Openvswitch Trunk does not work for me

engelant

Renowned Member
Jun 4, 2014
16
1
68
Hi there,

today, or better tonight, i decided to reconfigure my network using openvswitch. For testing i attached my two server interfaces to two switch ports.
The switch in use is a managed switch by TP-Link.

eth0 is connected to a port which runs in access mode
(untagged VLAN ID 10, primary VLAN ID 10)

eth1 is connected to a port which runs in trunk mode
(tagged VLAN IDs 1, 10, 101, primary VLAN ID 10)

with a wiped network configuration i tried the following:

Code:
ovs-vsctl add-br br0
ovs-vsctl add-port br0 eth0 vlan_mode=access tag=10
ovs-vsctl add-port br0 eth1
ovs-vsctl add-port br0 vlan10 tag=10 -- set interface vlan10 type=internal
ifconfig vlan10 10.0.14.17 netmask 255.255.255.0
ifconfig eth0 up
ping 10.0.14.1 #Works
ifconfig eth0 down
ifconfig eth1 up
ping 10.0.14.1 #Does not work

I am using Proxmox VE 3.2 with kernel 2.6.32-29-pve.
Also it might be worth a note, that both my interfaces use the intel e1000 driver.

Thanks in advance
 
Hi,
for my testing i wiped /etc/network/interfaces, leaving only eth0, eth1 and lo. ifconfig -a doesn't show br0. Do you think, this still could be the issue?
 
Hello engelant,

I cannot see any restriction for bridge names - br0 should work. The whole configuration seems to be ok. Strange that "ifconfig -a" does not show br0. Check if ovs service is running by

Code:
/etc/init.d/openvswitch-switch status

the response should look like this

Code:
ovsdb-server is running with pid 1179
ovs-vswitchd is running with pid 1188

check also by

Code:
ovs-vsctl show

whether the configuration is (still) correct.

If so verfiy with tcpdump what happens on the interfaces "vlan10", "br0" and "eth1" - if eth1 sends an ARP request but does not get any answer the problem is outside of the server.

Kind regards

Mr.Holmes
 
Hi again,
so first of all, my comment about ifconfig -a for br0 only appies after boot, as soon as I create the bridge br0 is created as supposed.
Just wanted to make clear, that for testing after right after boot i had no other interfaces but eth0, eth1 and lo.

Nobody could spot my error and help me, because as it turns out, the commands I wrote are just fine. Just fired up an other server with differen NICs and it works.

Key seems to be e1000 and this http://openvswitch.org/pipermail/discuss/2012-September/008122.html .

I suppose this won't get fixed for 2.6.32 . It might work with the newer 3.x kernel.
Update: it works with 3.10 Kernel.
 
Last edited:

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!