NSX Edge interface remains in DOWN state due to driver race condition during link status updates

Products

• VMware NSX

Issue/Introduction

- NSX Edge interface remains DOWN despite the physical or virtual switch showing an UP status.
- Logs indicating "Potential parallel command is sent down to backend" and "link update, read 0x0".

/var/run/log/vmkernel.log
YYYY-MM-DDTHH:MM:SS In(05) vcpu-6 - VMXNET3 user: Ethernet2: Potential parallel command is sent down to backend: lastcmd: 0x####### 
YYYY-MM-DDTHH:MM:SS In(05) vcpu-6 - VMXNET3 user: write get link status cmd                                                           
YYYY-MM-DDTHH:MM:SS In(05) vcpu-6 - VMXNET3 user: Read link status: 0x0      

Environment

VMware NSX

Cause

The root cause is a race condition involving parallel accesses to the command register during link status updates.
When multiple processes attempt to access the register simultaneously, the NSX Edge datapath may receive stale or invalid data, preventing the interface from initializing correctly after a status change.

Resolution

This is scheduled to be addressed in a future version.

The following is the workaround when the issue occurs:

1. Log in to the NSX Edge CLI using the admin credentials.
2. Execute the following commands:

set physical-port fp-ethX state down
set physical-port fp-ethX state up

Workaround to be performed from vCenter Server:

1. Log in to the vCenter Server
2. Locate the affected NSX Edge virtual machine.
3. Navigate to Edit Settings.
4. Uncheck the Connected box for the affected network adapter.
5. Click OK.
6. Re-open Edit Settings and check the Connected box again to re-enable the adapter.