[Gllug] Oddity in IPv6 routing

[John Winters]

Anthony Newman wrote:
> John Winters wrote:
>> I have set up a gateway machine on my LAN (called knight) which uses an 
>> IPv6 tunnel over my normal IPv4 connection to connect to the outside 
>> world.  (ISP is A&A who provide the IPv6 tunnel.)  knight is also 
>> running radvd which means all the other machines on the LAN pick up 
>> global IPv6 addresses and see knight as their IPv6 gateway.  IPv6 
>> forwarding is turned on on knight.
> 
> Unless you have another block of addresses routed over your tunnel, or 
> you bridge your tunnel device with your LAN segment, this is not really 
> expected to work anyway.

Can you expand on this?  Why shouldn't it work the way I now have it set 
up?  To repeat a parallel posting, I have the following configuration:


auto 6in4
iface 6in4 inet6 v4tunnel
          address <my IPv6 prefix>::1
          netmask 124
          endpoint 81.187.81.6
          ttl 64
          up ip link set mtu 1280 dev 6in4
          up ip route add default via <my IPv6 prefix>::2 dev 6in4

and then the rest of the machines on my LAN get global IPv6 addresses 
with the assistance of radvd (that is, a combination of my /64 prefix 
and their MAC addresses).

As far as I can see the two endpoint addresses of the tunnel (xxx::1 and 
  xxx::2) never actually get used for anything, but I've yet to find a 
way of configuring it without specifying some sort of addresses for the 
end points.

The only problem I can see is if one of the machines on my LAN actually 
wanted to send a packet to one of the tunnel end-point addresses then it 
would fail because it would think the address was local rather than on 
the other side of the router.  I can't think why it would want to though.

I do actually have a /48 allocation and could get a second /64 routed 
down the link if I wanted to, but as you say it seems very wasteful.

Cheers,
John
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug