[c-nsp] RANCID Spiking CPUs

[Justin M. Streiner]

On Mon, 9 Jun 2008, Nick Davey wrote:

> I've deployed rancid on a fairly large metro network, and am seeing some
> pretty high CPU averages. When RANCID runs the CPU's on a large number of
> our boxes spike to about 95% for several seconds. Although they have never
> hit 100%, or caused any issues (dropped OSPF hello's, stp bpdu's) I'm
> concerned that this could happen under the right combination of events this
> could result is dropped OSPF neighbor adjacency's or other badness.
>
> I've tried to replicate the high CPU issue by pasting the commands in
> manually however I haven't come anywhere close to the 95% I'm seeing when
> RANCID runs them. I'm assuming this is just the frequency at which the
> commands are run. Does anyone have any experience with this or any insight
> they can provide?

Keep in mind that interactive logins and the processes that are needed to 
service those sessions do carry a CPU cycle penalty.  Also, traffic _to_ a 
device, as opposed to through it, is normally process switched,  so that's 
another CPU hit right there.  How quickly the commands are run can 
certainly play into how much the CPU gets beaten up.

Also how are you looking at the CPU spikes?  If you're doing something 
like a "show proc cpu hist", that usually does a pretty good job of 
showing you some of the transient spikes that the 'legacy' CPU utilization 
OID can miss.  From what I recall, the 5 second CPU utilization OID in the 
CISCO-PROCESS-MIB gives results that more closely resemble the command 
above.

jms