NSX Edge BFD tunnels down due to Uplink Profile and physical NIC mismatch in VMware NSX

Products

• VMware NSX

Issue/Introduction

• BFD sessions remain in a DOWN or INIT state on the Edge CLI (verified via get bfd-sessions).
• ICMP pings between Edge TEPs and Host TEPs are successful (L3 connectivity is verified).
• Tunnels to some peers may be operational while others are down, indicating selective packet loss.
• The logical Uplink Profile is configured for multiple uplinks, but the host or Edge node only has a subset of those uplinks physically connected or active.

Environment

VMware NSX

Cause

NSX automatically load-balances Geneve and BFD traffic across all uplinks defined in the active Uplink Profile. When the profile specifies more uplinks than are physically connected, the NSX transport node continues to distribute traffic—including critical BFD heartbeat packets—out of the dead or disconnected interfaces. This results in "black-holed" traffic and failed tunnel establishment for the paths assigned to those inactive links.

Resolution

Reconcile the configuration to match the physical topology of your environment using one of the following methods:

1. Adjust Uplink Profile: Update the Uplink Profile (or the specific transport node override) to include only the physical uplinks that are currently connected and active. This forces all traffic onto the functional links.
2. Restore Physical Connectivity: If the design requires multiple uplinks, investigate and restore the physical switch ports or cabling to ensure all uplinks defined in the profile are active.
3. Verify Status: Once the configuration matches the physical hardware, verify the tunnel status:
   • Access the Edge or Host CLI.
   • Run the command: get bfd-sessions
   • Confirm that the tunnel sessions transition to an UP status.