Losing ethernet/fiber port Connectivity on 4600 AP Intermittently

Products ePMP
1

We’re having a strange problem that we’ve seen happening for several months.

We have four 4600s at one site. Initially, we were feeding those with CAT5e cable. Then we ran CAT6a and used SFP+ RJ45 modules in a MikroTik CRS309 switch and SFP+ RJ45 modules in the 4600s to feed the 4600s with up to 10 Gbps data. This did work, but one of the 4600s (we call it NWT2) would ‘disconnect’ data every few weeks. The link would stay up, but the data would stop flowing. We’d ‘bounce’ the port in the switch a few times, and the data would start up again.

We found that the SFP+/RJ45 modules were running very hot (which actually is normal) and we thought that this was causing the problem. So, we put some cooling on the switch and that seemed to have stopped the problem. Then it did it again.

Finally, we ran fiber from the bottom of the water tank to the top and are feeding all four 4600s with fiber. We’re still using the MikroTik switch at the bottom. This seemed to have stopped the problem, but again today, this happened again on the same 4600. The link was up, but data stopped flowing. I ‘bounced’ the port on the switch a few times and it came back to normal.

We replaced the 4600 and the MikroTik switch and the POE feeding the 4600 (over the old CAT5e - of course that cable is only being used to power the 4600, with data disconnected). And we’ve used different ports on the MikroTik switch, with no difference. The problem is still occurring (just much less often than before). It’s only happened once now since switching over the fiber into the 4600.

All four 4600s are set up the same, through the same switch. Yet only this one 4600 is causing a problem. All of the others are fine.

The only thing that I’ve noticed is that when this happens, the 4600 log shows several instances of SMs losing communications with the AP (which also doesn’t make sense to me) - “2026.07.21 17:54:54 +00:00 NWT-2-6 kernel: [1818069.728323] SM[bc:e6:7c:92:ef:7d] aid=14 peer=77 disassociated. Reason: 48 (COMMUNICATION LOST)”

Also, this is probably our heaviest-loaded AP with 41 SMs (4625 and 4616) on it as of today. We do have other 4600s with as many as 39 SMs on them, and have never seen this problem. But, those 4600s are still being fed with Cat5e, so the load on them is much less than this one that keeps giving us trouble.

Any suggestions on what might be happening and how to fix it?

2

I’m not sure if it’s exactly the same issue, but we are experiencing it with a PtP link using the 4600c.

Data flow stops intermittently; we’ve already replaced the cable and the SM radio, yet the problem persists.

We are going to replace the 4600c AP to see if that resolves the issue.

This has been happening for several months—the link is intermittent.

We’ve tested various firmware versions, and the same thing happens.

We’ve tested both TDD and ePTP modes, and the same thing happens.

We’ve changed frequencies, and the same thing happens.

We’ve changed the channel width, and the same thing happens.

The difference is that you are experiencing this with a PtMP link, whereas we are dealing with a PtP link. :frowning:

3

Hello @John_Rayfield ,

could you please download a techsupport file right after you have exercised the port and have restored the connection to your AP.
Just PM it to me.

4

They are letting us know how the inspection went. We already decided to replace the equipment, and everything is working perfectly.
We tried swapping the SM, but since the issue persisted, we decided to replace the AP, and now everything works perfectly. It wasn’t a cabling issue. :frowning: