Cisco 2960L management problem

[Petri Kemppainen]

HI all!

We are runinng version Cisco IOS Software, C2960L Software (C2960L-UNIVERSALK9-M), Version 15.2(7)E0a, RELEASE SOFTWARE (fc1) in our Cisco 2960L switches. 

At night time our management connection works fine but when employees comes to work switch management ip starts to response ping packet very slowly or connection is almost lost.. Other place we have huge problems with this because we are using radius authentication for clients. Management connection is almost useless. How i can debug what causes this behavior? Or is there so much traffic that causes this? ipv6 or what can cause this ? switch is almost useless at 8 to 15. Other time there is no problems.

[Arne Bier]

Can you get a console onto the switch to see what the CPU is busy with?

show proc cpu hist
show proc cpu

Perhaps you're doing https redirection for portal based authentication (guest)? That is known to cause a bit of CPU load. A bit of a long shot but worth looking into.

other than that, check the logs of the device when you're experiencing the issue

show logging

[Petri Kemppainen]

There is nothing special in logging. Normal link state information. Also cpu is 30 % and there is no memory problem either. It seems that when we add one vlan where users and windows workstations are causes this problem. Is there some specific debug  command where we can see what happens? Could it be ipv6 which causes this problem because workstations are windows workstations and ipv6 are enabled but not in use. Could this generate problems? 

-Petri

[Arne Bier]

The command, show proc cpu   Will Show the process that’s taking up the most cpu cycles. I don’t know why IPv6 has anything to do with it. If your Cisco device is not configured for IPv6 then it won’t be sending RA‘a or doing anything related to v6

Maybe a TAC case? They know loads more tricks.