On Feb 11, 8:35*pm, Unruh <unruh-s...@physics.ubc.ca> wrote:
> David Schwartz <dav...@webmaster.com> writes:
> >On Feb 11, 5:10=A0pm, Unruh <unruh-s...@physics.ubc.ca> wrote:
> >> No there cannot be more than one. the system uses the last one exclusivel=
> >y.
> >If the system uses the last one exclusively, then there can be more
> >than one. Otherwise, there would be no "last one".
> Lets see 10 years ago I had a default route on my system in California.
> Thus the one I have today is not the only one, since I had one 10 years
> ago. Ok if that is how you want to use the terms, that's fine by me.
> There is only one operative default route at anytime, and that default
> route is the last one you set up.
> Does that make you happy?
That is correct only for a strange definition of "operative". Sure,
only one is operative. The other is a backup.
> >> There is no fallover either.
> >That's not true. The system will not use a default route if its next
> >hop is unreachable. It would be impossible for it to do so.
> It will use it, and tell you that the next hop is unreachable.
No, it will not use it.
> >> Default means default. If nothing else works then use this one. this ONE.
> >Correct. And one of the things that might not work is another default
> >route whose next hop is unreachable.
>
> wormhole[root]>route add default gw 192.168.0.3
> wormhole[root]>ping info
> PING info.physics.ubc.ca (142.103.234.23) 56(84) bytes of data.
> From wormhole1 (192.168.0.2) icmp_seq=3 Destination Host Unreachable
> From wormhole1 (192.168.0.2) icmp_seq=7 Destination Host Unreachable
>
> --- info.physics.ubc.ca ping statistics ---
> 9 packets transmitted, 0 received, +2 errors, 100% packet loss, time 7999ms
>
> (the computer 192.168.0.3 is switched off, but is on a network
> 192.168.0.0 * * 0.0.0.0 * * * * 255.255.255.0 * U * * 10 * * 0 * * * *0 eth1
>
> Ie, that network gateway is down, but the system keeps trying.
Correct, because the next hop is not known unreachable. Pull the
network cable.
> If I unplug that network cable, then I get
> I get the same thing for about 10 sec, at which point that default route is
> disconnected. Ie, I believe it is NOT the network software that is "falling
> over" but that some helper program is noticing that the network is
> disconnected and then disconnecting that default route.
You are extrapolating from a very simple setup where something doesn't
work the way you want it to and concluding that it never works.
If the next hop of a default route is known unreachable, the default
route will not be used. Having more than one default routes provides
failover if you have a way to make the next hop on the failed route
unreachable. You can do this many ways. One of them is having an
external routing protocol that controls the reachability of the next
hop. Another way is to have an external routing protocol that controls
the default route itself.
If you have two equally perfectly good default routes, there is
absolutely no reason not to install both of them.
DS
|