Problem with bridging and freezing on network.

[derek]

Good morning experts,

I have a new install of Proxmox 3.3 on a Supermicro based server.

Every few minutes, for a few seconds my network connection freezes to my VMs. This appears to be a fairly common problem, however, I haven't found a solution in the forums.

Networking is configured as:

• eth0 - Connected to my private network (e1000e driver) - MAC 00:25:90:57:7d:8f - This is connected to a Trendnet TEG-240WS switch. BPDU passthrough enabled and I do not have STP configured on the switch.
• eth1 - Connected to my public network (e100e driver) - MAC 00:25:90:57:7d:8e
• eth2 and eth3 (currently disabled) are used a private 10gbe network (ixgbe driver)

# network interface settings
auto lo
iface lo inet loopback


iface eth0 inet manual


iface eth1 inet manual


iface eth2 inet manual


iface eth3 inet manual


auto vmbr0
iface vmbr0 inet static
    address  192.168.1.253
    netmask  255.255.255.0
    gateway  192.168.1.1
    bridge_ports eth0
    bridge_stp off
        bridge_fd 0


#bridge_stp on
#bridge_maxwait 0
#bridge_maxage 0
#bridge_fd 0
#bridge_ageing 0


auto vmbr1
iface vmbr1 inet manual
    bridge_ports eth1
    bridge_stp off
    bridge_fd 0


auto vmbr2
iface vmbr2 inet static
    address  10.0.0.1
    netmask  255.255.255.0
    bridge_ports eth2
    bridge_stp off
    bridge_fd 0


pre-up ifconfig eth2 mtu 9000
post-up ifconfig vmbr2 mtu 9000

I am receiving the common error:

Jan  3 11:04:23 pm01 kernel: eth0: received packet with own address as source address

Turning on STP options above, remove this error from the logs, however, the freezing is still present.
I have tried removing the 192 address from vmbr0 and the problem is still present. The VMs do not have duplicate MAC addresses.

Running a tcpdump on my network reveals a continual flooding of packets every two seconds.

root@pm01:~# tcpdump -i vmbr0 | grep STP
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on vmbr0, link-type EN10MB (Ethernet), capture size 65535 bytes
11:18:09.682608 STP 802.1d, Config, Flags [none], bridge-id 8000.20:e5:2a:0e:5c:7d.8001, length 43
11:18:11.682439 STP 802.1d, Config, Flags [none], bridge-id 8000.20:e5:2a:0e:5c:7d.8001, length 43
11:18:13.682431 STP 802.1d, Config, Flags [none], bridge-id 8000.20:e5:2a:0e:5c:7d.8001, length 43
11:18:15.682384 STP 802.1d, Config, Flags [none], bridge-id 8000.20:e5:2a:0e:5c:7d.8001, length 43

This problem occurs whether or not the VMs are on.
XEN Server and VMware didn't have any issues with the switch configuration, however, I understand their networking stacks are significantly different.

Does anyone have any insight into how to fix this? I would like to use Proxmox as my ESXi replacement, however, this is making my network completely unusable at the moment.

Thanks!

 

[derek]

Interesting, the bridge information was being propagated by my Netgear R6300 WAP. I didn't find any options in config, but after flashing that with DD-WRT, the STP option was turned off. The performance is back to normal (for now).

 

[sgargel]

I've the same (or similar) problem. I've experienced it after upgrade my proxmox nodes.

I've upgraded 3 nodes in cluster to "pve-manager/3.4-11/6502936f (running kernel: 3.10.0-14-pve)"
and one stand-alone to "pve-manager/4.1-1/2f9650d4 (running kernel: 4.2.6-1-pve)"

<code>
[56601.876509] vmbr0: received packet on eth0 with own address as source address
[56666.988265] vmbr0: received packet on eth0 with own address as source address
[56791.672522] vmbr0: received packet on eth0 with own address as source address
[56992.864270] vmbr0: received packet on eth0 with own address as source address
[57066.272488] vmbr0: received packet on eth0 with own address as source address
</code>

Did someone found any solution?

 

[sgargel]

Ok... i've isolated the stand-alone server and no errors are displayed. So.. maybe that upgrading proxmox only activate the logging of the issue.

But the issue resides very probably on the network side.

Edit:
seems the same as https://forum.proxmox.com/threads/new-install-eth0-received-packet-with-own-address-as-source.9183/

Edit2: Identified the problem by founding many CRC errors on a switch port. The problem was on a cheap AccessPoint "burned"