Proxmox 8.3.1 randomly freeze when GPU plugged in and cant see NIC in one PCI port

wavesswe

New Member
Dec 12, 2024
8
0
1
Hello!
I have come across a wierd behavior that i really dont know how to move forward.

I swapped case for my proxmox server.

Spec:
MB: x11ssh-f
RAM: 32 GB ECC
CPU: Xeon® Processor E3-1245 v6


I have had a 10GB nic plugged in in the 3rd PCI slot in the other case and it have worked without a problem (2 x SFP+)

When swappig case i swapped that out for an 2 x 10GB RJ45 instead that i have had in another working machine and also adding a ASUS GeForce GTX 670 2GB.

When adding this proxmox just freeze after some fiddeling i reached the GUI once. otherwise it freezes before that.

When unplugging the GPU the freezing stops.

The NIC shows up and are up but it says "No Carrier" I have tried to move it to the otherwise covered by GPU PCIE slot and there it works with a connection. But not in the slot where the old nic worked.

So now I removed the GPU and moved the NIC to slot nr2 and it works. But I really want the GPU to work and the NIC to work in the 3rd slot.


Please help!


System log from when i have been trying.

Please let me know if any more data is helpful.



This is the only thing that is not up to latest avalible status:

libpve-storage-perl (8.3.2)
 

Attachments

Last edited:
Still need help but to give an update on my findings so far.

Regarding the NIC issue:
Dec 13 23:21:59 proxmox kernel: bnx2x: [bnx2x_timer:5810(enp5s0f1)]MFW seems hanged: drv_pulse (0x5) != mcp_pulse (0x7fff)
Dec 13 23:21:59 proxmox kernel: bnx2x: [bnx2x_timer:5810(enp5s0f0)]MFW seems hanged: drv_pulse (0x7) != mcp_pulse (0x7fff)
Dec 13 23:21:59 proxmox kernel: bnx2x: [bnx2x_hw_stats_update:871(enp5s0f0)]NIG timer max (4294967295)

This seams to be the issue trigger. when in slot 2 (enp2s0f1 / enp2s0f0) this dosent show. But i dont know how to proceed. I doubt its a NIC HW issue. Something i can do or check from Bios side?




Regarding the GPU i will dig in to the settings mentioned here:

https://ubuntuforums.org/showthread.php?t=2199399
 
Update:

the troublesome PCI-E slot have the following info" PCH SLOT4 PCI-E 3.0 X4 (IN X8)"

Could that be the cause? or should it still work but slower?