Open danielschonfeld opened 4 years ago
Thanks for the report. Could there be any relation to any other event on the system or does it happen completely at random? If anyone else has seen this happen please provide any details you can about the circumstances so I can try to reproduce it in case it's possible on my hardware.
I knew when posting it, this wasn't going to be enough to go by. I'll update if i find more data points to learn from. Of course now, after having done it 3 times in a row, it ceased doing it.
I want to say that before, we only had one peer, namely the peer that had the endpoint dropped. And it was set up to be a "client" tunnel, channeling all traffic thru WG0.
Also, I have a feeling, tho I can't prove it, that the endpoint dropping had to do with rebooting the device and taking away power multiple times.
Will update further if I find something concrete as far as reproducing
I can confirm the same results after a reboot by power cycle or reboot command. The remote peer does not come until the peer is deleted and re created.
I've tested on an ER-X running 2.0.6 and e50-v2-v1.0.20200520-v1.0.20200513 and most versions going back to wireguard-v2.0-e50-0.0.20191219-2
It seems that I can only replicate the problem when PoE Pass through is enabled on eth4. If I turn off Passthrough the issue goes away.
fwiw in my case ETH4 passthrough was OFF. Weird...
also for me, all i had to do was issue a new set interfaces wireguard wg0 peer XXX endpoint
and it took it, and reestablished the tunnel. No need to delete the peer for me.
Periodically and haven't yet been able to pin point what triggers it, the endpoint configuration item vanishes from the configuration. This ofcourse renders the interface dead and not working. Using a URL not an IP for the endpoint.
Using version
e50-v2-v1.0.20200520-v1.0.20200513
on anER-X with 2.0.8-hotfix1