Need some help from the geniuses here

TL;DR LAG not working(ish) on MB8600 to PFSense WAN.

I recently got the Motorola (Zoom) MB8600 cable modem with the sole purpose of being able to LAG port 1 and 2 attached to my PF box.

I was able to get the LAG to work (sorta) with some notable exceptions. Here is where I need help. With the LAG setup I am maxing out at 760 Mbps from the PF box to the interwebs. With LAG turned off over 1 cable I am able to get 928 Mbps (don’t care about upload).

So I started to do some testing. I have attached a link to a quick diagram to hopefully help everyone understand how I have it setup. https://imgur.com/gallery/3DIPnKH

First, I tried to determine if possibly one of the cables was bad. When I unplugged port 1 going from the modem to the router everything seem to work ok. The connection stated alive but the speed did not increase. I plugged port 1 back in and did the same thing with port 2. Very interesting result, no connection. PF was not able to pull an IP address. I changed out the cable and had to same issue. Ran diagnostics on modem and the port is fine (also same port that is used when LAG is turned off).

To double check that the port was ok I turned off LAG and removed the LAG interface from PF and switched over the WAN to one of my Intel NIC ports. Tested everything. Worked just fine with speedtest coming back with 928 Mbps.

Why would I get capped out at 760 Mpbs with PF when I was able to find others out there with Linksys and ASUS getting 1.1+ Gbps (Due to over provisioning). Right now I would be happy to LAG working with 940 Mbps. I am not thinking that LAG is magical and will get me some crazy more bandwidth. I started this because I wanted to experiment with LAG and now I am just at a loss. Please help! TIA

1 Like

I have never tried using LAG with pfsense, try their forums.

I hate to be the bearer of bad news but that is not how LAGs work. LAGs can be used to provide link redundancy and increase bandwidth (aka traffic lanes) but do not necessarily increase throughput. The only way you would see throughput gains by implementing a LAG is if a single link is under a heavy enough load from multiple sources. If you connect a single computer to a LAG built from 2 1Gb links it does not have a 2 Gb connection, but a choice of which 1Gb link to use. If you connect 2 computers to a LAG built from 2 1Gb link each computer could use a separate 1 Gb link giving is access to a full 1Gb instead of having to share a a single 1 Gb link. High end storage typically prefers to use MPIO over LAG for this very reason.

Now given the right firmware in the modem it is possible that the modem itself is increasing its internal throughput because it is detecting two active links, however that would still require the carrier to allow that additional throughput, perhaps in the form of bursting.

1 Like

Thank you very much for that explanation. Do you have any idea why a speed without lag would provide 940 Mbps down and with lag would top out at 780 Mbps?

Hard to say for sure without testing each leg of the connection independently, as there are to many internal and external factors that come into play each time you run a speed test. However if I had to guess I would imagine that it would be the modem causing the slow down. If you really wanted to figure it out you could run iperf across each leg of the connection to see where the slowdown is.

Yes have this same problem too… And I think it’s not recommended to LAG the WAN connection too.