Skip to content

Supervisor with "NSX + DTGW/VNA"

This section describes the procedures for Troubleshooting Network Services into the VKS Namespace utilizing an "NSX + DTGW/VNA" architecture inside a vSphere environment.

VDS Architecture

Troubleshooting - Pod access down

As described in the Packet Walk - E/W Pod to Pod section, Pod accessing a Pod traverse the following path:

Logical and Physical View

Logical

Physical

Troubleshooting Steps

Step1: Traffic leaves the Source Pod

This one is internal to the K8s Node and should not be blocked.

Step2+3: Cross-Node Encapsulation/Decapsulation + Delivery

  • Validate the ESX host tunnels are UP

    Status Validation: TEP Tunnels

    Navigate to vCenter > Host and Clusters > [your vCenter Cluster] > Configure > Networking > Network Configuration.
    Ensure "Cluster Status" and "Host Status" are "Green", and ESX have at least 1 TEP IP Address:
    NSX Host Preparation Status

    Note: If no workloads have been deployed on logical networks yet, it is normal to have zero tunnels established on the ESX hosts.

  • Validate the ESX host tunnels accept large packets (MTU)

    Status Validation: Jumbo Frames 
    vmkping ++netstack=vxlan <remote-ESX-TEP-IP> -d -s 8900
    How to find remote-ESX-TEP-IP

    remote-ESX-TEP-IP are available on vCenter > Host and Clusters > [your vCenter Cluster] > Configure > Networking > Network Configuration.
    NSX Host Preparation Status

    Output example

    From the ESX CLI:

    [root@esx-01a:~] vmkping ++netstack=vxlan 10.1.3.207 -d -s 8900
    PING 10.1.3.207 (10.1.3.207): 8900 data bytes
    8908 bytes from 10.1.3.207: icmp_seq=0 ttl=64 time=1.234 ms
    8908 bytes from 10.1.3.207: icmp_seq=1 ttl=64 time=1.102 ms
    8908 bytes from 10.1.3.207: icmp_seq=2 ttl=64 time=1.098 ms
    --- 10.1.3.207 ping statistics ---
    3 packets transmitted, 3 packets received, 0% packet loss
    round-trip min/avg/max = 1.098/1.144/1.234 ms
    If vmkping doesn't work, consult with your Network Team to determine why routing does not reach the destination.

    • If standard "small" pings fail, the issue is typically related to routing misconfigurations or firewall blockages.
    • If "large" pings fail, the issue is typically related to Jumbo Frames (MTU) not being enabled across the physical fabric.