I have about 20 2960s of the same model running with almost identical configs, but one switch is running with almost exactly twice the 5 minutes CPU load all all the others.
sw-xxx#sho processes cpu sorted CPU utilization for five seconds: 22%/0%; one minute: 21%; five minutes: 21% PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process 108 16106500 2561179 6288 15.49% 15.10% 15.07% 0 Hulc LED Process 146 2863 220 13013 0.79% 0.47% 0.12% 2 SSH Process
It looks like the Hulc LED Process is taking most of the time.
I found this, albeit for the 3750 that states this can be caused by a flapping interface. I looked in the logs, but no interface seems to be dropping.
Does anybody know what can be causing this?