VMware SD-WAN Edge may experience a Dataplane Service failure and service restart when moving a USB modem from one USB port to another

Products

• VMware SD-WAN by VeloCloud

Issue/Introduction

• This KB article documents a known issue impacting all Edge models when moving a USB modem (hot-cut) from one USB port to another. 
• The issue is being tracked under ID #69194

Symptoms:

A VMware SD-WAN Edge Dataplane Service failure is observed when moving a USB modem (link up) from one USB port on an Edge to a different USB port. This is seen in a hot-cut scenario when a live USB link connected to the Edge is moved to another USB port on the same Edge.

This behavior can be observed on all Edge models meeting the above criteria.   

Environment

VMware SD-WAN by VeloCloud

Cause

• According to the core file associated with this crash, USB ports where incorrectly being bound to the DPDK AF_PACKET drivers. This driver does not support port removal and can eventually cause the Edge dataplane service to crash.
• A core file is generated every time a USB modem is moved from one USB port to a different USB port. 

Resolution

The problem has been addressed through internal ticket ID #69194.

The fix for this issue is found on the following VMware Edge releases:

• On release 4.2.2, this fix is included with the R422-20220419-GA which was released on 04-20-2022 and all later Edge builds in this release train.
• On release 4.3.1, this fix is included with the 4.3.1 Edge release R431-20211208-GA and all subsequent releases like 4.5.0 and 5.0.0.0. 

Workaround:
No workaround is available.

On an Edge not using software with the fix for #69194, any move of a live USB modem from one USB port to a different USB port should only be done in a maintenance window that minimizes the potential impact of a disruption of customer traffic. 

If the Edge is powered off prior to moving a USB modem to a different USB port, this would avoid the Dataplane Service failure and resulting system restart, but will result in additional downtime while the Edge is powered off and before it fully comes back from boot up after being powered on again.

Additional Information

Impact/Risks:
This issue can cause an Edge to experience a Dataplane Service failure with a resulting restart. An Edge Service restart results in a brief service disruption while the Edge's Dataplane Service recovers. The Edge also generates a new core file.