Another possibly somewhat similar issue to report.
In this scenario there's a "server" wg system that 3 peers connect to
in a site-to-site configuration; all systems are running OpenBSD
-current. No keepalive alive necessary as the sites are active 24/7.
This was setup when wg first entered the OpenBSD kernel and has
basically been flawless except for one hiccup which wasn't present
early on. When I originally set this up the peers all connected via an
alias address (wgendpoint) on the server and there was no issue with
reconnecting if a system went down and came back up, either due to an
upgrade, power outage, etc. At some point (it's been a while) when I
would do system upgrades to the current -current some of the peers
(usually 2 out of the 3 and not always the same ones) stopped
reconnecting upon reboot. Rebooting or restarting the wg interface
does not resolve the issue, but if I edit the peer's wgendpoint to use
the "servers" other address and restart the wg interface the
connection is established. By other address I mean the address the
peer wasn't using previously, if the peer was previously (before the
upgrade) using the alias I need to change it to the main address, if
it was previously using the main address I need to change it to the
alias.
Not sure what I would have to do if I didn't have two addresses to
switch between.
