An IP address is required for the vMotion network for each ESXi host in the VxRail cluster. A private address range is acceptable if you decide the vMotion network to be nonroutable. If your plans include the ability to migrate virtual machines outside of the VxRail cluster, that ability must be considered when selecting the IP address scheme.

Starting with VxRail version 7.0.010, you can choose to have the IP addresses assigned automatically during VxRail initial build. You can also choose to manually select the IP addresses for each ESXi host. If the VxRail version is earlier than 7.0.010, auto-assignment method by VxRail is the only option. 

For the auto-assignment method, the IP addresses for VxRail initial build must be contiguous, with the specified range in a sequential order. The IP address range must be large enough to cover the number of ESXi hosts planned for the VxRail cluster. A larger IP address range can be specified to cover for planned expansion. 

If your plans include expanding the VxRail cluster to deploy nodes in more than one physical rack, you can stretch the IP subnet for vMotion between the racks. Or you can use routing services in your data center for a multirack cluster.

• For the IP address auto-assignment method, record the IP address range.

  Network configuration table rows 51 and 52

  Enter the starting and ending IP addresses for vSphere vMotion.

• For the manual assignment method, record the IP addresses.

  Network configuration table rows 53–56

  Enter the IP addresses for vSphere vMotion.

• Enter the subnet mask and gateway. You can use the default gateway that is assigned to the VxRail External Management network. Or, enter a gateway that is dedicated for the vMotion network.

  Network configuration table row 57	Enter the subnet mask for vMotion.
  Network configuration table row 58	Enter the gateway for vMotion.