Proxmox Vlan issues after GPU removal

T

Thundercat

New Member
Jan 31, 2023
13
0
1
Hello everyone!

Im aware of the bug where removing a PCIE device your onboard NIC gets renamed. This has really messed with my setup and its taken me for an unwanted loop..

Ive "fixed" the issue using nano and renamed the last 2 interfaces in the networking document to the current name of my only nic. Now...

The machine gets access to the internet and can ping devices outside of it no issues, however when i reboot my machine it says i can access it by going to 10.10.3.70. The Vlan it should be on is 10.10.70.x (in this case 7, the other vms run on 10.10.70.11, 12, 13 etc).

I cant ping it from my management vlan and the boot up screen tells me its on the previously mentioned 10.10.3.70.

It isnt helped that im running this under a really annoying Ubiquiti Edgerouter 10x that wont show active ip-adresses under each vlan either. I have however segregated the vlans so that they shouldnt be able to access eachother.


Now, do anyone know if the 10.10.3.70 adress is correct or a static item shown on startup? Anyone otherwise have a good idea of how to fix this issue?
 
VictorSTS said:
The IP shown in the console when you boot Proxmox is the one set in /etc/hosts for your hostname. If you modified the IPs of the host, remember to change the IP there too.
Click to expand...
Thanks m8! I assumed it updated as it should have. After fixing the interface name i got it to work. Sortof.

I now have issues where the Proxmox box can be reached and it can reach out both locally (vlan), east/west but also outside of the LAN, which is good. Whats not good is that the VM's cant be reached and they cant reach out.

Ive set the new vmbr0 as Vlan aware and ive specified the vlan the box is on (i assume thats used to wrap the packets according to normal Vlan behaviour) but it still doesnt work. Ive rebooted the vm's multiple times without much success.
Ive even disabled the firewall rule for the vmbr0 on one of the machines but to no avail.

I cant ping the machines from the admin vlan outside of the proxmox box but i can ping the box itself.
 
VictorSTS said:
I need the exact network configuration of both the host /etc/network/interfaces and a VM, both qm config VMID and the configuration of the network in the guest OS.
Click to expand...
How do i get the "qm config VMID" information?
/etc/network/interfaces

auto lo
iface lo inet loopback

auto enp3s0
iface enp3s0 inet static
address 10.10.70.7/24
gateway 10.10.70.1
bridge_ports enp3s0
bridge_stp off
bridge_vids 2-4094
bridge_vlan_aware 1

iface enp3s0 inet static
address 10.10.70.7/24
gateway 10.10.70.1

auto vmbr0
iface vmbr0 inet static
address 10.10.70.10/24
bridge-ports none
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4094


Guest os (VM):
IP: 10.10.70.16
Default route: 10.10.70.1 (Router)
DNS: 1.1.1.1

Ls on /etc/network/ doesnt show the interfaces file.
 
VictorSTS said:
run qm config with the number (ID) of the VM.

That whole text that you posted before is your /etc/network/interfaces of the Proxmox host?
Click to expand...
VM:
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 161.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 196.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 196.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 222.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 222.
boot: order=scsi0;ide2
cores: 6
cpu: host,flags=+pcid;+aes
ide2: none,media=cdrom
memory: 14000
name: MServer
net0: virtio=86:29:23:52:CD:DB,bridge=vmbr0,tag=70
numa: 0
onboot: 1
ostype: l26
parent: A2023_03_11
protection: 1
scsi0: SSD1:vm-103-disk-0,discard=on,size=200G,ssd=1
scsihw: virtio-scsi-single
smbios1: uuid=15189f76-a6c1-4b8a-8ed6-96243de3a849
sockets: 1
startup: order=1,up=20
unused0: local-lvm:vm-103-disk-0
vga: memory=512
vmgenid: c311f051-d3df-4fad-b29e-8d9df5ac8f6e
root@host1:~#


"That whole text that you posted before is your /etc/network/interfaces of the Proxmox host?"
Yes, that was the interfaces file of the host, that is correct.
 
Thundercat said:
"That whole text that you posted before is your /etc/network/interfaces of the Proxmox host?"
Yes, that was the interfaces file of the host, that is correct.
Click to expand...
That file is a mess. Dunno how was it generated, but pretty sure it wasn't using the web UI:

Thundercat said:
auto lo
iface lo inet loopback

auto enp3s0
iface enp3s0 inet static
address 10.10.70.7/24
gateway 10.10.70.1
bridge_ports enp3s0
bridge_stp off
bridge_vids 2-4094
bridge_vlan_aware 1
Click to expand...
You can't have bridge configuration in a non bridge interface (must start with vmbr) [1]

Thundercat said:
iface enp3s0 inet static
address 10.10.70.7/24
gateway 10.10.70.1
Click to expand...
Iface enp3s0 is present twice in the network's file.

Thundercat said:
auto vmbr0
iface vmbr0 inet static
address 10.10.70.10/24
bridge-ports none
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4094
Click to expand...
This vmbr0 is missing bridge-ports, so it is not connected to any physical interface of the host, hence your VM's can't reach the world outside the bridge itself.

I suppose that you need something like:

Code: 
auto lo
iface lo inet loopback

iface enp3s0 inet static
address 10.10.70.7/24
gateway 10.10.70.1

auto vmbr0
iface vmbr0 inet static
address 10.10.70.10/24
bridge-ports enp3s0
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4094
 -> Check a few posts below, sorry.

Then use ifreload -a or reboot the host.

Thundercat said:
net0: virtio=86:29:23:52:CD:DB,bridge=vmbr0,tag=70
Click to expand...

Remove the vlan on the VM configuration: it is in the same network as the host but the host is not using a vlan, so I suppose that your switch isn't tagging packets with vlan 70.


[1] https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#_naming_conventions
 
Last edited:
  • Like
Reactions: Thundercat
VictorSTS said:
That file is a mess. Dunno how was it generated, but pretty sure it wasn't using the web UI:


You can't have bridge configuration in a non bridge interface (must start with vmbr) [1]


Iface enp3s0 is present twice in the network's file.


This vmbr0 is missing bridge-ports, so it is not connected to any physical interface of the host, hence your VM's can't reach the world outside the bridge itself.

I suppose that you need something like:

Code: 
auto lo
iface lo inet loopback

iface enp3s0 inet static
address 10.10.70.7/24
gateway 10.10.70.1

auto vmbr0
iface vmbr0 inet static
address 10.10.70.10/24
bridge-ports enp3s0
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4094

Then use ifreload -a or reboot the host.



Remove the vlan on the VM configuration: it is in the same network as the host but the host is not using a vlan, so I suppose that your switch isn't tagging packets with vlan 70.


[1] https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#_naming_conventions
Click to expand...
Il absolutely try that! Thanks a ton m8!
I havent gotten too deep when it comes to the networking side of the proxmox box, as you can clearly see. Its all a mess to me atm.

The network im running is a ubiquiti Edgerouter10x with vlans on every active port (switch in the same package) and firewall rules to deal with network segmentation/isolation.

The proxmox box is connected to the routers tacked on switch and is connected via vlan 70. I guess the vm on the proxmox box doesnt need to be told to run on that vlan if there is only one vlan to send the data through? I assume the hypervisor takes care of the trafic automagically.

Il try that setup and see if it works.
Thanks again! <3
 
VictorSTS said:
There's no automagic anywhere. Recheck vlan theory, you are probably using untagged vlan70 on the switch port of the Proxmox host.
Click to expand...
Cant get it to work. The only config i can get to work (Including restarts of the host and network service) is what i posted. I managed to get a bridge going for a short while but it still didnt send any data over to the VM.

What would a bridge look like for you? And what version of proxmox are you running?
 
Error code:

-- The unit networking.service has entered the 'failed' state with result 'exit-code'.
Nov 16 21:25:09 host1 systemd[1]: Failed to start Raise network interfaces.
-- Subject: A start job for unit networking.service has failed
-- Defined-By: systemd
-- Support: https://www.debian.org/support
--
-- A start job for unit networking.service has finished with a failure.
--
-- The job identifier is 820 and the job result is failed.
 
Your Proxmox install isn't ok. Just noticed these messages when you used qm config (sorry, dunno how did I overlooked that!):
Thundercat said:
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 161.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 196.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 196.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 222.
Use of uninitialized value $PVE::JSONSchema::pVE_TAG_RE in concatenation (.) or string at /usr/share/perl5/PVE/DataCenterConfig.pm line 222.
Click to expand...

You probably used apt update + apt upgrade, but Proxmox requires [1] apt dist-upgrade to install new packates and dependencies. Try running apt dist-upgrade to see if new packages get installed, reboot the host and then run qm config VMID again to see if those error messages are gone.

I'm posting a full configuration that I think you can use, as I made a typo in my example before and included an IP in the physical interface (check if the IP below is the one in /etc/hosts of your Proxmox server:

Code: 
auto lo
iface lo inet loopback

auto enp3s0
iface enp3s0 inet manual

auto vmbr0
iface vmbr0 inet static
address 10.10.70.10/24
bridge-ports enp3s0
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4094

If this still doen't work, recreate the configuration from the webUI from scratch.

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_system_software_updates
 
  • Like
Reactions: Thundercat
VictorSTS said:
Your Proxmox install isn't ok. Just noticed these messages when you used qm config (sorry, dunno how did I overlooked that!):


You probably used apt update + apt upgrade, but Proxmox requires [1] apt dist-upgrade to install new packates and dependencies. Try running apt dist-upgrade to see if new packages get installed, reboot the host and then run qm config VMID again to see if those error messages are gone.

I'm posting a full configuration that I think you can use, as I made a typo in my example before and included an IP in the physical interface (check if the IP below is the one in /etc/hosts of your Proxmox server:

Code: 
auto lo
iface lo inet loopback

auto enp3s0
iface enp3s0 inet manual

auto vmbr0
iface vmbr0 inet static
address 10.10.70.10/24
bridge-ports enp3s0
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4094

If this still doen't work, recreate the configuration from the webUI from scratch.

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_system_software_updates
Click to expand...
Ive tried this about 1 bilion times now and it just doesnt want to work. I can ping the device but i cant access the GUI. I can ping the VM's ip but it cant reach anything outside of its own little vm sphere.

Im honestly fed up with this, whats the best way of just cleaning this machine of the current install and reinstall proxmox on this node? Just a normal purge and install or?

And do you know if the GPU breaking everything bug is present on the current Proxmox build?
Really dont want to have to spend another couple of weeks trying to wrestle with a non functioning system the next time i try to change what gpu i have connected.
 
Im going to loose about 4 months of work due to this bug but i cant justify messing around with it any longer.
 
What's the current output of qm config VMID? As I said, the errors that the command gave you are related to a apt upgrade instead of apt dist-upgrade.

You can backup all your VMs to another disk, reinstall the host and restore the backups.
 
  • Like
Reactions: Thundercat
You must log in or register to reply here.

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Quick Navigation

Home Get Subscription Wiki Downloads Proxmox Customer Portal About

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!
Community platform by XenForo® © 2010-2024 XenForo Ltd.
Top Bottom
Back