Hello.
I had a new NVME drive I wanted to add for a while when I had the time as I knew I would have downtime to adjust newly enumerated devices.
I have two autostarting VMs on this node. Node was on latest pve 7.4, one (and only one) has a pcie gpu passed through. I had read to prepare and knew I had to adjust both pci and network devices. I wanted to do this before any upgrades to pve 8. One step at the time.
I proceeded with the screenshotting and saving config files of the VM, disabled autostart and shutdown the node. Added the nvme and started up.
Then modified the VM settings using the UI to get the new pci device id of the gpu and audio device. It would not start, no matter what combination of All functions, GPU, ROM. Pci-e was enabled as in the past. The other VM without passthroughs started fine.
I had updated /etc/network/interfaces to reflect the new interface identifier and ended with:
Before the interface was enp4s0 .
After many hours I decided to upgrade pve in-place and now the pci devices to passthrough appear slightly differently but I've been able to start it up. But the network interface in the VM won't get an ip address. Diagnosing is hard for me because the passthrough takes over the gpu & keyboard so no real console.
I appreciate this question might need to be asked in the networking section but I ask first as is a result of the upgrade/setup of VM.
What happens now is this:
Normal pve boot:
VM starts and I can interact with the desktop. No network. Interface is up. I can restart it, see its definition. No routes out of it nor into it. My dhcpv4 logs at the firewall don't have an entry. It seems to me it can't get out of the pv linux bridge. I can't interact with proxmox UI or ssh into it (no route to host, clearly not getting out of itself).
Rescue mode pve boot:
I can use the local consoie (as the VM hasn't taken over):
Interface and bridge will be down and after
And the firewall gives the lease, all good.
Here I am stuck. If I then resume booting to normal
then I'm back without network on the VM and out of the pve shell.
What I look for is advice how can I diagnose the problem I have, in the pve shell.
Also, before I re-enabled the autostart, the networking appeared fin in the UI. It reflected the new device, it was getting an IP address where I was reaching it. It has a static lease on the router/firewall.
Thanks!
I had a new NVME drive I wanted to add for a while when I had the time as I knew I would have downtime to adjust newly enumerated devices.
I have two autostarting VMs on this node. Node was on latest pve 7.4, one (and only one) has a pcie gpu passed through. I had read to prepare and knew I had to adjust both pci and network devices. I wanted to do this before any upgrades to pve 8. One step at the time.
I proceeded with the screenshotting and saving config files of the VM, disabled autostart and shutdown the node. Added the nvme and started up.
Then modified the VM settings using the UI to get the new pci device id of the gpu and audio device. It would not start, no matter what combination of All functions, GPU, ROM. Pci-e was enabled as in the past. The other VM without passthroughs started fine.
I had updated /etc/network/interfaces to reflect the new interface identifier and ended with:
auto lo
iface lo inet loopback
auto enp8s0
iface enp8s0 inet manual
auto vmbr0
iface vmbr0 inet static
address 192.168.5.2/24
gateway 192.168.5.1
bridge-ports enp8s0
bridge-stp off
bridge-fd 0Before the interface was enp4s0 .
After many hours I decided to upgrade pve in-place and now the pci devices to passthrough appear slightly differently but I've been able to start it up. But the network interface in the VM won't get an ip address. Diagnosing is hard for me because the passthrough takes over the gpu & keyboard so no real console.
I appreciate this question might need to be asked in the networking section but I ask first as is a result of the upgrade/setup of VM.
What happens now is this:
Normal pve boot:
VM starts and I can interact with the desktop. No network. Interface is up. I can restart it, see its definition. No routes out of it nor into it. My dhcpv4 logs at the firewall don't have an entry. It seems to me it can't get out of the pv linux bridge. I can't interact with proxmox UI or ssh into it (no route to host, clearly not getting out of itself).
Rescue mode pve boot:
I can use the local consoie (as the VM hasn't taken over):
Interface and bridge will be down and after
Code:
#systemctl restart networking
Code:
root@pve:~# ip -br -c link show
lo UNKNOWN 00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
enp8s0 UP f0:2f:74:1a:29:ee <BROADCAST,MULTICAST,UP,LOWER_UP>
vmbr0 UP f0:2f:74:1a:29:ee <BROADCAST,MULTICAST,UP,LOWER_UP>
root@pve:~# ip -br -c addr show
lo UNKNOWN 127.0.0.1/8 ::1/128
enp8s0 UP
vmbr0 UP 192.168.5.2/24 fe80::f22f:74ff:fe1a:29ee/64Here I am stuck. If I then resume booting to normal
Code:
#systemctl defaultWhat I look for is advice how can I diagnose the problem I have, in the pve shell.
Also, before I re-enabled the autostart, the networking appeared fin in the UI. It reflected the new device, it was getting an IP address where I was reaching it. It has a static lease on the router/firewall.
Thanks!
ff video=efifb