Virtual Machine Vlan tagg problem

[jgarcia]

Hi dear friends,

I am having a problem with tagging vlans on VMs in proxmox-ve-2.6.32: 3.4-150 (running kernel: 2.6.32-37-pve) every time i tried I received the following errors:

Added VLAN with VID == 105 to IF -:bond0:-
Cannot find device "bond0.105"
can't activate interface 'bond0.105'
/var/lib/qemu-server/pve-bridge: could not launch network script
kvm: -netdev type=tap,id=net0,ifname=tap10000i0,script=/var/lib/qemu-server/pve-bridge,downscript=/var/lib/qemu-server/pve-bridgedown: Device 'tap' could not be initialized
TASK ERROR: start failed: command '/usr/bin/kvm -id 10000 -chardev .....

ERROR: trying to add VLAN #105 to IF -:bond0:- error: File exists
can't add vlan tag 105 to interface bond0
/var/lib/qemu-server/pve-bridge: could not launch network script
kvm: -netdev type=tap,id=net0,ifname=tap10000i0,script=/var/lib/qemu-server/pve-bridge,downscript=/var/lib/qemu-server/pve-bridgedown,vhost=on: Device 'tap' could not be initialized
start failed: command '/usr/bin/kvm .....

Please do I need any special package to this version of proxmo, or something ?? Because on previous versions it worked like a charm using drbd and other stuff.

Thanks in advance.

 

[wolfgang]

Hi,
try to use a bridge on top of the bond.

 

[jgarcia]

Hi,

You are referring to this:
auto vmbr0
iface vmbr0 inet static
address x.x.x.x
netmask x.x.x.x
gateway x.x.x.x
bridge_ports bond0
bridge_stp off
bridge_fd 0

This is what I have now. Or it needs to be different??

Thanks.

 

[jgarcia]

From proxmox side, It should be able to create automatically the sub-interface once I put the vlan id in the VM vlan tag field.
Please correct me if I am wrong and they change something in this version, and now it's necessary to manually set this out in /etc/network/interfaces every time we add a new VM belonging to a new VLAN.

 

[jgarcia]

Hi dear friends,

Problem solved.
Because I am using ceph on top of the cluster, and I have a separated ceph network. From the proxmox side I needed to put directly an ip into my eth3 (untagged) and use that for ceph network without bridge/bond for all nodes.

Staying only with a bond and bridge interfaces for the rest of ethx also untagged.

And from the switch (huawei) side needed to configure an adittional portchannel for the ceph network.

Anyway is working now.

Thanks, and sorry for my english.