Internet Dropping PPPoE - Need to reboot router to get back online

[Icelvlan]

Hello,

I'm on the latest firmware for both my RT-AX88 and RT-AC3100 and I keep having issues with 5GHZ connection. When I switch to 2.4 GHZ everything is back to normal. I have 2 wireless routers in the house and 1 set as a mesh node.

AX88 - 388.1
AC3100 - 386.9

Issue

Mostly slowness like I am having issues with my internet connection. Sometimes things load sometimes they don't or take forever. Sometimes it works for a decent amount of time then it starts having issues.

Is there somewhere I can look to help diagnose the issue? I don't remember the issue starting after a particular update unfortunately.

 

[L&LD]

'Latest firmware' doesn't mean anything. Please be specific.

There is no RT-AX3100 router. Is the RT-AC88 the RT-AC88U?

What other features, scripts, or options do you have enabled/running on the main router? Which one is the main router?

How are the two routers connected?

How is the WiFi setup?

When was the last time you did a full reset to factory defaults (at least on the main router)?

How large is the area you're covering? Will one router be sufficient for that? (The RT-AC3100 had the best range of any router I've ever used, if that is the model you actually have).

What are your ISP speeds? Are they symmetrical, up/down?

 

[Icelvlan]

Apologies updated.

Routers connected by copper and copper has been tested 1G
Wireless is set up as 2 separate SSID (2.4 and 5) and 1 node is AImesh
1 at each end of the house, never had covered issues for a number of years. It also happens when I'm 10' from a router in open space.
Internet connection is 900M up 1G down
No additional scripts or features are enabled on the router.
Last time I did a factory default on primary router was probably before I moved over to 388.1. Maybe 1-2 years ago

 

[Tech9]

I moved over to 388.1. Maybe 1-2 years ago

You're good! RMerlin managed to release his 388.1 version only in Dec 2022.

 

[Icelvlan]

No I mean I last factory reset 1-2 years ago.

 

[Tech9]

Do it now for both routers + manual configuration after. You don't need Asuswrt-Merlin on the Node - it's an Access Point, no routing. For more stable Wi-Fi run the newer stock Asuswrt on your AX88U. It has fixes not yet implemented in Asuswrt-Merlin.

 

[Icelvlan]

So what I've tried so far.

Did a factory reset and reconfigured manually. For the time, I've left the nodes on Merlin as I just want to try one thing at a time. Seems that the 5Ghz issue is resolved.

Now anywhere between 4 hours - 2 days the internet randomly drops. The only thing that restores the connection is rebooting the router. Turning off/on WAN from webpage does nothing.

I'm using PPPoE and here are some logs that are showing up.

Apr 9 14:22:07 pppd[1428]: Timeout waiting for PADO packets
Apr 9 14:22:22 rc_service: httpds 1185:notify_rc restart_wan_if 0
Apr 9 14:22:22 pppd[1428]: Unable to complete PPPoE Discover
Apr 9 14:22:22 pppd[1428]: Hangup (SIGHUP)
Apr 9 14:22:22 pppd[1428]: Terminating on signal 15
Apr 9 14:22:22 pppd[1428]: Exit.
Apr 9 14:22:22 lldpd[1351]: removal request for address of 169.254.****%26, but no knowledge of it
Apr 9 14:22:22 avahi-daemon[1179]: Withdrawing address record for 169.254.*** on br101.
Apr 9 14:22:22 avahi-daemon[1179]: Leaving mDNS multicast group on interface br101.IPv4 with address 169.254.***.
Apr 9 14:22:22 avahi-daemon[1179]: Interface br101.IPv4 no longer relevant for mDNS.
Apr 9 14:22:37 dnsmasq[32159]: warning: interface br2 does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: warning: interface br1 does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: warning: interface pptp* does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: warning: interface ppp0 does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: asynchronous logging enabled, queue limit is 5 messages
Apr 9 14:23:08 kernel: CFG80211-ERROR) wl_cfg80211_change_station : WLC_SCB_AUTHORIZE sta_flags_mask not set
Apr 9 14:23:10 pppd[32111]: Timeout waiting for PADO packet
Apr 9 14:23:12 services: apply rules error(18891)
Apr 9 14:23:32 services: apply rules error(18891)
Apr 9 14:23:36 services: apply rules error(18891)
Apr 9 14:23:41 services: apply rules error(18891)

 

[drinkingbird]

So what I've tried so far.

Did a factory reset and reconfigured manually. For the time, I've left the nodes on Merlin as I just want to try one thing at a time. Seems that the 5Ghz issue is resolved.

Now anywhere between 4 hours - 2 days the internet randomly drops. The only thing that restores the connection is rebooting the router. Turning off/on WAN from webpage does nothing.

I'm using PPPoE and here are some logs that are showing up.

Apr 9 14:22:07 pppd[1428]: Timeout waiting for PADO packets
Apr 9 14:22:22 rc_service: httpds 1185:notify_rc restart_wan_if 0
Apr 9 14:22:22 pppd[1428]: Unable to complete PPPoE Discover
Apr 9 14:22:22 pppd[1428]: Hangup (SIGHUP)
Apr 9 14:22:22 pppd[1428]: Terminating on signal 15
Apr 9 14:22:22 pppd[1428]: Exit.
Apr 9 14:22:22 lldpd[1351]: removal request for address of 169.254.****%26, but no knowledge of it
Apr 9 14:22:22 avahi-daemon[1179]: Withdrawing address record for 169.254.*** on br101.
Apr 9 14:22:22 avahi-daemon[1179]: Leaving mDNS multicast group on interface br101.IPv4 with address 169.254.***.
Apr 9 14:22:22 avahi-daemon[1179]: Interface br101.IPv4 no longer relevant for mDNS.
Apr 9 14:22:37 dnsmasq[32159]: warning: interface br2 does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: warning: interface br1 does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: warning: interface pptp* does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: warning: interface ppp0 does not currently exist
Apr 9 14:22:37 dnsmasq[32159]: asynchronous logging enabled, queue limit is 5 messages
Apr 9 14:23:08 kernel: CFG80211-ERROR) wl_cfg80211_change_station : WLC_SCB_AUTHORIZE sta_flags_mask not set
Apr 9 14:23:10 pppd[32111]: Timeout waiting for PADO packet
Apr 9 14:23:12 services: apply rules error(18891)
Apr 9 14:23:32 services: apply rules error(18891)
Apr 9 14:23:36 services: apply rules error(18891)
Apr 9 14:23:41 services: apply rules error(18891)

Is your WAN MTU set to 1492 (required for PPPoE)? Maybe you forgot to set that again after resetting?

 

[Icelvlan]

I have it at 1500 right now, I thought some ISPs supported that? It work fine for a long time, I’m getting great speeds

 

[sfx2000]

common issue - it's a PPP dangling leg - the client side (your router) has disconnected due to a timeout, but the server side (your ISP) still has the session up...

can't have two sessions at the same time...

Call Customer Care at your ISP - they'e the ones that can attempt to fix it - anything from the client side is screaming at the wind - feels good perhaps, but doesn't fix the problem at hand.

 

[drinkingbird]

I have it at 1500 right now, I thought some ISPs supported that? It work fine for a long time, I’m getting great speeds

PPPoE almost always needs 1492. It has an extra 8 bytes overhead which is deducted from the 1500. You can test by doing:

ping -l 1472 -f google.com

If it fails with "fragment needed) your MTU is too high.

if ping -l 1464 -f google.com works, then your MTU should be 1492.

note the -l is a lowercase "L"

In fact some will argue that 1452 is the best size for PPPoE as it uses ATM technology and 1452 is a size that requires no packet padding (thus reducing the overhead a bit and making the connection more efficient). But I would start with 1492 (assuming the tests above show that is what is needed). It could be interfering with your DHCP or other PPPoE control messages causing the connection to hang. Considering the only change you made was firmware, it seems likely that maybe you had it set right before and just didn't remember this time.

 

[Icelvlan]

I just switched it back to 1492, I'll see how it is for the next few days.

 

[Icelvlan]

An update, I've noticed something right before I have issues. Wifi drops almost as if it is reset before the WAN disconnect and again, when it disconnects nothing outside of a router reboot bring the WAN back online. I'm going to try updating to 388.2 and see if it is better.

 

[sfx2000]

An update, I've been nothing something right before I have issues.

Call Customer Care at your ISP - they'e the ones that can attempt to fix it - anything from the client side is screaming at the wind - feels good perhaps, but doesn't fix the problem at hand.

 

[Icelvlan]

Call Customer Care at your ISP - they'e the ones that can attempt to fix it - anything from the client side is screaming at the wind - feels good perhaps, but doesn't fix the problem at hand.

What I'm curious about is why I cannot connect to wifi 5ghz and 2.4ghz as it is happening. I lose connection, then when I try to connect it says wrong password. A few minutes later I'm able to connect to wifi again.

 

[Icelvlan]

Also, what is the difference between the following for factory reset?

Rebooting router and holding reset button vs holding WPS button during reboot?

 

[sfx2000]

What I'm curious about is why I cannot connect to wifi 5ghz and 2.4ghz as it is happening. I lose connection, then when I try to connect it says wrong password. A few minutes later I'm able to connect to wifi again.

Because on your WAN side - you have a PPP dangling leg - the drop started, but PPP is stateful, so it needs the ack - if the line is already dropped, the ack will not be received...

Hence you have a stalled session on the other end, and until that session times out again and closes, you won't get in...

 

[drinkingbird]

Because on your WAN side - you have a PPP dangling leg - the drop started, but PPP is stateful, so it needs the ack - if the line is already dropped, the ack will not be received...

Hence you have a stalled session on the other end, and until that session times out again and closes, you won't get in...

WAN issues should not prevent you from connecting to WIFI, even if there is no internet access, it would not report back a bad wifi password.

Now I suspect their internet isn't actually going down, just wifi.

 

[drinkingbird]

Also, what is the difference between the following for factory reset?

Rebooting router and holding reset button vs holding WPS button during reboot?

You want do the WPS button method, that formats JFFS and fully restores back to factory. Personally I like to do it both before and after upgrading firmware (you'll need to configure just basic stuff to get back in and upgrade the firmware, then configure everything after the second reset).

Are you sure your internet is going down or is it just the wifi?

Are you using DFS channels for wifi? If so and it detects radar, it will disconnect 5ghz wifi for several minutes while it scans. Maybe your 2.4 has a different/wrong password configured so when 5Ghz goes into scan, 2.4 can't connect?

 

[Icelvlan]

My wifi is going down at the same time, the router stays up. As soon as I'm able to connect to wifi again, PPPoE is down and gives error. I don't understand how PPPoE would impact being connected to actual wifi of the router.

Could it be that if I didn't do WPS method, something broke? What is JFFS?