TINC Restarting every few seconds

[zombielinux]

I've got two networks Site A and Site B.
Both pfsense instances are the default gateways for each site.

I will eventually be adding more hosts to this TINC mesh VPN network (hence why I'm using TINC, and not OpenVPN or IPSEC).

Site A config:
WAN: siteA.fqdn.org
LAN: 172.19.0.0/16
Lan IP: 172.19.2.1

Site B config:
WAN: siteB.fqdn.org
LAN: 172.25.0.0/24
Lan IP: 172.25.0.1

The gateways can ping each other, but can't ping behind each other reliably (172.19.2.1 can't ping 172.25.0.10 for example). Some hosts work intermittantly, others fail with "ping: sendto: host is down". That may be beyond the scope of this question.

The main issue I'm having is a random restart of the "tincd" process on Site A. The default logging level is too low to get anything useful. Even killing the process and starting it manually only gets a cryptic "Got TERM signal" with no indication of the signal source.

Could it be some buried watchdog somewhere? Both hosts are running 2.4.4-1-RELEASE and tinc-1.0.35