[c-nsp] Redistributing static routes in BGP

[Peter Rathlev]

I'll second Adams point -- BGP will only announce your static route if
it's already in your routing table.

There's a caveat though: The static route might well be in your routing
table even when you'd expect it to have disappeared. A static route
pointing to a host on a connected interface might survive even if the
interface goes down, since a recursive routing lookup might find another
way to the host, like a supernet of the connected net. (Add an interface
to the static route would make this problem go away, but that's not
always an option.)

If what you want to check isn't wether there's a valid route to the next
hop (which would make the static valid), but if the next hop is
"reachable" and a more "pingable"-like sense then you're in different
situation. Depending on your platform, "Reliable Static Routing Backup
Using Object Tracking" (http://tinyurl.com/67pawg) might be worth
reading.

(Or convince Cisco to implement BFD for static routes in regular IOS...)

Regards,
Peter

On Wed, 2008-04-23 at 18:21 +0100, Adam Armstrong wrote:
> Gary Roberton wrote:
> > All
> >
> > I have a static route that I am redistributing into BGP.  However, I only
> > want to redistribute it if the next hop is available.  For example, if I
> > have ip route 88.0.0.0 255.0.0.0 10.35.1.1, I only want to redistribute it
> > if the 10.35.1.1 next hop is still available.  I won't go into detail here
> > unless it looks like I need to.
> >   
> Is this not default behavior already?
> 
> If the next-hop for the static goes away (if an interface went down, for 
> example), the static would be removed and thus the BGP announcement 
> would go away?
> 
> adam.
> _______________________________________________
> 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/