I was messing with one of the nodes on my proxmox cluster and experimenting with putting the management interface/GUI on it's own VLAN, and in the process managed to completely kill the network connectivity on the node even after undoing all my changes. I'll try to explain in a series of steps.
Originally, my /etc/network/interfaces file looked something like this:
I have an existing VLAN 10 on the switch this node is connected to, so I update the file to read like this:
This works! I connect the switch to a port that accepts tagged traffic for VLAN 10 and can access it from other ports that belong to the same VLAN. So far, so good.
Other devices use VLAN 10 however, and I want my management interface to be on it's own dedicated VLAN. I set up VLAN 100 on two ports for the switch (one connected to the node, the other to my main router where my PC is connected). Then I add VLAN 100 as an interface in the proxmox GUI. I haven't re-assigned the IP to it yet and while the VLAN exists in the switch no traffic should be getting sent on it yet as the PVIDs are all still set for VLAN 10.
Regardless, as soon as I hit the apply configuration on button in the Proxmox GUI the node's network connectivity goes down and stays down. I have to connect a monitor to it so I can access the node directly and undo all my changes to /etc/network/interfaces until it looks like it did originally. Even so, I can't connect to or ping the node regardless of if it's connected to the switch, my main router (which has no VLANs at all), or with a direct connection between the node and an end terminal like a laptop.
Here's a snapshot of the output from running ip addr on the node: https://imgur.com/a/aa2iXlg
Two things I notice are that the NIC and vmbr0 both show a perpetual state of DOWN even though the physical link is good, and the MAC addresses don't match. I am not sure if they did before, but in my other still-working node these equivalent values are the same. I'm not sure how that would've changed from what I did or how to make the MACs match, if that's the problem. Advice or other suggestions would be appreciated.
Originally, my /etc/network/interfaces file looked something like this:
Code:
auto lo
iface lo inet loopback
iface enp8s0 inet manual
auto vmbr0
iface vmbr0 inet static
address 192.168.0.116/24
gateway 192.168.0.1
bridge-ports enp5s0
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4092
I have an existing VLAN 10 on the switch this node is connected to, so I update the file to read like this:
Code:
auto lo
iface lo inet loopback
iface enp8s0 inet manual
auto vmbr0.10
iface vmbr0.100 inet static
address 192.168.0.116/24
gateway 192.168.0.1
auto vmbr0
iface vmbr0 inet static
bridge-ports enp5s0
bridge-stp off
bridge-fd 0
bridge-vlan-aware yes
bridge-vids 2-4092
This works! I connect the switch to a port that accepts tagged traffic for VLAN 10 and can access it from other ports that belong to the same VLAN. So far, so good.
Other devices use VLAN 10 however, and I want my management interface to be on it's own dedicated VLAN. I set up VLAN 100 on two ports for the switch (one connected to the node, the other to my main router where my PC is connected). Then I add VLAN 100 as an interface in the proxmox GUI. I haven't re-assigned the IP to it yet and while the VLAN exists in the switch no traffic should be getting sent on it yet as the PVIDs are all still set for VLAN 10.
Regardless, as soon as I hit the apply configuration on button in the Proxmox GUI the node's network connectivity goes down and stays down. I have to connect a monitor to it so I can access the node directly and undo all my changes to /etc/network/interfaces until it looks like it did originally. Even so, I can't connect to or ping the node regardless of if it's connected to the switch, my main router (which has no VLANs at all), or with a direct connection between the node and an end terminal like a laptop.
Here's a snapshot of the output from running ip addr on the node: https://imgur.com/a/aa2iXlg
Two things I notice are that the NIC and vmbr0 both show a perpetual state of DOWN even though the physical link is good, and the MAC addresses don't match. I am not sure if they did before, but in my other still-working node these equivalent values are the same. I'm not sure how that would've changed from what I did or how to make the MACs match, if that's the problem. Advice or other suggestions would be appreciated.
Last edited: