Proxmox 6.8.12-9-pve kernel has introduced a problem with network connection - enp0s31f6 Intel NIC

L

leandrosgf

New Member
Mar 11, 2025
6
0
1
Sao Paulo / Brazil
Hello All,

Has anybody experienced problems with the enp0s31f6 Intel network card with the new Proxmox VE 6.8.12-9 Kernel?
With this new kernel, the network suddenly stops after some hours. I don't seem to see any relevant information that I can explore.
I don't have the same behavior using the pinned 6.8.12-8-pve Kernel.

Thank you!
 
I don't know if it is relevant, but after the last 3-4 updates my Proxmox 8.3.5 stops responding on the network when the backup starts.
When I pull and put back the ethernet cable, the connection is restored, only the backup fails.
I have here two almost identical systems. One is fully up to date and hangs. The other list a bunch of updates, but is NOT freezing during backup.
One of the things offered is the new Kernel. See screenshot.
 

Attachments

  • 2025-04-01-chrome-004.png
    2025-04-01-chrome-004.png
    85.5 KB · Views: 13
  • Like
Reactions: leandrosgf
SelfMan said:
I don't know if it is relevant, but after the last 3-4 updates my Proxmox 8.3.5 stops responding on the network when the backup starts.
When I pull and put back the ethernet cable, the connection is restored, only the backup fails.
I have here two almost identical systems. One is fully up to date and hangs. The other list a bunch of updates, but is NOT freezing during backup.
One of the things offered is the new Kernel. See screenshot.
Click to expand...
Interesting, I will keep watching it and check If I can see ithis same behaviour.
Looking at other threads it vould be something wrong with E1000 driver and this specific new kernel.

Thanks for your comments!
 
There is a way to revert to the previous kernel version, but I am no linux pro. I need to check it out a bit more.
 
SelfMan said:
There is a way to revert to the previous kernel version, but I am no linux pro. I need to check it out a bit more.
Click to expand...
Me neither, but you can easily do it using this way:
1. First, list your available kernels with the command # proxmox-boot-tool kernel list
2. Once you have a list of your available kernels, choose one and copy the name of this kernel ( in my case, I used the "6.8.12-8-pve" )
3. Finally, pin your kernel for the next boot using this command:# proxmox-boot-tool kernel pin 6.8.12-8-pve --next-boot

You can unpin the kernel and go back to the latest one anytime using the command: # proxmox-boot-tool kernel unpin

Check out: How to pin and unpin a specific kernel
 
Hi and yeah, I had phind.com those steps already and had them ready to use in my notes. Thanks anyway.
The conclusion is that it worked. I am on the same "old" 6.8.12-8-pve kernel and the backup went on as usual. No hiccups, no network problems.
Thanks for pointing me to the kernel.
 
Same behaviour here. Network card is up but no connectivity (Intel I217-LM controller). I must disconnect and reconnect the network cable and then network access is reestablished.
Issues begin after last kernel update.
I've just reverted kernel to 6.8.12-8-pve , thanks @leandrosgf , will see if issues are gone.
 
Just FYI, I have two network adapters on the board 1+2.5Gb
Intel Corporation Device 0dc8 (rev 11) I219V - 1Gb (connected)
Intel Corporation Ethernet Controller I226-V (rev 04) - 2.5Gb disconnected/not used)
on a ASRock Z790M-ITX WIFI motherboard.
The second backup went just fine today, so I can confirm that the Kernel downgrade helped.
 
Last edited:
  • Like
Reactions: francesc_dj
I'm having the same issue with an I217-LM (rev 04). Downgrading the kernel has fixed it for me so far.
 
Last edited:
Hi,
I'm experiencing the same issues with the latest kernel and the I217-LM (rev 04).
There's a similar problem here that has been resolved LINK

Code: 
ethtool  -K enp0s31f6 gso off gro off tso off tx off rx off rxvlan off txvlan off sg off

Could this possibly be a solution here as well?
 
FYI - The PBS, which is running in my PVE as a VM, crashed during backup of the VM. Ping was ok, SSH allowed login, but that was it. No shell and no web admin. I had to forcefully shutdown the VM.
I did not downgrade the kernel in this VM yet and I am testing the reconfiguration of the network adapter as proposed by Hegse.
 
Last edited:
I have the same issue. Just installed Proxmox to one of my servers a few days ago. Before there was Windows Server running on the same hardware.
I have two Intel NICs: I210 for MGMT traffic and I219-LM for VM data traffic with VLAN tagging.
When I do an backup or move a lot of data between the VMs and an external destination, the VMs loosing connection until I shut/ no shut the port on the switch where the nic is connected.

I downgraded to kernel 6.8.12-8-pve but with no success, now changed the NIC setting with ethtool like mentioned by Hegse and could do the backup without problems.
Tonight I have some sync jobs running, let's see if the connection is broken again tomorrow or not. But for the moment it looks like that helped.

But I just bought a pcie card with dual 2,5G lan and Realtek chipset for this server. I have another Proxmox host with a quad 2,5G realtek pcie card and never had problems with it.
 
I've ran a backup using the 6.8.12-9-pve kernel with the settings recommended by Hegse and it ran without a freeze. i have to give it a few more days.
 
Silas95 said:
I have the same issue. Just installed Proxmox to one of my servers a few days ago. Before there was Windows Server running on the same hardware.
I have two Intel NICs: I210 for MGMT traffic and I219-LM for VM data traffic with VLAN tagging.
When I do an backup or move a lot of data between the VMs and an external destination, the VMs loosing connection until I shut/ no shut the port on the switch where the nic is connected.

I downgraded to kernel 6.8.12-8-pve but with no success, now changed the NIC setting with ethtool like mentioned by Hegse and could do the backup without problems.
Tonight I have some sync jobs running, let's see if the connection is broken again tomorrow or not. But for the moment it looks like that helped.

But I just bought a pcie card with dual 2,5G lan and Realtek chipset for this server. I have another Proxmox host with a quad 2,5G realtek pcie card and never had problems with it.
Click to expand...
Just FYI, the switch for the kernel downgrade is only temporary for the next boot. To make it permanent you have to use:
Bash: 
proxmox-boot-tool kernel pin 6.8.12-8-pve
 
SelfMan said:
Just FYI, the switch for the kernel downgrade is only temporary for the next boot.
Click to expand...
I know, I already made it permanent.
But I reverted back to 6.8.12-9-pve now cause I had a container running with docker compose and that broke with the old kernel version.
I keep investigating if the nic error still occurs.
 
Last edited:
You must log in or register to reply here.
Top Bottom