Strange behaviour on M4300 stack with RADIUS authenticated ports

Hello,

I have M4300 based switch stack that is configured to perform dot1x and MAC authentication on client ports.

There are 2 RADIUS (Windows NPS) servers configured. One is set as a primary.

The same two servers are also setup as RADIUS accounting servers.

In the RADIUS statistics, the primary server shows requests and no errors. The secondary server shows fewer access requests and quite a lot of timeouts

When I look at the RADIUS (NPS) event log on the secondary server I see lots of event ID 16

A RADIUS message with the Code field set to 4, which is not valid, was received on port 1812 from RADIUS client xxxxxxxxxxxxxxxxxxxxx. Valid values of the RADIUS Code field are documented in RFC 2865.

I have used WireShark to capture the traffic. I see that the switch stack is sending RADIUS accounting messages to port 1812 rather than the correct accounting port 1813. Also the messages are sent to the secondary RADIUS server

The RADIUS server accounting setup on the switch stack shows as correct  with the port as 1813

In the switch statistics for RADIUS account is shows failed accounting messages being sent to the primary RADIUS server.

This must be a software bug

Has anyone else experienced this?? And perhaps found a fix??

Firmware version 12.0.17.19