3560 High CPU

[RB117892]

We have a older 3560 access switch that has high CPU usage the past three days. Solarwinds is telling us the switch is constantly around 90% - 95%. The switch has been around in our environment quite some time and this is the first we have seen this issue with it. So maybe it's just dieing a slow death. When I do a sho proc cpu s | e 0.00, I see the following:

         

Any ideas or thoughts on some of these processes?

Thanks. 

[daniel.dib]

It's well known that the Hulc process can use CPU, it's a cosmetic bug as I remember it.

However the amount of CPU used for CDP and STP is worrying to me. Do you think there could be a bridging loop? Are your interfaces using more bandwidth than usual? Do you see a spike in broadcasts on the interfaces?

Can you also post show proc cpu hist. If you have support I would open a TAC case.

Daniel Dib
CCIE #37149

[RB117892]

From looking at my Solarwinds info, it does not appear the interfaces are using much for bandwidth or a spike in broadcasts. Here is a history of the CPU usage from Solarwinds as the problem has been going on for 3 days and the command on the cli only shows 60 mins from what I am seeing. I'd open a case with TAC but we no longer carry support on this device.

[RB117892]

Everything appears to be back to normal for the CPU load on the switch. Didn't do anything on our end, seems to have resolved itself.

[Leo Laohoo]

That's an old IOS you've got there.  If you can, upgrade to 12.2(55)SE8.