AC68U CPU in status page seems always high, but top seems fine?

[BatKing]

Hi,

my AC68U is running merlin 384.19 as AiMesh Router and another AC68U running as AiMesh Node. in the network map status page, both CPU is running from 10% to 100% up and down constantly. but if I ssh into the router and run top, the load avg is about 2.xx. CPU temperature is 82C which I think is normal. I only setup AiMesh since about 10 days ago. I updated the ac68u router from a 4 year old merlin (380) firmware to merlin 384.19 just trying to run AiMesh. even before Setup another AC68U (asus official 385 firmware) as node, the router CPU in status page is jumping like in the image below. The router does running Diversion and UiDiversion and YuZhi guest Wifi plugins. but disabling diversion doesn't seems to do anything to the Status page CPU load.

but from top it seems normal?

Mem: 114984K used, 140692K free, 11984K shrd, 3068K buff, 32136K cached
CPU:  7.5% usr  6.8% sys  0.0% nic 41.3% idle 43.5% io  0.0% irq  0.6% sirq
Load average: 2.59 2.43 2.52 1/123 27528
  PID  PPID USER     STAT   VSZ %VSZ CPU %CPU COMMAND
2252     1 nobody   S     2764  1.0   1  2.4 avahi-daemon: running [RT-AC68U-XXXX.local]
  237   236 admin    S     4948  1.9   0  2.0 nt_monitor
  263   236 admin    S     4948  1.9   0  1.4 nt_monitor
  286     1 admin    S     7136  2.7   0  0.9 httpd -i br0
  314     1 admin    S     6356  2.4   1  0.2 mastiff
  239     1 admin    S     5156  2.0   0  0.2 nt_center

of course the node's TOP look much better which I take it is only running AP but nothing else.

Mem: 55684K used, 200016K free, 2344K shrd, 480K buff, 8888K cached
CPU:  1.2% usr  0.6% sys  0.0% nic 97.3% idle  0.0% io  0.0% irq  0.6% sirq
Load average: 0.01 0.10 0.09 1/97 9794
  PID  PPID USER     STAT   VSZ %VSZ CPU %CPU COMMAND
  225     1 admin    R     4976  1.9   1  1.1 nt_center
   35     2 admin    SW       0  0.0   0  0.2 [kworker/0:1]
8400  8388 admin    R     1396  0.5   1  0.1 top
   36     2 admin    SW       0  0.0   1  0.1 [kworker/1:1]
  561     1 admin    S     5776  2.2   0  0.0 networkmap
  298   296 admin    S     5628  2.2   0  0.0 amas_bhctrl
  272     1 admin    S     5628  2.2   0  0.0 amas_bhctrl

 

[L&LD]

If you upgraded from 4-year-old firmware without doing a full reset afterward, that is where I would start.

I would not be trying to get a stable AiMesh network from the 384.19_0 firmware today though.

Rather, flash the latest RMerlin 386.1 Beta 5 firmware first (which is stable for many and includes the greatly improved AiMesh v2.0).

Download | Asuswrt-Merlin (asuswrt-merlin.net)

Then, follow the suggestions below to get your network running fast, stable and secure.

New M&M 2020

M&M 2020 Specifics

How to Connect an AiMesh Node

amtm Step-by-Step https://www.snbforums.com/threads/amtm-step-by-step-install-guide-l-ld.56237/#post-483421
(Note: 'installing' amtm isn't required in the guide above if using RMerlin firmware 384.15_0 or later).

If doing the above suggestions doesn't help, then you may need to also follow the suggestions below to get your network and router to a good/known state and one where the router is using the expected defaults of the firmware currently installed on it.

Fully Reset Router and Network

 

[BatKing]

Of course the 1st thing I did is reset the router to factory. in fact I am using the other AC68U updated to merlin 384.19 as the router, then setup everything as the old router setting manually, swap out the router, update it to Asus official 385 for AiMesh 1.0 and reset, setup as the Node. (I know there is 386 for AiMesh 2.0 but I just want a taste of AiMesh and stability only but not new fancy AiMesh 2.0 features).

So everything is reset, manual setup. No old config from update at all. Besides other than the status page CPU load issue, my only other issue (I am not sure if it is a issue yet) is the main router overtime will kick all the wifi device to my node (node and router maybe too close by) but no wifi disconnection has ever been detected.

 

[ColinTaylor]

The CPU load is being caused by I/O. The most likely cause of this would be some process reading or writing to an attached USB device, for example the Media Server performing a disk scan.

 

[L&LD]

Seems to me like all your issues are from mixing firmware bases and also using outdated AiMesh code (as you point out too).

If @ColinTaylor is correct about an attached USB drive, and removing said drive brings stability to the network, great.

Otherwise, I would be flashing both routers to at least the same firmware level, even if using older code seems a bit nuts to me, today.

How is AiMesh set up? In wired or wireless backhaul mode? If wireless, can you wire it (if only just to test)?

The following link may help too.

AiMesh Ideal Placement

 

[BatKing]

The CPU load is being caused by I/O. The most likely cause of this would be some process reading or writing to an attached USB device, for example the Media Server performing a disk scan.

Thank you. This gave me a hint. I don't have any USB share device on the router. but I do have a 8.0 GB USB flash drive on 2.0 port only for entware and swap. but I do recall I read somewhere the uiDivStats does go through the DNS log like no other day and store the statistic somewhere (which I guess maybe the USB storage). so uninstalling uiDivStats solved my issue. My router's CPU load now is a happy and cool.

 

[BatKing]

Seems to me like all your issues are from mixing firmware bases and also using outdated AiMesh code (as you point out too).

If @ColinTaylor is correct about an attached USB drive, and removing said drive brings stability to the network, great.

Otherwise, I would be flashing both routers to at least the same firmware level, even if using older code seems a bit nuts to me, today.

How is AiMesh set up? In wired or wireless backhaul mode? If wireless, can you wire it (if only just to test)?

The following link may help too.

AiMesh Ideal Placement

My house can run with only one router but there are one spot where my phone only gets about -80 RSSI, even my 2011 MBP can get a stable -79 RSSI wifi and get over 100mbps connection from speedtest on the 5Ghz, my iphoneX will switch to LTE. this is the only reason I am getting into AiMesh. but to run Ethernet backhaul, I have to place the router and node almost directly up floor and down floor.

Surprisedly this work well for most of the devices. I have 2.4 and 5Ghz on different SSID. my smart devices (6 only on 2.4Ghz) stay connected without noticeable issues. my iphoneX will roam base on which floor I am at and this solved my dead spot issue mentioned above. The only issue is my main 2019 MBP will roam to the node down floor (-65 db) rather than stay connected to the router in the same room (-38db). At first, I thought that maybe the MBP's issue. but lately I am thinking maybe its the router issue.
Every time this happens, all other wifi devices connected to the router seems also disconnected and rejoin (in wireless log connected column time is reseted to 0X:XX) and my 2019 MBP stay connect to the node. but all the rejoined devices doesn't report any error. and I won't know my MBP roamed to down floor node without looking my wifi RSSI value. Even my work VPN doesn't report any error. it almost seems like the router did a roaming assistance kick to every devices on the Router (RA is always disabled, as I don't want to mix in something before I test out the current behavior, so not sure why) . and when every other devices still pick the strongest signal which should be the router, my 2019 MBP just stay connected to the Node without bother to roam to the strongest Wifi signal AP.
One more thing, every Wifi devices, if I turn off and on the wifi, they always seems connect the strongest AP (router or node) initially.

Since I don't get any wifi disconnection error at all and AiMesh does solve my house's wifi weak spot issue. I think I will wait for 386 merlin's stable official release