ESXi Transport Node TEPs with APIPA addresses and down Geneve tunnels due to missing Transport VLAN in uplink profile

Products

• VMware NSX

Issue/Introduction

ESXi transport nodes configured for Overlay with 'Use DHCP' option for IPv4 Assignment (TEP).
No geneve tunnels at the ESXi are found to be 'UP'
Inspection of the ESXi hosts using esxcfg-vmknic -l reveals that the vmk10 and vmk11 overlay interfaces (TEPs) are holding IP addresses from the APIPA address space (169.254.#.#), indicating they have not received a DHCP lease.

Assigning a static TEP IP to the VMK interfaces also fails to bring the TEP up.

Environment

VMware NSX

Cause

NSX uplink profile was not correctly attached to the Transport Node Profile (TNP)/Host NSX configuration.
Or, more specifically, the Transport VLAN configuration was missing or incorrect within the uplink profile. This prevents the TEP broadcast DHCP discover packets from reaching the gateway where DHCP relay is configured or exchange the BFD packets between Transport nodes to bring the tunnel up.

Resolution

Update the correct Transport VLAN ID in the affected uplink profile.
This will allow the TEP interfaces to successfully broadcast DHCP discover packets, reach the DHCP relay configured gateway, and subsequently obtain valid IP addresses from the DHCP server and also establish the TEP communication between ESXi/Edge Transport Nodes.

Additional Information

To validate if the VXLAN vmkernel interfaces are sending out a broadcast DHCP discover packet, the following commands can be used:

For TEP interface: pktcap-uw --vmk vmk10 -o - | tcpdump-uw -enr -
For Uplink interface (to check for bootp): pktcap-uw --uplink vmnic<> --capture UplinkSndKernel,UplinkRcvKernel -o - | tcpdump-uw -enr - | grep -i "bootp"