This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] Xen 3.0 Xend loses network - eth0 -> veth0?

Hi Keir, and others, I didn't have time last week to further investigate
the problem, but I'm able to do now.

When I do the network scripts by hand, as I already mentioned,
things are going well and the interface and routing configuration are
okay. this means correct routes via veth0 etc. Also eth0 got its
FE:FF:FF:FF:FF:FF correctly.

The arp table looks good actually:

normal situation without xen:
megalosaurus:~# cat /proc/net/arp
IP address       HW type     Flags       HW address            Mask
Device         0x1         0x2         00:01:02:0A:25:0D     *
eth0       0x1         0x0         00:00:00:00:00:00     *
eth0       0x1         0x0         00:00:00:00:00:00     *
eth0       0x1         0x2         00:13:49:10:0B:E0     *

And with Xen and the veth0 interface:
IP address       HW type     Flags       HW address            Mask
Device       0x1         0x0         00:00:00:00:00:00     *
veth0       0x1         0x0         00:00:00:00:00:00     *
veth0         0x1         0x2         00:01:02:0A:25:0D     *
veth0       0x1         0x2         00:13:49:10:0B:E0     *

Looks perfectly right, I don't know what the problem could be actually.
Everyting looks right and they reply to each others arp requests, that
means on ARP level they see each other and reply to each other with the
ARP replies.


Although very unlikely, it seems Keir Fraser stated on Jun 6 that :

> On 5 Jun 2005, at 21:53, Jurgen Stroo wrote:
> >
> > The thing is, after I've got the following configuration it is still
> > -NOT-
> > working:
> >
> > megalosaurus:~# ifconfig -a
> > eth0      Link encap:Ethernet  HWaddr FF:FF:FF:FF:FF:FF
> The MAC address should be FE:FF:FF:FF:FF:FF. The one above is the
> broadcast address. Not sure if it actually matters, but worth fixing
> anyway as it might confuse the bridge.
> I had problems with ARP when I first started testing veth0. It turned
> out that domain0 was not applying the information in ARP packets to the
> ARP cache. Since the ARP cache wasn't kept up to date, I saw very weird
> IP behaviour.
> I fixed that particular bug, but maybe there's another one. It is
> perhaps worth looking at /proc/net/arp and seeing how it compares with
> other machines on your network.
>   -- Keir

Xen-devel mailing list