PS
I reread this thread. My pfsense has a point to point network to my layer 3 switch. There is no asymmetrical network. pfsense forwards all traffic to my layer 3 switch so the L3 switch can route it.
All my web pages now pop faster on the screen when using the RV320 router than when using my pfsense router on an old Xeon server motherboard. Two years ago when I tested pfsense using testing web pages it was as fast or faster than the RV320 which now no longer holds true. So I have moved back to my old Cisco RV320 router. I just unplug one router and plug the other one in and I can tell the difference. I do miss the time server in pfsense.
Thank you for help. We use static routing earlier but from time to time something is change so we want enable any routing protocol. Cisco routers also use ripv2. I resolved my problem. I added parameter "passive" in /etc/gateways to interface which i don't want advertise.
Probably this, assuming you're on net30 topology which you appear to be:
https://doc.pfsense.org/index.php/Why_can%27t_I_ping_some_OpenVPN_adapter_addresses
Disable monitoring for that, or make sure you have a route pushed for the tunnel network and then setup a monitor address for .1 in the tunnel network, which might work.
As long as all the pfSense interfaces are separate with separate gateways (such as /30 networks with one address on each side) you should be able to put them all into a load balance group.
