Multi-WAN not failing back over properly after one uplink fails

[pjw]

I recently upgraded to 25.7 and things have been mostly fine.

My setup is:

- WAN uplink to broadband modem
- WAN uplink to Starlink

I have multi-WAN rules setup to push my work traffic out my Starlink WAN primarily.  It involves an alias with my work's Wireguard target endpoints, and I push anything bound to them to the Starlink WAN.  Everything else in the house defaults to the broadband WAN uplink.  When either link fails, the traffic is configured to failover to the other WAN.

After upgrading to 25.7, when my Starlink WAN uplink fails, traffic isn't properly failing over.  Also, when the Starlink WAN comes back up, the traffic isn't failing back at all.  All my Wireguard tunnels stay down.  This is using either UDP or TCP.  I have to reboot my OPNsense box to unwedge things.

This does seem like a regression in 25.7 with multi-WAN.  I'm happy to provide configs or logs to help debug, or try patches.  Just let me know how I can help!

[TrustedComputer]

I have a couple of clients with multi-wan on 25.1, they are not upgraded yet (thanks for posting!). Please do update when you find a fix.

[franco]

Fact: the multi-WAN code on 25.1.12 and 25.7.x is exactly the same.


Cheers,
Franco

[TrustedComputer]

Thanks, franco!