Prepare a cluster for VXLAN in NSX

Preparing a cluster for VLAN is very simple, it only takes a few clicks and will install the vibs required for NSX on the hosts in the cluster. The only prerequisite, is that you need to configure an IP Pool that the vmkernel adapters will use for VXLAN communication.
Step 1. Log into the vSphere Web Client and navigate to Networking & Security -> Installation -> Host Preparation
Step 2. Select Install under Installation for the cluster you would like to prepare for VXLAN.
Step 3. When prompted, click Yes
Step 4. Once the cluster shows Ready, click Configure under VXLAN
Step 5. Specify a VLAN (if applicable), MTU (1600 or higher), IP Pool, and Teaming Policy then hit OK
You should see the VXLAN status show Enabled


  1. mokhtar -  May 23, 2015 - 2:55 pm 115

    Hi Sir ,
    here in VXlan creation i think if i put vlan for examble 50 so all Vxlan will encapsulated in side this vlan ?? also physical switch must support this vlan ?

    • Sean Whitney -  May 26, 2015 - 11:27 am 120

      Hi Mokhtar,

      Yes, the VXLAN will be encapsulated inside VLAN 50, and your physical switch(s) must support it end to end!

      Hope this helps!


  2. Rajeev -  February 8, 2016 - 1:36 am 391

    I configured the MTU to be 1600 for the VXLAN & also the MTU size in my physical switches.
    Even after this when i do the ping test for the logical switch it fails.
    What could be the reason for this.

    • Sean Whitney -  February 9, 2016 - 4:20 pm 399

      Can you manually test using the following command?

      From one ESXi host:

      vmkping ++netstack=vxlan -d -s 1573

      If that works you may be hitting a GUI issue. If it doesn’t MTU is not enabled end to end.

  3. Rajeev -  February 11, 2016 - 4:03 am 404

    I manually ping & did the test as you have mentioned,
    It was working fine. I used the VTEP IP as destination to Ping.
    Let me know if this is the right procedure,


Leave A Comment

Your email address will not be published. Required fields are marked (required):

Back to Top