My bridge does not work

L

langloispy

Guest
Hi everyone,

I have create a bridge over a bond on 2 nic (which are connected on two different switches). When I create a VM (openvz), the VM is able to ping the host but cannot ping anyone else outside the host.

Here is a sample of my interfaces files: http://pastebin.com/YdEaxTrf

How can I debug this problem or do you have some hints for me?

Thanks!

pveversion -v
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LANG = "fr_CA.utf8"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
pve-manager: 1.6-2 (pve-manager/1.6/5087)
running kernel: 2.6.32-4-pve
proxmox-ve-2.6.32: 1.6-19
pve-kernel-2.6.32-3-pve: 2.6.32-14
pve-kernel-2.6.32-4-pve: 2.6.32-19
qemu-server: 1.1-18
pve-firmware: 1.0-8
libpve-storage-perl: 1.0-14
vncterm: 0.9-2
vzctl: 3.0.24-1pve4
vzdump: 1.2-7
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.12.5-1
ksm-control-daemon: 1.0-4
 
Last edited by a moderator:
Ok... I got it to work but I still don't understand why it did not work previously:

I have reconfigure my nic bonding to use the active-backup mode instead of the balance-rr mode. Why the balance-rr does not work?
 
Ho no. It fails again. Using the active-backup mode, I was able to ping the VM and the host. I then tried to update my VM (apt-get update) and now both host and vm can't be pinged... Help!
 
New development. I have undo the bond config. Now I got a simple bridge vmbr0 over my eth5 (which is connected to the net). When I create a VM (openvz) and tries to connect to the net from the VM (apt-get update), it crashes the entire server (vm + host) ... How can I debug this problem? Here is what I see on the raritan console: http://ow.ly/i/44r3

It is really a openvz related problem... I tried with a debian 6 and ubuntu 8.04 in an openvz container and both crashes. When I tried with a KVM ubuntu 10.04 with virtio drivers, everything works fine.

Thanks
 
Last edited by a moderator:
I tried with the kernel 2.6.24 but this kernel does not recognize my network cards.

Does this mean that I'm very lucky that the version 1.6 (with 2.6.32 support) was released at the begin of the month, otherwise I would not be able to run PVE on that hardware?

NIC:
01:00.0 Ethernet controller: Broadcom Corporation NetXtreme II BCM5709 Gigabit Ethernet (rev 20)
07:00.0 Ethernet controller: Intel Corporation 82576 Gigabit Network Connection (rev 01)

Here is the content of the last entries from /var/log/messages before a crash:
Sep 23 16:33:13 pve02-fl kernel: CT: 101: started
Sep 23 16:33:14 pve02-fl kernel: device veth101.0 entered promiscuous mode
Sep 23 16:33:14 pve02-fl kernel: vmbr0: port 2(veth101.0) entering forwarding state

Is there some other thing I could try to understand how openvz crashes the host?
 
dietmar said:
That config looks quite strange - why do you assign an ip to eth0 (instead of bond0)?
Click to expand...

Sorry, this pastebin contains two files... the first partof the paste is the /etc/network/interface from the HOST, the second part is from the VM...
 
No.. At first I suspected the bonding to be the cause but it was not it. (See post #4). In fact that thread should be renamed: openvz + network = crashes host. Could it be a problem with the network card driver? That server will soon be transferred on a 10Gb switch so I will also change my NIC accordingly (OCE10102-NX emulex 10 gig)... I will then be able to validate that hypothesis. But if you want, I can help you to debug with that particular configuration until I got my new cards.

I also have a question about the emulex card... I don't know if it is supported natively by the kernel but I know that the manufacturer provides linux drivers that needs to be compiled. If it is not supported, will I need to compile those drivers in the openvz to update its kernel or do I only need to compile them on the host?

Thanks!
 
Last edited by a moderator:
The most stable OpenVZ kernel is 2.6.18 - please can you try that to? To debug the problem, you need to find a way to reliable reproduce the bug - so how can I trigger the bug (exactly)?
 
Hi Dietmar,

Here is what I did in detail. I hope this is enough information for you to reproduce the bug.

Hardware:
DELL R710, 32Gb RAM
NIC: 07:00.0 Ethernet controller: Intel Corporation 82576 Gigabit Network Connection (rev 01)

Using the PVE 1.6 ISO released on the Sep. 02, 2010 I have installed PVE on that hardware. The installation work like a charm, the NIC is automatically configured. I have succesfully updated the host with apt.

Then I have created a OpenVZ from the Debian 6.0 appliance. The NIC is bridged.
I can connect to it without any problem from the VNC console.
I have edited the /etc/network/interface to add eth0 as DHCP.
I issued a "/etc/init.d/networking restart" to obtains an IP.
I then received a valid the IP.
I ping the host from the openVZ. Everything works fine.
I tried to update the repo of the OpenVZ with the command "apt-get update". This command crashes the host.

pve-manager: 1.6-2 (pve-manager/1.6/5087)
running kernel: 2.6.32-4-pve
proxmox-ve-2.6.32: 1.6-19
pve-kernel-2.6.32-3-pve: 2.6.32-14
pve-kernel-2.6.32-4-pve: 2.6.32-19
qemu-server: 1.1-18
pve-firmware: 1.0-8
libpve-storage-perl: 1.0-14
vncterm: 0.9-2
vzctl: 3.0.24-1pve4
vzdump: 1.2-7
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.12.5-1
ksm-control-daemon: 1.0-4

When I try to use a kernel older than 2.6.32-4-pve, my network card was not recognized. I wanted to try to recompile the Intel drivers for 2.6.18 kernel but meanwhile, I have received my 10Gb NIC. The tech guy is now removing my server from the rack to install my new NIC so I can't try to recompile the driver for now. I hope to do it as soon as possible.

This is where I am.
About drivers compilation. Since the OpenVZ shares the host's kernel, I suppose that I only need to compile it on the host to render the OpenVZ functionnal. Am I right?
 
Thanks Tom, I will test it as soon as my server will be plugged.
 
hello,

i have same problem Intel® 82576 controller,

i try test debit on NOD:

:~# wget http://test-debit.free.fr/image.iso
--2010-10-08 21:30:57--
http://test-debit.free.fr/image.iso
Resolving test-debit.free.fr... 212.27.60.49
Connecting to test-debit.free.fr|212.27.60.49|:80...
connected.
HTTP request sent, awaiting response... 200 OK
Length: 678526976 (647M) [application/x-iso9660-image]
Saving to: `image.iso'

100%[================================================== ====
=============================>] 678,526,976 56.0M/s in
13s
Click to expand...
i try test debit on VM:

~# wget http://test-debit.free.fr/image.iso
--2010-10-08 19:49:03--
http://test-debit.free.fr/image.iso
Resolving test-debit.free.fr... 212.27.60.49
Connecting to test-debit.free.fr|212.27.60.49|:80...
connected.
HTTP request sent, awaiting response... 200 OK
Length: 678526976 (647M) [application/x-iso9660-image]
Saving to: `image.iso'

0%[ ] 344,347 17.5K/s eta 8h 20m
Click to expand...
Code: 
dmidecode 2.9
SMBIOS 2.6 present.

Handle 0x0002, DMI type 2, 15 bytes
Base Board Information
        Manufacturer: Supermicro
        Product Name: H8DGU
        Version: 1234567890
        Serial Number: 1234567890
        Asset Tag: To Be Filled By O.E.M.
        Features:
                Board is a hosting board
                Board is replaceable
        Location In Chassis: To Be Filled By O.E.M.
        Chassis Handle: 0x0003
        Type: Motherboard
        Contained Object Handles: 0
My serveur http://www.supermicro.com/Aplus/motherboard/Opteron6100/SR56x0/H8DGU.cfm

I try kernel v3 v4 and last test pve test, no result :/

Linux ns213267.ovh.net 2.6.32-4-pve #1 SMP Fri Oct 8 12:50:10 CEST 2010 x86_64 GNU/Linux


firmware ? driver controler ?

have you solution ? tks


Edit: :~# uname -a
Linux ns213267.ovh.net 2.6.18-3-pve #1 SMP Mon Sep 20 14:57:48 CEST 2010 x86_64 GNU/Linux

wget http://test-debit.free.fr/image.iso
--2010-10-10 22:54:21-- http://test-debit.free.fr/image.iso
Resolving test-debit.free.fr... 212.27.60.49
Connecting to test-debit.free.fr|212.27.60.49|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 678526976 (647M) [application/x-iso9660-image]
Saving to: `image.iso.1'

100%[======================================>] 678,526,976 45.4M/s in 16s

2010-10-10 22:54:37 (41.1 MB/s) - `image.iso.1' saved [678526976/678526976]
Click to expand...


this kernel is ready for my controler but this kernel have you bug hack "kernel 64bit" root accès ? ?
 
Last edited:
dietmar said:
Please can you test with the latest kernels from the pvetest repository (uploaded today)?
Click to expand...

Hi, my server still down.. I hope to test your kernel next week. How can I setup the pvetest repo? In the source.list, Do I replace the line

deb http://download.proxmox.com/debian lenny pve

for

deb http://download.proxmox.com/debian lenny pvetest

Or do I need to add two lines:

deb http://download.proxmox.com/debian lenny pve
deb http://download.proxmox.com/debian lenny pvetest

thanks
 
You must log in or register to reply here.

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!

Quick Navigation

Home Get Subscription Wiki Downloads Proxmox Customer Portal About

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!
Community platform by XenForo® © 2010-2024 XenForo Ltd.
Top Bottom
Back