Re: ip route-cache

From: Kevin Baumgartner (kbaumgar@xxxxxxxxx)
Date: Tue Sep 19 2000 - 13:44:35 GMT-3


   

Sure but again like you say you have to have the appropriate hardware. And on t
he smaller
routers Netflow switching isn't available. My discussion was assuming that Netf
low switching
wasn't available on the router and that only Fast switching and process switchi
ng was available.

   One of the interesting troubleshooting tips I have found is to sometimes tur
n off fast switching
if trying out a new IOS feature and it doesn't work. The reason being is that t
his feature
may be coded to use process switching rather than fast switching. I have been h
it on this a
few times. May not be all that useful for the lab test but is very useful to kn
ow this when
troubleshooting a customer problem and things just aren't working.

  Kevin

At 08:36 AM 9/19/00 -0700, you wrote:

>However then Netflow switching does become an option assuming the appropriate
hardware.
>
> Kevin Baumgartner <kbaumgar@cisco.com> wrote:
>No that isn't exactly what I am saying. If you are running fast switching
>on a interface and then for example apply a access-list that does tcp ports
>filtering you are in effect turning off fast switching. This is because
>the router has to look into each packet as it goes in or out of this
>interface. So you can't fast switch the packet in this case.
>
>Kevin
>
> >
> > Kevin,
> > Are you saying that fast switching causes access-lists to be bypassed or
> > were you getting at something different?
> > Thanks,
> > George
> >
> > At 06:44 PM 9/13/00 -0700, Kevin Baumgartner wrote:
> > >
> > >All this does is to allow enabling of fast switch
> > >or by doing "no ip route-cache" to disable fast switch and
> > >doing process switching.
> > >
> > > The issue is that some features will not allow fast switchi! ! ng.
> > >For example access-lists. The router need to look into each packet
> > >to see if there is match on the access-list. So in this case fast switchin
g
> > >can't be used.
> > >
> > >With fast switching it looks at the first packet and then
> > >finds out what interface to send it to. Then all other packets who destina
tion
> > >is the same address will be fast switched out to the same interface and
> > >will no have to look in the packet at all. This is much faster than
> > >process switching.
> > >
> > > Also a quick troubleshooting thing to try if
> > >packets are not going out a interface is to turn
> > >off fast switching. Especially if you are try some new
> > >features and then don't work.
> > >
> > > Kevin
> > >
> > >At 08:05 PM 9/13/00 -0500, you wrote:
> > >
> > >>Hi,
> > >>
> > >>Can anyone help me o! ! ut here?
> > >>
> > >>I've been wondering about the aforementioned command. I realize, after
> > >looking at the docs that "ip route-cache" enables fast switch processing b
ut
> > >I'm curious about the parameters and the ability to disable it.
> > >>
> > >>Can anyone provide me with some solid, factual, reasons as to why you wou
ld
> > >want to turn fast switching off or turn autonomous switching on, etc? Also

> > >why can you use it on a Vlan interface in a switch - this one sort of has
me
> > >at the moment?
> > >>
> > >>If this is something basic that I'm overlooking.......well....I was going

> > >to apologize but we all have to take a trip back to the basics sometime do
nt
> > >we.....so I wont :)
> > >>
> > >>Thanks
> > >>
> > >>Jamie
> > >



This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:24:58 GMT-3