[c-nsp] Safe to enable CEF on live 3725 production router?

[Rodney Dunn]

On Tue, Aug 01, 2006 at 10:49:38PM +0200, Mikael Abrahamsson wrote:
> On Tue, 1 Aug 2006, Rodney Dunn wrote:
> 
> > If it's route cache (fast switching) it's probably pretty unlikely
> > turning on CEF will fix it although it should still be done.
> 
> Well, actually I'd say the opposite. If you're fast-switching or process 
> switching you'll see much more jitter than if you're CEF switching. 

Process yes fastswitching no. It's impossible except for one thing.
If you are fastswitcing and the cache ager expires the fastcache
entry which would force the first packet to be process switched
you would see some jitter.



That's 
> my experience from troubleshooting QoS issues. Enabling CEF is the 
> preferred way to go if you want proper QoS and predictable behaviour over 
> time from the router.

For the QOS we do recommend CEF just in general but it should work the
same in the fastswitching path.

> 
> First thing to enable on a cisco router should be CEF (if at all 
> possible), and if you have issues with CEF, log a case and get it fixed by 
> new code. Running fast-switching or process switching is not the way out, 
> it's only a stop-gap meassure until you can get the CEF-problem solved.

Yep because in later 12.4T code there will no longer be a fastswitching
path. CEF or process is all there will be.

> -- 
> Mikael Abrahamsson    email: swmike at swm.pp.se
> _______________________________________________
> 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/