When designing the management network, single points of failure should be minimized. Redundant network connections and multiple network interfaces connected to separate physical switches should be configured to provide connectivity.

The storage that hosts the management components should be configured to support the capacity and performance of the management components. The storage should also be configured to be highly available so that a disk or path failure does not interrupt management operations.

In environments where the vCenter Server provides provisioning, such as a virtual desktop or self-service cloud environments, vCenter uptime is critical.

If the vCenter Server is running on a virtual machine, it can be protected with HA. If the host that vCenter is running on or the operating system crashes, the vCenter Server is restarted on a surviving host. There will be some downtime associated with the failure, but when they are designed correctly, the vCenter Server services will be quickly restored.

Sufficient resources should be dedicated to the vCenter Server and its components. We discussed the correct sizing of the vCenter Server earlier in this chapter. Sizing vCenter correctly and reserving resources ensures not only the performance, but also the availability. If a virtual machine is running on the same host as the vCenter Server or one of its components and it consumes too many of the host resources, it may impact the performance and availability of the vCenter Server services. Applying resource reservations to the vCenter Server will prevent resource contention.

Another means of preventing resource contention is to design a separate cluster to host the management components. Management cluster design will be discussed in the Designing a separate management cluster recipe.