Before you configure an NSX-T transport node using vSphere Distributed Switch (VDS) as a host switch, ensure that the VDS created on a vCenter Server 7.0 or a later version is configured to manage NSX-T traffic.

High-level tasks to configure a cluster or a standalone managed host using a VDS switch.
Important: To create a VDS switch supporting NSX-T networking, the following conditions must be met:
  • vCenter Server 7.0 or a later version
  • ESXi 7.0 or a later version

Prerequisites

  • Verify that ESXi hosts have the required number of physical NICs to meet networking requirements. For example, if you plan to configure teaming policies and remote span port mirroring, ensure that a free physical NIC is available to avoid uplink conflicts.
  • Ensure that the MTU value of the physical switch port or LACP port is set to 1600.

Procedure

  1. In a vCenter Server, create a VDS. For more information about creating a VDS, see the vSphere Networking documentation.
    • Set the MTU value for the VDS to at least 1600.
      Note: On VDS 7.0 or later, the default MTU size is 1500. To prepare a VDS for NSX-T overlay networking, the MTU size of the VDS must be at least 1600. Starting in NSX-T Data Center 3.2.1, if the MTU size of the VDS is below 1600, NSX Manager notifies you that the MTU size will be automatically increased to 1600.
    • Connect the switch to hosts that you want to prepare for NSX-T networking.
    • Assign physical NICs to uplinks on the VDS.
  2. In NSX-T, add an uplink profile that defines a teaming policy mapping NSX-T uplinks with VDS uplinks.
  3. In NSX-T, prepare an ESXi host using VDS as the host switch.

    At the end of the configuration, the host is prepared as NSX-T transport node with VDS as the host switch.

What to do next

Configure the host as a transport node. See Prepare ESXi Cluster Hosts as Transport Nodes.