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-users] Yet another question about multiple NICs

Felix Kuperjans wrote :
Answers within quotes:

- It's interesting that dom1's firewall output shows that no packages
were processed, so maybe you didn't ping anything since the last reboot
from dom1 or the firewall was loaded by reading it's statistics...
You requested for the outputs "when <my> system has just started".
Hence no packet, I guess. But shouldn't there be at least those exchanged
for the ssh connection to the dom1 ?
Anyway, after one minute or so, I get on the dom1:
# iptables -nvL
Chain INPUT (policy ACCEPT 23 packets, 884 bytes)
 pkts bytes target     prot opt in     out     source

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source

Chain OUTPUT (policy ACCEPT 4 packets, 816 bytes)
 pkts bytes target     prot opt in     out     source
That looks better.

Still no reasons why you can't ping local machines from the dom1 (and
sometimes even not from dom0). Have you tried pinging each other, so
dom0 -> dom1 and vice versa?
Yes I tried, and it has always worked while dom0's eth1 was up.
So it's only impossible to ping the domU from other machines on the
network (and vice versa)?
I think Fajar is probably right with his guess that your physical
switches are managed. That means they do traffic filtering on their
ports based on the mac addresses.
Which switch models do you use on your two networks?

I already answered Fajar in this thread: when the FIRST vif of dom1 is connected to dom0's eth1, then the behaviour on that switched LAN is normal, while the traffic on the routed LAN of dom0's eth0 issues the bug. So my issue is definetely related to the instanciation of the SECOND interface of dom1, whatever network it is connected to. Or there is some kind of black magic underneath...

The only remaining thing that denies communication would be ARP, so the
output of:
# ip neigh show
on both machines *directly after* a ping would be nice (within a few
seconds - use && and a time-terminated ping).
Nothing on a machine when not connected. But when connected (here the
$ ip neigh show dev eth1 lladdr 00:16:36:e0:81:2c REACHABLE dev eth0 lladdr 00:16:38:4c:04:00 DELAY dev eth0 lladdr 00:16:36:e0:81:2e STALE dev eth0 lladdr 00:1b:24:3d:ca:95 REACHABLE dev eth0 lladdr 00:16:38:28:b5:39 REACHABLE
ARP seems to work at least on the Domain-0 *if* one of those IP
addresses is the one of the domU...
Can you try doing this on the DomU when pinging a host in the network?

I did ! As requested ! And as you know, dom0 and dom1 are alternatively connected. When dom0 is connected ( on eth0 and on eth1) I get the trace above, but nothing from dom1. When dom1 is connected, I get a similar trace from dom1, but nothing from dom0 (I mean nothing on network


Xen-users mailing list