WAN was restored

[vnangia]

AC68U, with a dual WAN setup, with failover and failback enabled, normal (not aggressive) DHCP query frequency. Under 378.55, I would lose connectivity every few seconds and I'd see an entry in the system log that read like so:

Jul 31 22:32:11 WAN(0) Connection: WAN was restored.
Jul 31 22:32:26 WAN(0) Connection: ISP's DHCP did not function properly.
Jul 31 22:32:26 stop_nat_rules: apply the redirect_rules!
Jul 31 22:32:44 WAN(0) Connection: WAN was restored.
Jul 31 22:32:44 start_nat_rules: apply the nat_rules(/tmp/nat_rules_vlan2_vlan2)!
Jul 31 22:33:01 WAN(0) Connection: ISP's DHCP did not function properly.
Jul 31 22:33:01 stop_nat_rules: apply the redirect_rules!
Jul 31 22:34:08 WAN(0) Connection: WAN was restored.
Jul 31 22:34:08 start_nat_rules: apply the nat_rules(/tmp/nat_rules_vlan2_vlan2)!
Jul 31 22:34:20 WAN(0) Connection: ISP's DHCP did not function properly.
Jul 31 22:34:20 stop_nat_rules: apply the redirect_rules!
Jul 31 22:34:57 WAN(0) Connection: WAN was restored.
Jul 31 22:34:57 start_nat_rules: apply the nat_rules(/tmp/nat_rules_vlan2_vlan2)!
Jul 31 22:35:21 WAN(0) Connection: ISP's DHCP did not function properly.
Jul 31 22:35:21 stop_nat_rules: apply the redirect_rules!
Jul 31 22:35:59 WAN(0) Connection: WAN was restored.
Jul 31 22:35:59 start_nat_rules: apply the nat_rules(/tmp/nat_rules_vlan2_vlan2)!
Jul 31 22:36:18 WAN(0) Connection: ISP's DHCP did not function properly.
Jul 31 22:36:18 stop_nat_rules: apply the redirect_rules!
...

I reverted to 378.54_r2 and found that while there were similar entries, the connection didn't actually go down - it kept functioning. I've now done a hard reset to stock and re-set up the device from scratch, but still experiencing the connection loss under 378.55, but not 378.54_r2. Any thoughts?

ETA: The only thing different that I can see on the DHCP side in the changelog is that you can now have 31-day leases.