[c-nsp] bgp holdtime adjustment

[Keegan Holley]

On Tue, Nov 9, 2010 at 8:34 PM, Chris Evans <chrisccnpspam2 at gmail.com>wrote:

> Is the port going down? Or it stays up during this failure.  Different
> failures require different mechanisms...
>

Correct, basically if you have to wait for the peer to timeout because there
was no interface state transition it takes 180s.  Also if you are peered via
loopback (admittedly strange for a provider link) the same will happen.  It
is actually recommended to have much shorter timers because of this fact.
 Try changing them to 10/30 or 15/45.  Also, with BGP the timers are
negotiated to the lowest value changing them should bring your upstream peer
along for the ride.

On Nov 9, 2010 8:00 PM, "John Elliot" <johnelliot67 at hotmail.com> wrote:
> >
> > Hi,
> >
> > We have 2 upstreams, in two seperate locations(both only advertising
> default route to us) - both of our edge routers have a floating static
> default to the other edge should we lose bgp - The problem we have is that
> if we lose one upstream, it is taking 180sec for that router to lose the
> bgp
> default route(So all connections terminating on that router are down for 3
> minutes), and revert to floating static.
> >
> > Is adjusting the neighbour timers someting to consider to reduce this
> downtime?
> >
> > Thanks in advance.
> > _______________________________________________
> > 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/
> _______________________________________________
> 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/
>
>
>