Configuring both ipv4 and ipv6 with ifupdown

Hi,

In sarge, the following /etc/network/interfaces configuration used to
work:

----
auto eth0
iface eth0 inet static
address 192.168.0.1
netmask 255.255.255.0

iface eth0 inet6 static
address 2002:some:addr::1
netmask 64
----

while in etch if fails with:

# invoke-rc.d networking restart
Reconfiguring network interfaces...SIOCADDRT: File exists
Failed to bring up eth0.
done.

I am ready to file a bug report against ifupdown with severity serious
as it is a regression with respect to working configurations in sarge,
though I ask here before in case I miss something.

I have read bugs #360155, #395826 and #168776 though none of them
seems to be the same as this one.

Thanks for your help,
Frédéric Lehobey


--
To UNSUBSCRIBE, email to debian-ipv6-REQUEST.DeleteThis@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster.DeleteThis@lists.debian.org

Hi,

On Thu, Jan 11, 2007 at 04:34:19PM +0000, Jeroen Massar wrote:
> Frederic Lehobey wrote:

> > In sarge, the following /etc/network/interfaces configuration used to
> > work:
> >
> > ----
> > auto eth0
> > iface eth0 inet static
> > address 192.168.0.1
> > netmask 255.255.255.0
> >
> > iface eth0 inet6 static
> > address 2002:some:addr::1
> > netmask 64
>
> I have a couple of hosts where that works perfectly well with separate
> 'iface eth0 inet...' and 'iface eth0 inet6...' lines.

OK (I understand them to be running etch). Thanks.

By the way the configuration that does not work (in etch) has more
IPv4 fields set up than the working configuration in sarge, namely:

broadcast 213.251.134.255
network 213.251.134.0
gateway 213.251.134.254

I am considering this given the kind of problems that come around from
googling for the 'SIOCADDRT: File exists' error I try to understand.

> Also note that RFC1918 and 6to4 (2002::/16) don't mix, as they

(actually the etch host has a real 2001:... address and direct IPv6
connectivity, it is not using 6to4).

> technically can't work, unless you route 6to4 globally of course or are
> using the IPv4 address of another (public) host in the 6to4 address...

Yes. The documentation of the working (and different) sarge setup is
there: http://proxience.net/dokuwiki/doku.php?id=doc:ipv6 (and, by the
way, credits you as a source for inspiration: thanks twice).

> > # invoke-rc.d networking restart
> > Reconfiguring network interfaces...SIOCADDRT: File exists
> > Failed to bring up eth0.
> > done.
>
> Mine work, but I usually use 'ifdown eth0' and 'ifup eth0', never
> invoke-rc.d, can you try it with 'ifdown eth0' and then 'ifup eth0' ?

Actually it is a remote server and very boring to get back on track if
I loose my ssh access. So I am extremely careful in my
experiments. Actually, 'invoke-rc.d networking restart' is a mild
substitute for a reboot (but the problem shows up too at reboot with
the IPv6 address not being enabled).

> Also interesting to do is activating debugging.
>
> As the above in effect only calls:
>
> ifdown -a --exclude=lo || true
> if ifup -a --exclude=lo; then
> log_action_end_msg $?
> else
> log_action_end_msg $?
> fi
>
> try a 'ifdown -a --verbose --exclude=lo' and then a
> 'ifup -a --verbose -exclude=lo' to see what the results are.

Thanks for the hint, I am still investigating.

I have actually workarounds with
up ifconfig eth0 inet6 add 2001:...
or
up ip -6 ...
but I would prefer (with respect to bug reporting) understand the real
cause of the problem.

Thanks,
Frédéric


--
To UNSUBSCRIBE, email to debian-ipv6-REQUEST.DeleteThis@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster.DeleteThis@lists.debian.org