[c-nsp] IP Input high CPU utilization

[Rodney Dunn]

David,

Punting to the RSP to do CEF switching is one thing. Process switching
is a much larger problem. What is that route-map doing?

Remember that the switching path also depends on the egress
interface for the traffic. 

And it can get a bit more complicated with subints as we don't
have per subint CEF flags. If it's punt on one it's punt on all on
the 75xx.

Rodney

On Wed, Jan 24, 2007 at 02:32:41PM -0500, David Coulson wrote:
> This has been bugging me for a while, and I thought it had it cracked -
> But no, I don't.
> 
> I have two 7507s with PA-FEs on VIP2s and GEIPs. On one router I have my
> vlan 100 on a FE (core1), on the other it is on g 4/0/0.100 (core2).
> Life is good. Both routers run <5% CPU all the time. core1 is generally
> the active router in the HSRP pair.
> 
> I set core2 to be the active router, and life continued to be good. CPU
> usage didn't change at all, and everything was smooth. I moved the
> config off the FE onto g 4/0/0.100 on core1, and again, things kept
> working and CPU usage was low. At this point core2 is still the active
> router in the configuration.
> 
> So, I decided to make core1 the active router in HSRP. It started to
> utilize 100% cpu on core1 due to the 'IP Input' process. I know this is
> caused by process switched packets, however since the config for the
> interfaces is the same on both routers, I can't figure out why it would
> behave differently on the two routers. For all intents and purposes,
> these two routers are identical (same IOS, same RSPs, same VIP/PAs).
> Even when I have HSRP move onto core2, we still have outbound traffic to
> that interface on core1, and traffic which would previously have been
> routed via core1 (e.g. directly connected customers) will come in on
> another GigE interface sub-interface and end up on core1 anyway.
> 
> Is my best bet to turn it up on core1, clear the counters, and watch
> 'show int switching' for a while and see why it is process switching the
> packets?
> 
> This is my config for the interface. Seems pretty basic and there isn't
> anything in there that I know causes the packets to be process switched.
> 
> interface GigabitEthernet4/0/0.100
>  encapsulation dot1Q 100
>  ip address a.b.c.d 255.255.255.0
>  no ip proxy-arp
>  ip ospf cost 100
>  ip ospf priority 60
>  ip policy route-map in-vlan100
>  no cdp enable
>  standby 10 ip a.b.c.e
>  standby 10 priority 110
>  standby 10 preempt
> end
> 
> The config for the FE is identical, except for the encapsulation command
> - It comes in untagged on the FE.
> 
> Anyone got a clue that might point me in the right direction? Is there a
> debugging method available to figure out why the router would process
> switch something?
> 
> 
> 
> David
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/