I have made a switch from PfSense to RouterOS, so I am very new to Mikrotik. Setting up load-balancing and failover was a pretty straightforward task in PfSense. However, in RouterOS, I have come to realize there are many ways one could undertake accomplishing this. I would just like to know what generally is the best practice for setting up LB and FO for one PPPoE WAN and one static IP WAN. I would like one to ping 8.8.8.8 and the other 1.1.1.1 and not their gateways. For the finer details I will bother you all again, if I don’t end up resolving the matter on my own. Many thanks.
simply stating failover is not enough detail.
-simply stating load balancing is not enough
What is primary, what is secondary.
What do you want to have happen if primary or secondary faiils.
Why say failover and then load balancing>
Do you mean both are roughly equal interfaces Not a primary or Secondary but
perhaps one should handle more of the wan traffic by a a certain amount… but takes all the load if one fail.s
No specific lan or subnet requirements for the above scenarios??
Thank you for the quick reply. The PDF looks very informative. Although I already have done some research on ECMP, Nth LB, PBR, and PCC. I understand each have their own benefits and downsides.
Well
simply stating failover is not enough detail.
-simply stating load balancing is not enough
Why Failover? Redundancy: If WAN1 goes down (packet loss goes to 100% or latency above ~500ms) switch to WAN2 for default gateway. When WAN1 comes back online then switch BACK to WAN1 from WAN2. In PfSense this is to go both ways (as in Tier 1 to Tier 2 and Tier 2 to Tier 1). I also don’t want either WAN to ping their gateways (next hops). I want them to ping an actual address on the web.
Why Load Balancing? Utilizing two pipelines at the same time: If an app has the capacity to open enough connections, then both WANs should be available to it (such as an IDM). If load balancing automatically addresses failover in Mikrotik then the need for failover essentially becomes moot. But in PfSense default gateway would not automatically switch, unless specified under routing rules. That is why I was asking for both.
What is primary, what is secondary.
What do you want to have happen if primary or secondary faiils.
Static Primary. PPPoE secondary.
Do you mean both are roughly equal interfaces Not a primary or Secondary but
perhaps one should handle more of the wan traffic by a a certain amount… but takes all the load if one fail.s
Assigning a weight ratio (as in 2 to 1) was beyond the scope of what I wanted to ask on the get go. But should this matter, primary’s connection speed is 50% faster than the secondary. Overall, it wouldn’t matter if a connection goes from one or the other (but it has to come back through the same gateway and go out the same [such as for banking connections [anyway, this is besides the point and easily done with mangle]).
No specific lan or subnet requirements for the above scenarios??
For LAN subnets, you don’t need to intervene. As the implementation in that link relies on interface grouping/lists instead of per subnet basis (which is equivalent to a brain seizure), so you can create a LAN interface list containing all your bridged VLANs or whatever and it will work automatically.
Sorry it took me a while to get things working. I am wondering what is the benefit of assigning TCP,UDP (80,443) to PCC and the rest to Nth. Utilizing only PCC across the board seems - at first glance - to function as expected. Looking forward to your input.