ISE Switch log:Not responding/Being marked alive

[aehtibarov]

Hello guys. Need your help. We run 4 ISE appliance 2 for admins (Primary, Secondary) and 2 PSN nodes. In switch configured radius-group pointing both nodes. However, switches repeatedly give same error logs:

Dec 13 17:29:28: %RADIUS-4-RADIUS_DEAD: RADIUS server XXXX is not responding.

Dec 13 17:30:28: %RADIUS-4-RADIUS_ALIVE: RADIUS server XXXX is being marked alive.

When using only one i see that both of them work separately. I put ping on both server and no timeout or any cut. I am already losing my mind because of these errors. Waiting for your reply.

Below you can find config file.

[ognyan.totev]

Hi , 

check the system mtu ,if it is not 1500  you will see exact same problem . I have same issue on our deployment and i fix it with system mtu.

[mnagired]

Is this happening on all switches across your network?
If not, was there any config changes done on this specific switch from which this started?
Can you ping the radius server using Vlan 333 as source IP address?
By default, server listens to 1812 and 1813 & 1645 and 1646 ports, Where the default changed?
Any firewall on the path and hope those ports are opened?

Can you remove aaa new-module and reconfigure back, may be Mode of operations is a culprit.