I've noticed that once the xen-guest is running a new ethenet interface is waked UP and it's binded to the internal switch:
[root@localhost ~]# brctl show
bridge name bridge id STP enabled interfaces
xenbr0 8000.feffffffffff no vif1.0
You can even see the MAC of the virtual interface of the guest:
[root@localhost ~]# brctl showmacs xenbr0
port no mac addr is local? ageing timer
1 00:16:3e:71:08:b4 no 6.22
1 fe:ff:ff:ff:ff:ff yes 0.00
But this setup is missing a second interface to bridge with, isn't it?
In fact DHCP request are not being honored with this setup so I've tried to manually add the physical ethernet interface of xen0:
[root@localhost ~]# brctl addif xenbr0 eth0
Having done so, the bridge can actually see both segments, now
[root@localhost ~]# brctl showmacs xenbr0
port no mac addr is local? ageing timer
2 00:0e:a6:ce:7c:0e yes 0.00
2 00:10:db:4d:ab:02 no 0.63
1 00:16:3e:71:08:b4 no 42.57
1 fe:ff:ff:ff:ff:ff yes 0.00
And even ports number make sense:
[root@localhost ~]# brctl showstp xenbr0
xenbr0
bridge id 8000.000ea6ce7c0e
designated root 8000.000ea6ce7c0e
root port 0 path cost 0
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 0.00 bridge forward delay 0.00
ageing time 300.01
hello timer 1.72 tcn timer 0.00
topology change timer 4.02 gc timer 0.06
flags TOPOLOGY_CHANGE TOPOLOGY_CHANGE_DETECTED
eth0 (2)
port id 8002 state forwarding
designated root 8000.000ea6ce7c0e path cost 19
designated bridge 8000.000ea6ce7c0e message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 0 hold timer 0.72
flags
vif1.0 (1)
port id 8001 state forwarding
designated root 8000.000ea6ce7c0e path cost 100
designated bridge 8000.000ea6ce7c0e message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 0 hold timer 0.72
flags
And in fact, now DHCP assignement worked, and the guest got his IP during the install process, we can see my Netscreen DHCP server leases report:
172.20.76.11 00163e7108b4 23 hour 57 minute
So at least layer2 connectivity is working with the outside world, but the guest can not yet mount the NFS share, nor xen0 can ping the xen-guest 172.20.76.11 newly assigned IP address. Nor can other external boxes ping the xen-guest, by the way...
IPTABLES is disabled, and its modules unloaded, and SELinux is also disabled...
Isn't this somewhat wierd?
Thanks...
On 4/7/06, Raul Saura <raul.saura@xxxxxxxxx> wrote:
Hi, again.
I've forgoted to note that I've had to disable SELinux in order to get xend daemon running.
Is this normal?
thank again.
Raúl.
---------- Forwarded message ----------
From: Raul Saura <raul.saura@xxxxxxxxx>
Date: Apr 7, 2006 8:14 PM
Subject: Network problem during FC5 XEN Guest install
To: fedora-xen@xxxxxxxxxxHello,
I'm trying to create a xen-guest install using xenguest-install.py script, following the indications from http://fedoraproject.org/wiki/FedoraXenQuickstartFC5 .
I've had to shrink mi xen0 domain to 128MB in order to allow guest to start install, otherwise I get the "The privileged domain did not balloon!" error-message.
I've exported the FC5 DVD through NFS from the xen0 domain itself hoping it will be visible from the guest during the install process.
Now I'm in the first stage of the anaconda Installer and DHCP IP learning does not seems to work, in fact no network seems to be available from the guest, because when I manually configure an IP, the guest can not see the NFS export either...
I've been tcpduming for a while, and I can not see any traffic from/to the guest. But in fact I don't even known if i should be seeing any traffic from the xen0 eth0 interface, should I?
Any help will be welcomed.
Thanks a lot.
Raul.
