-backed networking for workload clusters. Standard Switches (VSS) do not provide the centralized management or the necessary feature set to handle the dynamic networking requirements of Kubernetes pods and services within the SDDC.

Second, the cluster must have vSphere HA and Fully Automated DRS (D) enabled. vSphere HA ensures the resilience of the Supervisor Control Plane VMs; if a host fails, HA restarts the control plane nodes on remaining hosts to maintain the Kubernetes API ' s availability. Fully Automated DRS is critical because the Supervisor ' s lifecycle manager relies on DRS to handle the initial placement and ongoing balancing of the control plane VMs and workload resources. If DRS is set to " Partially Automated " or is disabled, the system cannot automatically ensure that the Kubernetes control plane is optimally placed across the cluster ' s compute resources, leading to potential deployment failures or performance bottlenecks. These requirements ensure that the VCF 9.0 infrastructure provides the enterprise-grade stability required for modern, containerized applications.