Newbie question. Proxmox host losing connection when starting VM

[r0ck3t42]

Hello,
I am a very green proxmox newbie installing for the first time. System specs below
AMD Ryzen 9 7950X3D
128 GB RAM
ASRock X670E motherboard with 2.5 and 1G Realtek interfaces
Integrated AMD Raphael GPU
GeForce 4070 dGPU

Installation ran great and came up with now problems.
My /etc/network/interfaces

auto lo
iface lo inet loopback

iface enp110s0 inet manual

iface enp113s0 inet manual

auto vmbr0
iface vmbr0 inet static
        address 192.168.1.150/24
        gateway 192.168.1.1
        bridge-ports enp110s0
        bridge-stp off
        bridge-fd 0

iface wlp112s0 inet manual

I first created a linux vm with the following config.

args: -cpu 'host,+kvm_pv_unhalt,+kvm_pv_eoi,hv_vendor_id=NV43FIX,kvm=off'
bios: ovmf
boot: order=ide2;net0
cores: 4
cpu: host,hidden=1,flags=+pcid
cpuunits: 1024
efidisk0: local-lvm:vm-100-disk-0,efitype=4m,pre-enrolled-keys=1,size=4M
hostpci0: 0000:7f:00,pcie=1
ide2: local:iso/garuda-dr460nized-gaming-linux-zen-240428.iso,media=cdrom,size=4794856K
machine: q35
memory: 16384
meta: creation-qemu=8.1.5,ctime=1725579402
name: Von-linux
net0: virtio=BC:24:11:EA:E5:8C,bridge=vmbr0
numa: 0
ostype: l26
scsihw: virtio-scsi-single
smbios1: uuid=d892a29d-7e76-4c14-b512-7a5c24100854
sockets: 1
unused0: local-lvm:vm-100-disk-1
vmgenid: e45fc0d8-e4f8-44ec-90cd-054731286f3e

I plan to pass through the integrated GPU to it but it is not set on there yet.
When I start the VM I get no error messages and I lose connectivity to the Proxmox host. The computer is still running. I have to reboot to get it back.

I thought that it was a network issue.
Next I configured a Windows VM.
I already have a name drive with windows installed so I passed that through and I also passed through the 1G interface and the GeForce 4070.
Here is the config

agent: 1
bios: ovmf
boot: order=scsi2;ide0;ide2
cores: 4
cpu: x86-64-v2-AES
cpuunits: 1024
efidisk0: local-lvm:vm-101-disk-0,efitype=4m,pre-enrolled-keys=1,size=4M
hostpci0: 0000:6e:00,pcie=1
hostpci1: 0000:01:00,pcie=1
ide0: local:iso/virtio-win-0.1.262.iso,media=cdrom,size=708140K
ide2: local:iso/Win11_23H2_EnglishInternational_x64v2.iso,media=cdrom,size=6638374K
machine: pc-q35-8.1
memory: 16384
meta: creation-qemu=8.1.5,ctime=1725584216
name: von-win11vm
numa: 0
ostype: win11
scsi2: /dev/disk/by-id/nvme-Samsung_SSD_990_PRO_2TB_S7KHNU0X628156R,discard=on,size=1953514584K
scsihw: virtio-scsi-single
smbios1: uuid=ccff4fdd-8ce5-41c5-b660-4c1c1402a051
sockets: 1
tpmstate0: local-lvm:vm-101-disk-2,size=4M,version=v2.0
unused0: local-lvm:vm-101-disk-1
vmgenid: 950c2286-ae6f-444c-b194-f0dbf9692132
args: -cpu 'host,+kvm_pv_unhalt,+kvm_pv_eoi,hv_vendor_id=NV43FIX,kvm=off'
cpu: host,hidden=1,flags=+pcid

However, again as soon as I start the VM I lose connectivity to the host machine. With this one it doesn't appear to be a network configuration issue so I'm not sure what I could be doing wrong here.

Any help would be appreciated.

Edit: A little more information. I started the windows vm and then rebooted and looked at the logs and see the following

Sep 06 10:47:54 proxmox pvedaemon[5133]: start VM 101: UPID:proxmox:0000140D:00020F31:66DB323A:qmstart:101:root@pam:
Sep 06 10:47:54 proxmox pvedaemon[1425]: <root@pam> starting task UPID:proxmox:0000140D:00020F31:66DB323A:qmstart:101:root@pam:
Sep 06 10:47:54 proxmox kernel: r8169 0000:6e:00.0 enp110s0: Link is Down
Sep 06 10:47:54 proxmox kernel: vmbr0: port 1(enp110s0) entered disabled state
Sep 06 10:47:54 proxmox kernel: r8169 0000:6e:00.0 enp110s0 (unregistering): left allmulticast mode
Sep 06 10:47:54 proxmox kernel: r8169 0000:6e:00.0 enp110s0 (unregistering): left promiscuous mode
Sep 06 10:47:54 proxmox kernel: vmbr0: port 1(enp110s0) entered disabled state

So my question is why is starting the vm taking down the enp110s0 interface? Especially since it is not tied to vmbr0 at all?

 

[leesteken]

Maybe the network device and the device you are passing through are in the same IOMMU group? https://pve.proxmox.com/wiki/PCI_Passthrough#Verify_IOMMU_isolation

 

[r0ck3t42]

Maybe the network device and the device you are passing through are in the same IOMMU group? https://pve.proxmox.com/wiki/PCI_Passthrough#Verify_IOMMU_isolation

Thanks for the response I'll take a look at the link. However, the NIC is group 18, The NIC for the management port is group 15 and the 4070 is group 12.

 

[r0ck3t42]

I reinstalled Proxmox and I was able to install a linux vm before configuring any passthrough. As soon as I configured the passthrough I get the same issue happening as soon as I start the vm. Here is what I configured for the passthrough. I followed this guide

/etc/default/grub

GRUB_DEFAULT=0
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR=`lsb_release -i -s 2> /dev/null || echo Debian`
GRUB_CMDLINE_LINUX_DEFAULT="quiet iommu=pt"
GRUB_CMDLINE_LINUX=""

/etc/modules

vfio
vfio_iommu_type1
vfio_pci

/etc/modprobe.d/kvm.conf

options kvm ignore_msrs=1 report_ignored_msrs=0

/etc/modprobe.d/vfio.conf

options vfio-pci ids=1002:164e,1002:1640,10de:2786,10de:22bc
softdep radeon pre: vfio-pci
softdep amdgpu pre: vfio-pci
softdep nouveau pre: vfio-pci
softdep nvidia pre: vfio-pci
softdep nvidiafb pre: vfio-pci
softdep nvidia_drm pre: vfio-pci
softdep drm pre: vfio-pci

/etc/modprobe.d/blacklist.conf

blacklist radeon
blacklist amdgpu
blacklist nouveau
blacklist nvidia
blacklist nvidiafb
blacklist nvidia_drm

IOMMU, VFIO and blacklist all seems to working correctly.

 

[r0ck3t42]

not getting a lot of help on this but i wanted to share my own troubleshooting efforts. I removed the iGPU from the vfio, from the backlist and everything. I was able to get the windows vm installed and get passthrough working for the Nic, GeForce 4070, NVME drive, and wireless/bluetooth card.
so the conclusion from that is that there most be a conflict with the iGPU and another controller even though it doesn’t shoe it in the IOMMU groups. Does anyone have an idea of how i would go about troubleshooting that?

 

[leesteken]

so the conclusion from that is that there most be a conflict with the iGPU and another controller even though it doesn’t shoe it in the IOMMU groups. Does anyone have an idea of how i would go about troubleshooting that?

Integrated graphics is always more complicated because of the tight integration with the CPU and main memory. Sometimes it crashed the host, somethimes the (Windows) drivers jsut assume stuff that you cannot provide with a VM. But I don't have any specific ideas about troubleshooting this except searching for other people's work-arounds with the exact same hardware.