[c-nsp] cisco 6500 sup2 high load

[Alex Rubenstein]

On Sun, 3 Oct 2004, Adam KOSA wrote:

> Akiron#sh processes cpu
> CPU utilization for five seconds: 5%/4%; one minute: 46%; five minutes: 75%

If this is a s2/m2, something is wrong already.

The / number should be 0 or 1% at most, even at many gigabits/sec.


> That 75% bothers me, but i don't know the reason.  This never happened
> before.  The only process which has >1% cpu usage is:
>
> PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
> 41     1018884   5980897        170  2.12%  4.78%  6.92%   0 IP Input

Somewhere, you've lost hardware cef. What version of code are you running?


> which makes sense.  I still don't know what to do not to let this happen
> again.  This incident occured several times during the past 48 hours (3-4
> times).  During this mrtg shoved minimal traffic on the gbit link.

NO -- this does not make sense.

esd1.nyc>sho proc c
CPU utilization for five seconds: 7%/1%; one minute: 8%; five minutes: 8%

  69    26907128 199909790        134  0.24%  0.24%  0.21%   0 IP Input

This switch is handling about 3 gb/s as I write this.

What does a 'sho catalyst' show (assuming this is integrated)?






-- Alex Rubenstein, AR97, K2AHR, alex at nac.net, latency, Al Reuben --
--    Net Access Corporation, 800-NET-ME-36, http://www.nac.net   --