Cisco 2960x Switch Marking Both Radius Servers Dead

[Alex Pfeil]

We have an issue where a 2960x is marking our ISE Policy nodes down at 6:00 AM every morning. There may be a handful of other switches doing this as well. It would not make sense for the ISE policy nodes to be down. Has anybody had a similar issue where the policy nodes actually stopped responding or does this sound like a switch bug? I do think it could be ISE related because it happens near the same time every morning. The 2960x is running 15.02a(EX5).

RADIUS-3-ALLDEADSERVER: Group ISE_RADIUS: No active radius servers found. Id 88

[ognyan.totev]

Hello, 

I have similar issues with other model but I fixed it . I noticed that the MTU on the problem switches was less than 1500 after i change it and reboot the switch everything start to work as expected.

[Alex Pfeil]

Thanks for the reply. I actually did some reporting on the ISE policy nodes that are being marked as dead and they are actually going offline at 6:00 AM. The reports show that the server latency spikes to over 5 seconds at that time. Since the latency is checked only 1 every hour, I believe that some type of process is happening at 6:00 AM causing the issue on the Radius servers. I opened a TAC case and I will update this thread when it is resolved.

[Alex Pfeil · Accepted solution]

I ended up rebuilding one of the ISE policy nodes with the correct size hard drive. Over time, we have done some upgrades and the hard drive requirements went from 200 to 300 GB.