No WAN address after reboot with PPPoE VLAN7

[Lumpy]

I've got some issues with my PPPoE-Setup with the 26.1 series. Afair, this wasn't an issue with the RCs but started with the final 26.1 and is still present with 26.1.1. I'm not entirely sure what I'm looking at or how to debug this.

I use PPPoE with VLAN 7 on a fibre connection provided by Deutsche Glasfaser but with 1&1 as the ISP. Everytime I reboot my appliance it doesn't come up with an ip address on the wan interface. I have to restart the ONT to get it working again.

Log is attached but I don't see anything really unusual at first sight.

[meyergru]

That looks like an unstable ethernetconnection, probably because of using a RealTek card. Did you install the vendor driver os-realtek-re?

[Lumpy]

Yes, "os-realtek-re" is used, but this wasn't an issue with the 25.7 series. I had it running with 25.7 for months without issues. I can't get my head around it why it is needed to reboot the ONT to get everything back up and running. :(

Is there anything else I can do to help debug this?

[meyergru]

Did you try to disable and re-enable the WAN interface instead? It is strange that the ONT must be restarted, I would guess that this only initiates a retrain of the ethernet connection, which could also be triggered by a short cable disconnect. If that does not suffice, then it really is an ONT instability.

[Lumpy]

Yeah, I tried to disable/enable the WAN interface but that didn't do the trick. Rebooting again doesn't work either. It's stuck until I restart the ONT. But like I said, this wasn't an issue with 25.7 and isn't an issue with Mikrotik RouterOS on this device.

At first I thought it has something to do with hostwatch because it started with the newer hostwatch versions but that doesn't really make sense because disabling neighbour discovery doesn't do the trick once wan is stuck.

I'm a bit out of options here.

[meyergru]

If only an ONT restart does the trick, it is definitely the ONT. If it started with 26.1 only, then there must be something that "knocks out" the ONT - I doubt that it is something that has to do with the way how OpnSense handles the internet connection.

I would try to disable the hostwatch service completely - matter-of-fact, I still some some very strange behavior when it is active - like suddenly, all IPs in any subnet become active.

If the ONT gets overwhelmed by many ARP neighbors, for example, it might get a problem. ONTs usually see only two MACs.

Do you have an ONT whose management interface you can reach via IP? Is it still active when the problem arises?