** Andy Smith <andy@???> [2013-02-10 15:10]:
> On Sun, Feb 10, 2013 at 11:59:30AM +0000, Paul Tansom wrote:
> > I've noticed that my internal Ubuntu servers have odd default routes for and fe80:: address:
> >
> > ::/0 fe80::204:edff:febc:b011 UGDAe 1024 0 1 eth0
>
> It is okay to have default route through a link-local address. You
> would also get one on BitFolk if you didn't do any static routing
> and just let RA set it up for you. As long as your default route is
> reachable and works, does it matter which address on it is used?
I initially started investigating this when I was failing to download updates through aptitude or apt-get, before that I hadn't really noticed. When I 'ping6' the IP address above I get 'connect: Invalid argument', although 'ping6 -I eth0' works fine. All my routing is static on the servers is static so I had to work out where this new route had come from!
> > It seems that this is picked up from my router somehow and I've found that there is a need to use the following:
> >
> > echo 0 > /proc/sys/net/ipv6/conf/default/accept_ra
>
> A need to use that in order to achieve what outcome? Since that will
> disable RA, you shouldn't pick up a default route by RA, yes. Was
> disabling RA what you wanted to do?
My understanding was that this stopped the interface picking up any addressing or routes from the router for IPv6
> > Having looked at the configuration on my Bitfolk server I see the same sort of setup, although more extensive:
> >
> > pre-up echo "/sbin/modprobe ipv6" && /sbin/modprobe ipv6 || true
> > post-up echo "/proc/sys/net/ipv6/conf/default/forwarding=0" && echo 0 > /proc/sys/net/ipv6/conf/default/forwarding || true
> > post-up echo "/proc/sys/net/ipv6/conf/all/forwarding=0" && echo 0 > /proc/sys/net/ipv6/conf/all/forwarding || true
> > post-up echo "/proc/sys/net/ipv6/conf/default/accept_ra=0" && echo 0 > /proc/sys/net/ipv6/conf/default/accept_ra || true
> > post-up echo "/proc/sys/net/ipv6/conf/all/accept_ra=0" && echo 0 > /proc/sys/net/ipv6/conf/all/accept_ra || true
> > post-up echo "/proc/sys/net/ipv6/conf/$IFACE/accept_ra=0" && echo 0 > /proc/sys/net/ipv6/conf/$IFACE/accept_ra || true
> > post-up echo "/proc/sys/net/ipv6/conf/default/autoconf=0" && echo 0 > /proc/sys/net/ipv6/conf/default/autoconf || true
> > post-up echo "/proc/sys/net/ipv6/conf/all/autoconf=0" && echo 0 > /proc/sys/net/ipv6/conf/all/autoconf || true
> > post-up echo "/proc/sys/net/ipv6/conf/$IFACE/autoconf=0" && echo 0 > /proc/sys/net/ipv6/conf/$IFACE/autoconf || true
> >
> > I'm curious about two things:
> >
> > 1. why post-up and not pre-up
>
> If you do it early, the changing interface state can cause the
> settings to change again anyway.
>
> > I've tried the single line in pre-up and post-up and the full set of commands above. On one occasion I did loose the default route to the fe80:: address, but generally I don't (with no configuration change!). This is on two servers, one 10.04 and the other 12.04.
>
> So you're saying you have all the above lines in your
> /etc/network/interfaces but still end up with a default route to an
> fe80 address set up by RA?
Yup.
** end quote [Andy Smith]
--
Paul Tansom | Aptanet Ltd. |
http://www.aptanet.com/ | 023 9238 0001
======================================================================
Registered in England | Company No: 4905028 | Registered Office:
Crawford House, Hambledon Road, Denmead, Waterlooville, Hants, PO7 6NU
