xen DomU can't connect to network,

c-- · 07-22-2009, 10:20 PM

I don't know much about redhat/trixbox system configuration, so I'm hoping I've missed something simple here, but googling my errors hasn't gotten me anywhere.

I've got a trixbox domain running under xen on ubuntu 9.04 jaunty and the networking is behaving very strangely. On boot, it is unable to get a dhcp lease from the external network and gives me some errors about "device not found" for the nics. However, the interfaces appear to be set up correctly. I see the vif interfaces on the domo, which are bridged to the physical interface of the system, and the corresponding eth0 and eth1 are active in the domu.

So the interface doesn't have an address, when I run dhclient manually I get:

Code:

DHCPREQUEST on eth1 to 255.255.255.255 port 67
send_packet: Message too long

ifconfig in the domu shows:

Code:

eth0      Link encap:Ethernet  HWaddr 00:16:3E:2E:89:F2  
          UP BROADCAST RUNNING MULTICAST  MTU:0  Metric:1
          RX packets:541 errors:0 dropped:0 overruns:0 frame:0
          TX packets:3 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:82556 (80.6 KiB)  TX bytes:726 (726.0 b)

eth1      Link encap:Ethernet  HWaddr AA:13:37:00:00:01  
          UP BROADCAST RUNNING MULTICAST  MTU:0  Metric:1
          RX packets:529 errors:0 dropped:0 overruns:0 frame:0
          TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:78872 (77.0 KiB)  TX bytes:1452 (1.4 KiB)

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:322 errors:0 dropped:0 overruns:0 frame:0
          TX packets:322 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:28964 (28.2 KiB)  TX bytes:28964 (28.2 KiB)

Not surprising, the MTU is 0. so I run "ifconfig eth0 mtu 1500". This is where things get more frustrating. Now dhclient outputs:

Code:

Listening on LPF/eth1/aa:13:37:00:00:01
Sending on   LPF/eth1/aa:13:37:00:00:01
Listening on LPF/eth0/00:16:3e:2e:89:f2
Sending on   LPF/eth0/00:16:3e:2e:89:f2
Sending on   Socket/fallback
DHCPREQUEST on eth1 to 255.255.255.255 port 67
send_packet: Message too long
DHCPREQUEST on eth0 to 255.255.255.255 port 67
DHCPNAK from 192.168.61.1
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 6
send_packet: Network is down
DHCPREQUEST on eth1 to 255.255.255.255 port 67
send_packet: Message too long
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 15
DHCPOFFER from 192.168.61.1
DHCPREQUEST on eth0 to 255.255.255.255 port 67
DHCPACK from 192.168.61.1
RTNETLINK answers: Network is unreachable
bound to 192.168.61.24 -- renewal in 121772 seconds.

Somehow, the network is down/unreachable, and I manage to get a valid dhcp lease from the network, and ifconfig shows exactly what it did in the first place: no address, I'm offline, mtu is 0

ifconfig on the domo shows:

Code:

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:12902 errors:0 dropped:0 overruns:0 frame:0
          TX packets:12902 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:3133256 (2.9 MB)  TX bytes:3133256 (2.9 MB)

peth0     Link encap:Ethernet  HWaddr 00:01:29:50:18:21  
          inet6 addr: fe80::201:29ff:fe50:1821/64 Scope:Link
          UP BROADCAST RUNNING PROMISC MULTICAST  MTU:1500  Metric:1
          RX packets:2304699 errors:0 dropped:0 overruns:0 frame:0
          TX packets:1867813 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:1057298288 (1008.3 MB)  TX bytes:485823089 (463.3 MB)
          Base address:0xa000 Memory:fc000000-fc020000 

vif15.0   Link encap:Ethernet  HWaddr fe:ff:ff:ff:ff:ff  
          inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
          UP BROADCAST RUNNING PROMISC MULTICAST  MTU:1500  Metric:1
          RX packets:13 errors:0 dropped:0 overruns:0 frame:0
          TX packets:659 errors:0 dropped:94 overruns:0 carrier:0
          collisions:0 txqueuelen:32 
          RX bytes:2352 (2.2 KB)  TX bytes:100149 (97.8 KB)

vif15.1   Link encap:Ethernet  HWaddr fe:ff:ff:ff:ff:ff  
          inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
          UP BROADCAST RUNNING PROMISC MULTICAST  MTU:1500  Metric:1
          RX packets:10 errors:0 dropped:0 overruns:0 frame:0
          TX packets:665 errors:0 dropped:94 overruns:0 carrier:0
          collisions:0 txqueuelen:32 
          RX bytes:1836 (1.7 KB)  TX bytes:100963 (98.5 KB)

and "brctl show" on the domo

Code:

eth0            8000.000129501821       no              peth0
                                                        vif15.0
                                                        vif15.1

and the xen config itself for the trixbox domain:

Code:

name = "trixbox"
memory = "512"
disk = [ 'tap:aio:/home/xen/trixbox.img,xvda,w', ]
vif = [ 'bridge=xenbr0', 'mac=aa:13:37:00:00:01' ]
bootloader="/usr/bin/pygrub"
vcpus=1
on_reboot = 'restart'
on_crash = 'restart'

dhcp = "dhcp"

If anyone can even point me in the right direction I'd be grateful, from everything I've read on xen the config scripts are all doing their jobs. Is this a problem with the network configuration on the trixbox host itself?

Hari Gold · 07-24-2009, 07:21 AM

I think you want to change the line

vif = [ 'bridge=xenbr0', 'mac=aa:13:37:00:00:01' ]

to

vif = [ "bridge=xenbr0,mac=aa:13:37:00:00:01" ]

and restart the vm. This is python syntax, so the line in your config file defines two interfaces (eth0 and eth1). Not one as I think you intended. This may not fix your problem yet, but at least the diagnostics will start to make sense.

c-- · 07-24-2009, 12:24 PM

yeah, I found that mistake, now I've got one interface on the DomU. DHCP still fails every time, but the interface works fine when statically configured, so at this point I think I just need to ditch trixbox.

Hari Gold · 08-02-2009, 04:59 AM

I would be more inclined to suspect dhcp server misconfiguration at this point than suspecting the dhcp client. Run

ps axuww|grep dnsmasq

on Dom0 to see how xen started the server...