Intel 82574L Causing Network Issues?

[Lunar]

I've been testing Proxmox for about a week and I'm having a consistent issue with my Intel 82574L NIC. I've searched around and it seems other users have been experiencing the same issues. Basically my node will function without Proxmox just fine but once I get Proxmox up and running it will run for a bit then the entire server will become unreachable. I'm using a Hetzner dedicated server if that helps. I have 2 interfaces, eth0 and vmbr0 which is linked to eth0. I've been trying to troubleshoot it and I can't figure out how to solve the problem. Any help would be greatly appreciated.

 

[Lunar]

Willing to pay at this point for a fix

 

[tom]

If you have network issues in a datacenter, you should ask the datacenter for help.

Hetzner is not a Proxmox Hosting Partner but they run a lot of Proxmox VE Servers.

 

[Lunar]

If you have network issues in a datacenter, you should ask the datacenter for help.

Hetzner is not a Proxmox Hosting Partner but they run a lot of Proxmox VE Servers.

It doesn't seem to be a network issue at the datacenter. Debian 8 install works fine, it only starts having problems when I install Proxmox and create vmbr0.

 

[Lunar]

I also had them test all the hardware on the server. Nothing came up.

 

[t.lamprecht]

I also had them test all the hardware on the server. Nothing came up.

If you install it and create vmbr0 then you mean that you instal Proxmox VE on top of (plain) Debian? If so please post your network (after setting up Proxmox VE) config so we see if there is something wrong.

 

[Lunar]

If you install it and create vmbr0 then you mean that you instal Proxmox VE on top of (plain) Debian? If so please post your network (after setting up Proxmox VE) config so we see if there is something wrong.

Yea I installed it on top of Debian 8.

auto lo
iface lo inet loopback

iface lo inet6 loopback

auto eth0
iface eth0 inet static
    address  (ip was here)
    netmask  255.255.255.224
    gateway  (ip was here)
    up route add -net (ip was here) netmask 255.255.255.224 gw (ip was here) eth0
# default route to access subnet

iface eth0 inet6 static
    address  (ip was here)
    netmask  64
    gateway  fe80::1

auto vmbr0
iface vmbr0 inet static
    address  (ip was here)
    netmask  255.255.255.255
    bridge_ports eth0
    bridge_stp off
    bridge_fd 0

iface vmbr0 inet6 static
    address  (ip was here)
    netmask  64

 

[Lunar]

I noticed something. The server is up as long as I'm connected to it. Once I close all my connections to it like SSH and the Proxmox web interface it goes down.

 

[t.lamprecht]

I assume you followed:
https://wiki.hetzner.de/index.php/Proxmox_VE/en#Network_configuration
It seems you have a mixed Bridged/Routed configuration, could you mark the IP place holder with letters so we can differentiate where same IPs are used?

Could it be that you have another program managing the network still installed from the Debian installation? E.g. NetworkManager, that would explain such behavior.. If so remove them.

I'd generally recommend using the Proxmox VE ISO provided from us and else using a minimal Debian installation as a start, to avoid pulling dependencies which may give you problems.

 

[mir]

Could it be that you have another program managing the network still installed from the Debian installation? E.g. NetworkManager, that would explain such behavior.. If so remove them.

If OP have installed a graphical desktop from Debian on the server he will automatically get networkmanager or wicd as part of the installation.

 

[Lunar]

I have changed the network config but still having the same issue. I also want to mention the system is encrypted with LUKS and unlocked with dropbear.

auto lo
iface lo inet loopback

iface lo inet6 loopback

iface eth0 inet manual

auto vmbr0
iface vmbr0 inet static
    address  5.x.xxx.44
    netmask  255.255.255.224
    gateway  5.x.xxx.33
    bridge_ports eth0
    bridge_stp off
    bridge_fd 0

iface vmbr0 inet6 static
    address  2a01:xxx:xxx:xxxx::2
    netmask  64
    gateway  fe80::1

 

[Lunar]

I assume you followed:
https://wiki.hetzner.de/index.php/Proxmox_VE/en#Network_configuration
It seems you have a mixed Bridged/Routed configuration, could you mark the IP place holder with letters so we can differentiate where same IPs are used?

Could it be that you have another program managing the network still installed from the Debian installation? E.g. NetworkManager, that would explain such behavior.. If so remove them.

I'd generally recommend using the Proxmox VE ISO provided from us and else using a minimal Debian installation as a start, to avoid pulling dependencies which may give you problems.

I'm using a bridged config and it works fine without LUKS. When I encrypt the system and use dropbear to unlock the disk everything works fine. But once I install Proxmox and configure the network it is stable until I close my ssh session. So basically everything goes fine until I close my ssh client. Then the entire system is unreachable. I have tried using dropbear to unlock, and then openssh on the main system. And just dropbear as well. It doesn't seem to make a difference. This is why I'm wondering if it could be Proxmox related, because that's what seems to be causing the issues.

I don't believe it's anything that was installed with Debian as I installed the bare minimum. And I'd like to use the official ISO but it doesn't support LUKS..

 

[Lunar]

I've discovered why the connection is dropping. Not sure how to fix it though...

Example here when I look at ARP

Address                  HWtype  HWaddress           Flags Mask            Iface
5.x.xxx.33               ether   0c:96:12:f5:f0:a7   C                     eth0
Entries: 1    Skipped: 0    Found: 1
.....
Address                  HWtype  HWaddress           Flags Mask            Iface
5.x.xxx.33                       (incomplete)                              eth0
Entries: 1    Skipped: 0    Found: 1

Any ideas why it would be doing this?