Re: IPv6 linklocal address on bridged interfaces

Michael Richardson scrisse:
 
> >>>>> "Luca" == Luca BRUNO <[email protected]> writes:
>     Luca> Hi all, I'm currently scratching my head with this
>     Luca> doubt/curiosity: why do Linux bridged interfaces have an
>     Luca> ipv6 linklocal address? is it fine?
> 
>     Luca> In particular, I'm currently experiencing troubles in such a
>     Luca> scenario with two virtual machines bridged to a virtual
>     Luca> switch.
> 
>     Luca>  [VM-A] ---- br0 ---- (virtual switch) --- br1 --- [VM-B]
> 
>     Luca> I can see that after upping brX, the interface gets a ipv6
>     Luca> linklocal address.
> 
> What's a virtual switch?
> Do you mean a router? ("layer-3 switch")

To give more details, the scenario above is composed by two LXC
environments connected through the NS-3 simulator, which is in fact
doing layer 1 and 2 simulation/switching.

From the documentation, it looks like the preferred method to give
connectivity to the lxc containers is to bridge together the guest-tap
with the host-tap; that's why I did it in the first place.
But as it's giving me additional nuisances, I'll try to get rid of it
and glue directly together NS-3 to guest-taps.

> I'd prefer if the member devices of a bridge had their link-local
> addresses removed by the kernel when they get added to the bridge.

With your and Peter's usecase, I see that it may be a saner default.
Indeed, I was just puzzled at first as I was looking for a plain
transparent L2 bridge. 
In the end, I think I may have used used a wrong approach/setup for my
goal, so I'll look to simplify it.

Cheers, Luca

-- 
 .''`.  ** Debian GNU/Linux **  | Luca Bruno (kaeso)
: :'  :   The Universal O.S.    | lucab (AT) debian.org
`. `'`  			| GPG Key ID: 3BFB9FB3
  `-     http://www.debian.org 	| Debian GNU/Linux Developer

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAk9xcMwACgkQRqobajv7n7M3kACfcJvgs/lnlnQvafqWErLHzU0r
XZoAoKnKAqcx1l2HEj7dfXvwrPn3UoZH
=hUjA
-----END PGP SIGNATURE-----