Load balancing OnDemand based on primary WAN load

axsense · Feb 13, 2018, 2:20 PM

Hello,

Use case:
WAN1 100/100
WAN2 50/50

I would like to configure Load balancing so that as long as the load is <80Mbit it completely uses WAN1, but above 80Mbit WAN2 is also being used for load balancing. When the load resumes below 80Mbit all traffic goes back to WAN1 only.
Is this possible? The best I know of is to use Weight parameter to prefer WAN1, but it is not the same.

Axel.

Btw: Regarding best practices, why there are three port groups when LB is being used? (one for LB, two for failover WAN1->WAN2, WAN2->WAN1). Usually in clusters when one member is down, then it is out of capacity. When it is back online, then it is sharing the total load. Why I should create three ports groups for two WANs in this use case?

axsense · Feb 13, 2018, 7:39 PM

According to my searches, PFSense only uses Round Robin. So the use case of load level based LB is not an option. Maybe there is a service for that(?)

And for the second question: those failover groups are absolutely unnecessary if you only plan to use Load balancing. I think that quite a many PFSense has those unnecessary configurations just because they are mentioned everywhere just for an example of three kinds of setup.

It's a bit shame that PFSense has nothing but round robin.

Axel.