logging into the web gui via a second bridge and vpn

[leksand]

The problem with logging into the web interface via vmbr2 ip via a vpn located on a virtual machine hosted on proxmox itself

it hangs for a long time and then somehow opens.

it feels like there is a problem with routing and the return packets are not reaching.

adding a static route on the pve to the vpn clients subnet via the internal ip of the vpn server did not help

ping passes without delay
traceroute runs as needed - vpn gateway, address on the internal network without packet loss.

pve
eno1 and eno2 are in different vlans and do not overlap

cat /etc/hosts
127.0.0.1 localhost.localdomain localhost
10.245.0.246 pve.local
10.0.111.252 pve.local

cat /etc/network/interfaces
auto lo
iface lo inet loopback
iface eno1 inet manual
iface eno2 inet manual

auto vmbr0
iface vmbr0 inet static
address 10.245.0.246/24
gateway 10.245.0.254
bridge-ports eno1
bridge-stp off
bridge-fd 0

auto vmbr2
iface vmbr2 inet static
address 10.0.111.252/24
bridge-ports eno2
bridge-stp off
bridge-fd 0

vpn vm

cat /etc/network/interfaces
auto lo
iface lo inet loopback

allow-hotplug ens18
iface ens18 inet static
address 10.0.111.15/24
gateway 10.0.111.254
dns-nameservers 10.0.111.254

auto ens19
iface ens19 inet static
address pub-ip/28 (it is used to connect VPN clients)

 

[leksand]

there is a feeling that all virtual machines that have a network via vmbr2 (which is without a gateway and uses an eno2 network card) have problems

 

[leksand]

The first time I recreated a VM from the same template. Now, instead of recreating the VM, I deleted the network card and added a new one - and everything worked.

Before that, the network was working, Internet access seemed to be there - ping, traceroute were on the Internet, but apt did not work. Removing the network card and adding a new one solved the problem.

Is there any way to understand the reason? (the vm was from a template, debian 12)