Testing: RBR Reboot unreliable, IPv6 uses more memory on boot loader 2016.01_rtm-6.0.2.21

Today I did some testing on IPv6, memory usage on the router, and satellite hop count/connectivity issues.

As of this morning, my RBR 960 and 2xRBS 960s were up 9 days - all working fine:

RBR: 10.7%, memory 622MB (very consistent for a number of days);
2 x RBSs - CPU 6%-7%, memory 375MB, hop count 1

No restarts on any device - all show up for 9 days.

So this is what I did and the results:

1. Disabled IPv6. All Still working fine, RBR memory stayed around the same 620MB.
2. So I rebooted via the router admin page. By about 8 Minutes of reported uptime on the RBR, both RBSs reported Blue. The RBR admin page both showed the satellites with "Good" status in Attached devices list.

Yet ---

(a) Both RBSs showed hop count blank!
(b) I tested a couple of devices on the 3rd floor via WiFi, they had no Internet access
(c) Tried a Ookla Speed test on my phone and Mac - could not find a server. Browser fails to access the internet too.
(d) Yet as reported from my Mac and phone, the WiFi signal was strong, and the Mac showed transmit rate > 1.4 Gbps.

3. So Next I powered off the 2 RBSs. Waited about 3-5 minutes.
4. While the RBSs were off, I turned on IPv6 again. RBR memory usage went from 350MB to 440MB almost immediately.
5. Next I powered on just the 3rd floor RBS. 5 minutes after that the router is at 487MB. I suspect it's going to ~ 520-540MB.
6. Once the RBS on 3rd floor turned blue. I checked things. Hop count is still blank! Nothing can connect to it still - no Wifi on 3rd floor devices.
7. I waited another 6 minutes to see if the RBS sorted itself out. Nope! Hop count blank, nothing can connect to WiFi on the 3rd floor still.
8. So at this point I powered off that RBS again, and power cycled the router - I think it left the running satellites in a state that they can't connect to the internet or devices can't connect to them. Once the router was up, I powered on the 1st satellite. Once I got the "blue" LED, I tested. All was good! But literally 1 minute later, I noticed my iPhone went to 5G, so I checked - hop count was blank again! But then, 1 minute later again, I noticed my iPhone was back on WiFi - and hop count was 1. Crazy - just a temporary glitch.
9. I went ahead and powered on the other satellite. That came up fine as well.

Once everything was up, I tested everything. The router has been up for an hour. Its CPU is at 17%, but it will continue to drop until it gets around 10%. Memory is at 514MB.

RBS #1 is at 12.89% (that will drop too), memory at 358MB, hop count 1! RBS #2 is at 11.64% (that will drop too), memory at 371MB, hop count 1 as well! IPv6 is working fine too. Backhaul status is "Good" via Ethernet for both satellites.

Conclusions:
A. Reboot of the router doesn't work - I suspect the boot loader version 2016.01_rtm-6.0.2.21 - it leaves the running satellites in a state that they can't connect to the internet, or devices can't connect to them, even though devices show a strong WiFi signal, and router web page shows backhaul status of Good. The tip is hop count blank on the satellite debug page - if you see this, no connections are going through that RBS.
B. IPv6 enabled on the latest v6 boot loader, 960 series at least (maybe other devices), uses about 175MB more memory - instead of 350MB it's 525MB give or take. Once it gets there it's stable, no obvious memory leaks. I did find if I "enabled" IPv6 again while it was already enabled, it went up to 625MB and stayed there for days - stable, but that's possibly a bug that whatever data/cache etc the router created for IPv6, it did it again leaving the old memory still there.

****SO DON'T DO A REBOOT VIA THE ROUTER ADMIN WEB PAGE - ALWAYS power down satellites, then router, wait 2 minutes, power up router, when it's up power up satellites one by one.