Identify which interface Proxmox is using for ping request

[logui]

I have Proxmox installed on an appliance with 6 network interfaces, the management interface is vmbr0 who is linked to the physical interface enp1s0, I have OpenWrt running on a VM, using also vmbr0 as the LAN interface. The second bridge vmbr1 is linked to enp1s0, and it is being used by OpenWrt as the WAN interface.

When I perform a ping google.com in the Node's console I get an IPv6 address as a response instead of an IPv4 address, that tells me Proxmox is using vmbr1 for the ping instead vmbr0, is there a way/command that will help me trace which interface Proxmox is using for internet access?

In theory Proxmox itself (not the VM) should be using vmbr0 (enp1s0) for internet access, also, the Proxmox default gateway is an IPv4 address (OpenWrt router's address). Somehow Proxmox is skipping the router's LAN network (vmbr0) and directing traffic through the WAN directly (vmbr1), therefore that traffic can't be filtered by the Firewall.

I need a way to track this down, any help will be appreciated, thank you

 

[_gabriel]

post your pve /etc/network/interfaces + ip a
into CODE tags please.

 

[logui]

post your pve /etc/network/interfaces + ip a
into CODE tags please.

I was able to find two workarounds:

Use PCI passthrough for the WAN interface assigned to the VM, Proxmox can not see the interface, unfortunately with this one I can’t migrate the VM

Therefore I default to fully disable IPv6 in Proxmox following the sysadmin documentation, with this option no IPv6 is assigned to the vmbr1/WAN interface, therefore Proxmox can’t use it, and it defaults to use vmbr0/LAN interface using IPv4.

My configuration is default, cluster with 3 nodes and vmbr0 and vmbr1 on each node, that way I can migrate the OpenWrt router VM across nodes for node maintenance with almost no downtime .

To me, this is an issue or situation where Proxmox/host is prioritizing IPv6 over IPv4 for its own network activities, like apt, ping, etc.