Where would you start?

[Nebulaz]

TLDR: Internet goes down constantly. I am monitoring the outages, have updated firmware and cut the fat on AMTM add-ons, nothing works. How to proceed?

LONGER VERSION:
My internet seems to go out repeatedly through out the day. The issue, I don't know, and that's what I'd like help figuring out. However, I do know, that if I reboot the router (Asus - 86U, w/ Merlin 384.17) upon reboot, the issue is resolved.

At an unknown time later, sometimes immediately, sometimes not, it goes down again. Kids are having fun with it (slightly) as I have Net Uptime Monitor running on a machine, and it makes what they like to shout when they hear it, "dad, the wifi fairy farted again!".

In my ideal set up:

• NordVPN configured as a client on the router UDP w/ Strict Policy, Not blocking clients in tunnel goes down, LAN DNS Filter policy Router.
  
• 2GB Swap file, (brand new USB drive, passes health check)
  
• Diversion Standard (4.1.11) w/ default blocking list. Pixelserv-TLS set to 192.168.1.2
• Skynet (7.1.6) default blacklist, no added whitelist.
  
• YazFi (4.0.3) Running three, 2/4GHz Guest Networks, with values, Enabled = True, all with different ranges, 168.2.0, 168.3.0 etc., DHCP 2-254, DNS is my gateway, redirect to VPN = true, VPN client = 1, Two / One way = false, Client Isolation = True.

Ideally, I'd have all of this running. The issue has been ongoing since March, I was running 384.15 at the time it started. I have since did a dirty .16 and a clean .17 install, nothing got better. During the fresh install of .17, I added one thing at a time, running just Diversion, just Nord, just Skynet and it occurred without any installed and with each installed.

I'd like to figure out if its the modem or router. Modem, doesn't need to be reset router does. Internet goes out regardless of load, time of day and connect clients. I have monitored during two 4K streams and 15 clients AND basic web browsing with one client.

If this was your network, where would you go next. I do not have another router and modem for substitution testing. Is this the only next step? Are there log details I can look for?

Really appreciate any guidance.

 

[dave14305]

Is the internet connectivity truly down or is it a DNS issue?

I would troubleshoot without a VPN Client, but you may not be in a position to do that.

If you want to post your syslog to pastebin and share it here or via PM we could look for issues.

 

[ugandy]

i had same symptom for a while. not dns since ping test does not use dns(?). in my case i solved it when i noticed that whenever this happened my upstream was fully maxed by one of the clients (causing degradation on downstream for everybody else). enabled QoS, problem solved. your problem may be different.

 

[Nebulaz]

The issue occurred a few minutes ago, Net Uptime Monitor logged it at 15:46:41, however it only runs every ten seconds. There is a weird issue after it goes at 15:47 in the log, where the details go to May 4 and 2200 hours, then once the reboot process is finished everything goes back to May 3.

<iframe src="https://pastebin.com/embed_iframe/94Ds0YAC" style="border:none;width:100%"></iframe>

 

[dave14305]

Better link to pastebin: https://pastebin.com/94Ds0YAC

A lot of noise in the log about auto-channel changing. Can you setup fixed wireless channel numbers?

 

[Nebulaz]

Okay,

Made the following changes to static:

2.4GHz - Channel Bandwidth: (20MHz), Channel: 11
5.0GHz - Channel Bandwidth: (40MHz), Channel: 36

Anything else in that log seem out of the ordinary?

 

[Nebulaz]

This error occurred several times as well.

May 3 15:51:00 avahi-daemon[2482]: Host name conflict, retrying with RT-AC86U-11
May 3 15:51:00 avahi-daemon[2482]: Registering new address record for 192.168.1.2 on br0.IPv4.
May 3 15:51:00 avahi-daemon[2482]: Registering new address record for 192.168.1.1 on br0.IPv4.
May 3 15:51:00 avahi-daemon[2482]: Registering new address record for 127.0.1.1 on lo.IPv4.
May 3 15:51:00 avahi-daemon[2482]: Registering new address record for 127.0.0.1 on lo.IPv4.
May 3 15:51:01 avahi-daemon[2482]: Withdrawing address record for 192.168.1.2 on br0.
May 3 15:51:01 avahi-daemon[2482]: Withdrawing address record for 127.0.1.1 on lo.
May 3 15:51:01 avahi-daemon[2482]: Withdrawing address record for 127.0.0.1 on lo.
May 3 15:51:01 avahi-daemon[2482]: Host name conflict, retrying with RT-AC86U-12

 

[Nebulaz]

i had same symptom for a while. not dns since ping test does not use dns(?). in my case i solved it when i noticed that whenever this happened my upstream was fully maxed by one of the clients (causing degradation on downstream for everybody else). enabled QoS, problem solved. your problem may be different.

Did you just enable the built in QOS or an addon?

 

[ColinTaylor]

Do you have another router, AP or repeater connected to this router? If so, turn it off.

Do you have any additional network switches attached?

 

[dosborne]

There is a weird issue after it goes at 15:47 in the log, where the details go to May 4 and 2200 hours, then once the reboot process is finished everything goes back to May 3

Not weird it is just using a "default" date until the NTP code kicks in on reboot. Think of it like the old Jan 01 1980 back in the DOS days.

 

[ugandy]

Did you just enable the built in QOS or an addon?

initially i just used the bandwidth limiter option and limited that client's upload to half my line limit. i've since switched to adaptive Qos and then added the freshJr addon. set my connection limits to 90% of real line speed. my upload no longer maxes and i no longer lose download connection. whenever the upload line saturated, i'd get the wan red light...
from your logs, looks like you have something else going on...

 

[Nebulaz]

Do you have another router, AP or repeater connected to this router? If so, turn it off.

Do you have any additional network switches attached?

I have a 68U in Mesh mode. I have a Netgear gs308e, with DHCP disabled, on the same subnet as the gateway, (I usually have separate vlans, but disabled them) with loop detection enabled.



Sent from my Nokia 7.1 using Tapatalk

 

[Nebulaz]

With the static channels, it happened again. I will take the 68U and the netgear offline for tomorrow for verification.

https://pastebin.com/DqK92VYw

 

[Nebulaz]

These three lines always occur right before the connection drops. W10.3 is my third 2.4GHz guest network.

May 3 21:18:00 wlceventd: WLCEVENTD wlceventd_proc_event(386): wl0.3: Deauth_ind 14:91:38:85:41:FA, status: 0, reason: Deauthenticated because sending station is leaving (or has left) IBSS or ESS (3)
May 3 21:18:00 wlceventd: WLCEVENTD wlceventd_proc_event(401): wl0.3: Disassoc 14:91:38:85:41:FA, status: 0, reason: Disassociated because sending station is leaving (or has left) BSS (8)
May 3 21:18:00 wlceventd: WLCEVENTD wlceventd_proc_event(386): wl0.3: Deauth_ind 14:91:38:85:41:FA, status: 0, reason: Deauthenticated because sending station is leaving (or has left) IBSS or ESS (3)

 

[dave14305]

These three lines always occur right before the connection drops. W10.3 is my third 2.4GHz guest network.

These are usually normal, to log that a device has left the WiFi network (too far, powered off, etc.). Do you know which device has that MAC address?

 

[ColinTaylor]

With the static channels, it happened again. I will take the 68U and the netgear offline for tomorrow for verification.

That's a good idea. The fact that your log is flooded with "br0: received packet on ethX with own address as source address" messages suggests to me that you have created a network loop.

 

[Nebulaz]

Both ancillary devices have been removed from the loop. So I have a static channel, QOS monitored, single routing device network.

Crossing fingers, wish me luck.

Not to look for bad omens, but my reading older threads suggests that my 2018 model year my be a stroke of bad luck with the 2.4 channel as well.

Sent from my Nokia 7.1 using Tapatalk

 

[Nebulaz]

So I never got the issue resolved. Disconnects continue. So I am going to just buy a new 86u. I remember reading somewhere that I should avoid versions made before 2018, I think. Anyone have a link or know exactly what I think I remember seeing

Thanks everyone for the help.



Sent from my Nokia 7.1 using Tapatalk

 

[L&LD]

@Nebulaz you are correct. Look for a version manufactured in China with HW version A1.1 or later and made during 2019 or later too.

No need to find that link (but if you must, a quick search is a few clicks away).

 

[Nebulaz]

A1.1 only seems to be associated with the AX88, are you certain that is for the 86u?

Sent from my Nokia 7.1 using Tapatalk