On Tue, 2013-12-17 at 20:45 +0100, Tore Anderson wrote:
> * Anders Blomdell
> 
> > Will look into VRRP, but 4.1 from http://tools.ietf.org/html/rfc579 makes me
> > suspicious:
> > 
> >    In the IPv6 case (that is, IPvX is IPv6 everywhere in the figure),
> >    each router has a link-local IPv6 address on the LAN interface (Rtr1
> >    is assigned IPv6 Link-Local A and Rtr2 is assigned IPv6 Link-
> >    Local B), and each host learns a default route from Router
> >    Advertisements through one of the routers (in this example, they all
> >    use Rtr1's IPv6 Link-Local A).
> > 
> > won't NetworkManager pick up those routes as well, and [still] mess up 
> > routing?
> 
> Well, with VRRP, only the master router will send out Router
> Advertisements for the virtual router. The backup(s) will not. See RFC
> 5798 section 6.4.2 (325), 6.4.3 (630), and 8.2.3.
> 
> So from the hosts' point of view, only one default router exists. Even
> the MAC address is a virtual one that fails over, so they don't even
> have to go through NUD.
> 
> I know this works very well with Juniper's VRRPv3 implementation, at
> least. I'm not aware of any open-source VRRPv3 implementation for Linux,
> I'm afraid.
> 
> > Would [nm-policy.c: update_ip6_routing] be a good starting-point for looking
> > into this?
> 
> I don't really know, since I'm not really a NM hacker. My role here is
> to bitch and moan about all things IPv6 in the hopes that the actual NM
> hackers try to shut me up once in a while by actually implementing what
> I'm bitching about. ;-)

And I actually appreciate you doing this.  I hope you keep doing it.

Dan

_______________________________________________
networkmanager-list mailing list
[email protected]
https://mail.gnome.org/mailman/listinfo/networkmanager-list

Reply via email to