AC5300 Dual WAN does not work

[endurance]

I am using Merlin 386.7_2 on a RT-AC5300. trying to setup dual WAN with fallback.

As long as the primary WAN is up all works. If I manually reconnect the Prim WAN port to my secondary WAN gateway this works as well (just to check connectivity).

First issue:
Now first problem if you setup dual wan failover the secondary wan port stay in cold standby. A failover does not work. Manually disabling and enabling internet on secondary WAN brings it up to hot standby.
Should it not always be hot standby? Note using DHCP on both connections.

Second issue:
If it failover to secondary WAN no traffic is working and it states not internet connection (but there is secondary WAN is working wo issues).
Only after fallback WAN traffic comes back to normal.

What is the wanted designed behaviour for case 1?
Someone who got the failover and fallback working? any tipps?

thanks in advance

 

[endurance]

I changed DHCP to static but no change in behaviour

If prim WAN looses connection secondary does not pick up it show no internet connection (but I can connect to the management IF via this network so connection is there).

ping from WAN router works as well
Pinging 192.168.59.53 with 64 bytes of Data through LAN
Receive reply from 192.168.59.53, time=1.5ms
Receive reply from 192.168.59.53, time=10.0ms
Receive reply from 192.168.59.53, time=10.0ms
Receive reply from 192.168.59.53, time=10.0ms
Receive reply from 192.168.59.53, time=10.0ms
Packets: Sent = 5, Received = 5, Lost = 0 (0% loss)
rtt min/avg/max/mdev = 1.5/8.3/10.0/3.4 ms

 

[Smokey613]

Thread 'Dual WAN Failover Script ***v1.6.0 Released***'
https://www.snbforums.com/threads/dual-wan-failover-script-v1-6-0-released.82623/

 

[endurance]

Testing - thanks for the hint

 

[endurance]

Tested the script and checked also the native dual wan failover. None of both worked for me and I think it is related to criterias to switch over which seems purely on ping yes/no and ideally for longer duration.
So the easy case is now working with the native version. Means if for several minutes the prime WAN is no longer connected it fails over and also fallback works.
In most of my cases the prime WAN has only degraded performance and partial drops. So I would need something build on % of failures and latencies.