[Users] HN eth link down in bridged environment

[Matei Conovici]

Hello,

I am using debian 2.6.32-5-openvz-amd64 (2.6.32-21, 2.6.32-23) with
bridged veth networking.
When the veth devices are attached to the bridge, the following happens:

[  825.400527] CT: 808: started
[  825.491506] device veth808.0 entered promiscuous mode
[  825.491577] vzbr0: port 6(veth808.0) entering forwarding state
[  825.940139] e1000e: eth0 NIC Link is Down
[  825.960899] vzbr0: port 1(eth0) entering disabled state

then, 300 seconds later (it is always 300s),

[ 1127.620861] e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow
Control: None
[ 1127.620922] e1000e 0000:00:19.0: eth0: 10/100 speed: disabling TSO
[ 1127.621870] vzbr0: port 1(eth0) entering forwarding state

and everything works for a while. Then the eth0 link again goes down
for 300 seconds, a.s.o.

[ 1866.448119] e1000e: eth0 NIC Link is Down
[ 1866.465934] vzbr0: port 1(eth0) entering disabled state
[ 2168.196858] e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow
Control: None
[ 2168.196902] e1000e 0000:00:19.0: eth0: 10/100 speed: disabling TSO
[ 2168.197861] vzbr0: port 1(eth0) entering forwarding state
[ 2845.696123] e1000e: eth0 NIC Link is Down
[ 2845.714422] vzbr0: port 1(eth0) entering disabled state
[ 3147.340854] e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow
Control: None
[ 3147.340899] e1000e 0000:00:19.0: eth0: 10/100 speed: disabling TSO
[ 3147.341815] vzbr0: port 1(eth0) entering forwarding state

If I don't add the veth devices to the bridge (e.g. don't start the
containers), I have no problems.
I've also updated the e1000e driver from 1.0.2-k2 to 1.2.10-NAPI and
this didn't make a difference.
eth0 is a:

00:19.0 Ethernet controller: Intel Corporation 82566DM-2 Gigabit
Network Connection (rev 02)

There is currently no firewall on the HN.

bau:~# brctl show vzbr0
bridge name	bridge id		STP enabled	interfaces
vzbr0		8000.0018512d2cf8	no		eth0
							veth103.0
							veth401.0
							veth801.0
							veth804.0
							veth808.0

bau:~# ifconfig
[...]
eth0      Link encap:Ethernet  HWaddr 00:1c:c0:4d:25:53
          inet6 addr: fe80::21c:c0ff:fe4d:2553/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:19098 errors:0 dropped:0 overruns:0 frame:0
          TX packets:14945 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:1904150 (1.8 MiB)  TX bytes:2640343 (2.5 MiB)
          Interrupt:20 Memory:e0400000-e0420000

venet0    Link encap:UNSPEC  HWaddr
00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
          UP BROADCAST POINTOPOINT RUNNING NOARP  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

veth103.0 Link encap:Ethernet  HWaddr 00:18:51:a1:b6:0b
          inet6 addr: fe80::218:51ff:fea1:b60b/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:180 errors:0 dropped:0 overruns:0 frame:0
          TX packets:6892 errors:0 dropped:148 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:6576 (6.4 KiB)  TX bytes:271195 (264.8 KiB)

veth401.0 Link encap:Ethernet  HWaddr 00:18:51:2d:2c:f8
          inet6 addr: fe80::218:51ff:fe2d:2cf8/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:1186 errors:0 dropped:0 overruns:0 frame:0
          TX packets:8091 errors:0 dropped:106 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:300639 (293.5 KiB)  TX bytes:529016 (516.6 KiB)

[...]

(why the TX dropped on the veth interfaces?)

Any clues as to what is happening here ?

TIA,

Matei