Router restart all of a sudden then problems

👍

The story continues...

This morning I got up and saw that nothing was connected to one of the two RBS850s.

I checked the memory/CPU, and memory was ok, but CPU was 27% - the other RBS was 8% and the router was 11%. All three stayed at their respective CPU usage levels for the next 2 hours.

That offending RBS has an Apple TV box plugged into it via Ethernet, that continued to work. It was just WiFi connections that were refused.

None of the RBR/RBSs rebooted overnight/this morning on their own.

Using the WiFi analyzer, I sat 5 feet from the offending RBS. The signal was very strong, hovering around 94%. The other RBS and RBR signal for 5 GHz are 30% and 22% in other parts of the house. As a reminder, backhaul is wired.

The other funny thing, the Hop Count is blank! The other RBS shows 1. What does no hop count indicate?

I unplugged the offending RBS for about 2-3 minutes, then plugged it back in. All seems fine again, devices are connected to both RBSs and the RBR. Walking around with my iPhone or MacBook, they seem to switch nicely between router/RBS.

Once the RBS was up, the CPU settled in at ~ 8%  - like the other RBS. Hop count showed as 1.

So all I know:
- For a mysterious reason an RBS stops connecting any devices via WiFi.
- CPU usage is markedly higher at a sustained level - ~ 27% vs. 8%. What's it doing?
- 5 GHz signal strength is very strong and accurate - so devices see it, I suspect they attempt to connect to it and are rejected?
- Instead of having a hop count of 1, it is blank.
- Power cycle fixed it.

---

donawalt wrote:

I checked the memory/CPU, and memory was ok, but CPU was 27% - the other RBS was 8% and the router was 11%. All three stayed at their respective CPU usage levels for the next 2 hours.

---

How often was the debug.htm page refreshed?  (My experience is that the debug page does not refresh itself.  Once displayed, nothing changes unless the user refreshes the page.)

27% is not an unusual CPU measurement.  Often, it appears that the process of displaying the debug page causes a temporary increase in CPU load and when the page is refreshed, the CPU reading is much lower.

I re-logged in and checked it over the course of an hour or so, so I thin kit was real. I would not have wondered about only 27% either, yet the other two were 8% (RBS) and 11% (RBR) and after the power cycle, the offending RBS was back to 8% too. So it seems something was amiss. I just checked now, 5 hours after the power cycle, and it's at 7%. So 27% seems like an anomaly (along with no hop count)

I did do a 5 minute packet capture from the RBS debug page before I power cycled it, I don't know if that will do any good/if anyone can decipher it...

Might swap RBS around to see if you notice this issue follow the one RBS or if it stays at this one locations.

Also what power setting do you have the RBR set for under Advanced Tab/Advanced Settings/Wireless Settings

Who is your ISP? Your issues sound a lot like what I had gone through most of last year while Xfinity was upgrading equipment in our area.  Once they finished last November everything settled down and I have not had any glitches or weird restarts since then.  It's hard to imagine what a modem could do to cause the Orbi to reboot, but from my experience it can happen.

Might try 50% on the power sometime. 

Ya this is a odd one for sure, You system should not be seeing these odd events. 

Maybe BullMoon has it, I am on Comcast, and they have been putting the mid-level split/higher upload speeds into the neighborhood. It was working before they pulled the firmware on the modems because of issues, so maybe they are making changes. Thanks for that!

Kinda of hard to believe the ISP changes would cause issues at the RBS though. 

True - although the RBR could be instigating it. Even that is a little  hard to believe as BullMoon said. I could see the modem causing the router to reboot. But the madness of what happens on soft reboots is all Netgear I think. Like I said at one time in the past, they really need to test/debug the setup/restart etc scenarios, on the forums this is the stuff that causes people problems. Unless you restart everything, or upgrade firmware, etc. in exactly the perfect way, the system has problems. That's not right.