Troubleshoot VXLAN, VTEP, and VNI configuration and connectivity

There are a couple different ways to test connectivity for these components. I like to start out by testing connectivity using the GUI first, then command line second. Once these components are configured, you can test connectivity via the vSphere Web Client by navigating to Networking & Security -> Logical Switches and double click the logical switch to test connectivity.
From there, click on the Monitor tab and then then select Hosts -> Ping. you will be presented with the following screen. Select both the Source and Destination host and leave VXLAN standard as the size of the test packet. By default VXLAN needs a size of 1600 MTU. Go ahead and click start test, and the test will send and receive 3 packets and display the latency and whether or not the test was successful.
You can also use the command line to test connectivity. Log into the two hosts that you would like to test connectivity to and from via SSH, and run the following commands. the -s will specify the size of the packet, remember VXLAN requires 1600 MTU so we will test a jumbo frame slightly larger than 1500 MTU (You can’t do 1600 as the header information is about 28k). You can find all of your vmkernel information under Networking & Security -> Installation -> Logical Network Preparation. Notice I have specified the interface vmk3.

~ # vmkping ++netstack=vxlan -s 1570 -d -I vmk3
PING ( 1570 data bytes
1578 bytes from icmp_seq=0 ttl=64 time=2.017 ms
1578 bytes from icmp_seq=1 ttl=64 time=3.062 ms
1578 bytes from icmp_seq=2 ttl=64 time=0.962 ms

As you can see, my communication is successful. If the communication between the two hosts is unsuccessful you need to start troubleshooting why. Again, it’s important to remember that you need to have jumbo frames enabled end to end with a size of 1600 minimum. You can confirm that this is the issue if you run a vmkping with a smaller packet size, as shown below. If the below ping works with the smaller (1470) size, but not 1570, then you have a jumbo frame issue.

~ # vmkping ++netstack=vxlan -s 1470 -d -I vmk3
PING ( 1470 data bytes
1478 bytes from icmp_seq=0 ttl=64 time=1.258 ms
1478 bytes from icmp_seq=1 ttl=64 time=1.021 ms
1478 bytes from icmp_seq=2 ttl=64 time=1.420 ms

You also want to ensure that the VLAN is configured correctly. As show in the previous image under Networking & Security -> Installation -> Logical Network Preparation you will see my VLAN is 0. Make sure that they are both on the same VLAN, and that the physical network is trunking those VLANs. If everything looks good ok here, check out the following articles to ensure that everything is configured correctly.
Manage and Report on a Logical Switch using CLI
Manage and Report on NSX installation using CLI
Manage and report on NSX infrastructure using CLI


  1. haris minanda -  December 1, 2016 - 2:21 am 541

    hi, I found funny result while testing my VTEP.
    i can ping using 1570 and 1470 size.
    what is it mean ?
    is my physical already jumbo frame or not ?

    • Sean Whitney -  February 20, 2017 - 7:57 am 561

      Hi Haris,

      Yes, that means jumbo frames are already enabled for you, just make sure the ping is traversing the phsyical network, and not just same host same network, otherwise just the VDS has jumbo frames enabled.


  2. Michal -  March 1, 2018 - 7:04 am 633

    Hi Sean,

    I have a problem with jumbo frames problable, becasue I have no communication between two hosts

    [root@DualEduSVR01:~] ping ++netstack=vxlan -d -s 1470 -I vmk3
    PING ( 1470 data bytes
    1478 bytes from icmp_seq=0 ttl=64 time=0.805 ms
    1478 bytes from icmp_seq=1 ttl=64 time=0.381 ms
    1478 bytes from icmp_seq=2 ttl=64 time=0.379 ms

    — ping statistics —
    3 packets transmitted, 3 packets received, 0% packet loss
    round-trip min/avg/max = 0.379/0.522/0.805 ms
    [root@DualEduSVR01:~] ping ++netstack=vxlan -d -s 1570 -I vmk3
    PING ( 1570 data bytes

    — ping statistics —
    3 packets transmitted, 0 packets received, 100% packet loss

    How should I solve this issue? Is it on the swithc site problem or on the vDS site?

    • Sean Whitney -  March 23, 2018 - 2:04 pm 637

      Hi Michal, the issue is most likely on the physical infrastructure. You will want 1600 MTU end to end from the VTEPs. What is the physical topology between the ESXi host VTEPs?

      • Raghav -  March 18, 2020 - 3:42 am 692

        Hi Sean,

        I cannot ping using 1470 and 1570 between the hosts. vmkernel port MTU is already set to 1600.

        Do I need to enable jumbo frames on physical switch to make it work?

        • Sean Whitney -  May 9, 2020 - 8:58 am 693

          Hi Raghav – Yes, you need to enable it on the physical switches, end to end from the VTEPs/TEPs.


Leave A Comment

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

Back to Top