nsx-node-agent pod in failed status after ncp upgrade to 4.2.1

Products

• VMware NSX

Issue/Introduction

• OpenShift enviroment
• After upgraded the NCP version 4.2.1, some nodes (not all) the nsx-node-agent pods (more specifically container NSX-OVS) fails.
• On the Worker node logs we can see log like: 
  
  2025-03-13T12:06:37Z worker-node NSX 1 - [nsx@6876 comp="nsx-container-node" subcomp="NSX-OVS" level="ERROR"] .usr.local.bin.start_ovs Unable to find the system default ethernet connection name

Environment

VMware NSX 4.2.1

 

Cause

The default connection profile used by the OVS uplink interface does not have connection.interface-name property set.

Resolution

This is a known issue impacting VMware NSX.

Workaround:

1. How to check if the OVS uplink interface has connection.interface-name property configured.

   [root@worker-node]# nmcli c
   NAME UUID TYPE DEVICE
   Wired connection 1 <UID_for_Wired_Connection> ethernet ens192
   lo xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx loopback --

   Using the <UID_for_Wired_Connection> we can check the connection.interface-name:

   [root@worker-node]# nmcli c s <UID_for_Wired_Connection> | grep interface-name
   connection.interface-name: --        <<<<<< interface is not setted

2. Using the next command on the worker node command line:
   

   [root@worker-node]# nmcli c m <UID_for_Wired_Connection> connection.interface-name ens192 
   

3. Check again for the new settings
   

   [root@worker-node]# nmcli c s <UID_for_Wired_Connection> | grep interface-name