NSX TEPs fail to acquire IP address via DHCP after reboot

Products

• VMware NSX

Issue/Introduction

• After a host reboot, NSX TEP interfaces (vmk10/vmk11) lose their assigned IP addresses.
• In the NSX Manager UI, the Transport Node status shows "Faulty TEP" or "Down."
• VM migration (vMotion) fails for VMs on affected hosts.
• Running esxcfg-vmknic -l shows TEP interfaces with an APIPA address (169.254.x.x) or 0.0.0.0.
• The following error is observed in the NSX UI: "No valid v4 TEP IP for a long time of vmknic on VDS"

Environment

VMware NSX

Cause

This issue occurs when the ESXi host's VMkernel interface fails to receive a DHCPOFFER from the DHCP server during the DORA process. This is typically caused by:

1. DHCP server exhaustion or failure.
2. Firewall rules blocking UDP ports 67/68.
3. Incorrect DHCP Relay (IP Helper) configuration on the physical switches.

Resolution

1. Verify Host Logs: Check /var/log/dhclient.log for repeated discovery attempts without a response: dhclient-uw: DHCPDISCOVER on vmk10 to 255.255.255.255 port 67 interval 10 dhclient-uw: No DHCPOFFERS received.
2. Validate Physical Network: Ensure the TEP VLAN is correctly trunked to the host's physical uplinks (vmnics).
3. Coordinate with Network Team: Investigate the DHCP server and relay agents to ensure they are processing requests from the ESXi host.
4. Best Practice: To prevent future outages caused by DHCP environment issues, Broadcom recommends migrating TEP assignment from DHCP to a Static IP Pool.