Put Modem in Bridge Mode/ Disable CG-NAT and another error

[JT Strickland]

RT-AC86U 384.16 B3, RT-AC68U aimesh node w/ 384.16, Diversion, uiDivstats, Skynet, AiProtection, Scribe, uiScribe, Conmon, spdMerlin, scMerlin, Nsrum, 250/10 broadband with Netgear CM-1000 docsis 3.1 modem.

I couldn't log into my router this morning, like it was froze up. I switched it and the modem off then back on, then had a lot of these in the log:
Apr 13 09:52:30 RT-AC86U-8F38 kernel: br0: received packet on eth3 with own address as source address

And this is what was in the skynet log, and quite a few of them:
Apr 13 07:58:38 RT-AC86U-8F38 Skynet: [*] Private WAN IP Detected 192.168.100.10 - Please Put Your Modem In Bridge Mode / Disable CG-NAT

We had some strong storms yesterday, and the power and cable internet was off for a few hours.
I found a post on the dd-wrt board doing a search, and they said they resolved the first issue by changing the MAC address for eth3 and w1.1 in the Port Setup of the bridging section, and it worked like a charm. It may be the power outage that brought it on, but I have had the "own address" error for a while, just not near as many in a row.

Can anyone confirm this is what needs doing, and how do I go about doing it? Can my ISP put the modem in Bridge Mode and disable CG-NAT? Also, is changing the MAC address something I can do, or should I get someone to do it? It has also been eth5 in the past, now eth3, so I don't know. I could sure use some guidance.
TIA,
jts

 

[ColinTaylor]

The first message can be ignored whilst your network boots up. Once everything settles down they should disappear. If they don't it could indicate a troublesome or misconfigured client or a client that is "bouncing" between different APs.

The second message can usually also be ignored. When a cable modem boots up it typically hasn't connected to the internet so it gives the router a temporary IP address from its own subnet (192.168.100.x). A short while later once it has connected to the internet the interface is reset and the router is given its "proper" IP address. If the cable modem can't connect to the internet (because of ISP issues) you will often see the modem constantly restarting itself until the ISP resolves the upstream issue.

So this isn't CGNAT.

 

[JT Strickland]

I appreciate it, that makes me feel better. I looked back closer at the times and that does look like the case for the second message, but the first one is more random. It was a lot of times in a row before I shut if down. I wonder if my switch placement between the router and the node could be the culprit? I attached a sketch of my setup now and what it could be, or one way. I did it that way at first to have more expansion room and take a little work out of the node. My network has been running so smooth lately, or before the storm. Thanks again.

 

[ColinTaylor]

The switch placement is fine. However, if you were to connect the router to the node by Ethernet as well then that would create a loop which would be a problem.

The second message has been cropping up on Broadcom based routers for over a decade now! IMHO these are usually caused by clients switching from one AP to another. The router receives a MAC from the client on one interface and thinks "hang on a minute, didn't I just talk to that client on my other interface?".

So you can expect to see those messages when clients roam from one node to another, or one AP to another (i.e. 5GHz to 2.4GHz). If you get a lot of them (for stationary devices) that might indicate that the nodes/APs are too close together, or the roaming isn't setup correctly.

Unfortunately, a lot of this behaviour is down to the client and not the router/AP.

 

[JT Strickland]

It may be me that's causing that. My little office is less than 100' from our house, and whenever I go back and forth it switches my iPhone, and anything else I might be carrying, from the node to the router and vice versa, sometimes several times a day, and the same with my wife when she walks the dog or something, so that may be it. It hadn't occured to me.

It is nice, though, to just keep right on going and stay on the same network, albeit with two different hosts. If that is all the problem is, then it is worth it. I had two 68U's for nodes but decided I didn't need b0th and put one in the closet.
thanks again,
jts