RT-AX86U-Pro, WAN disconnect every 36 hours

[Nick24]

Hi all,

First time posting, but been reading the forum for years now, but never had anything to contribute myself.

Long story short, I've been dealing with an issue with periodic disconnects of my WAN connection for over 6 months or more. I've gone through the complete cycle of diagnosis with my ISP and now I'm back thinking my equipment is at fault.

My setup:

1. Coax internet connection with my ISP (Videotron in Quebec, Canada).
2. ISP supplied modem is an XB7 made by Arris (https://www.videotron.com/en/support/internet/equipment/helix-fi-gateway-2)
3. Modem is configured in 'Bridge Mode', connected via a CAT7 3ft cable to the 1G WAN port of my Asus Router
4. Asus Router is a RT-AX86U Pro running latest Merlin build 3004.388.8_2

My symptom:

1. Every 36 hours (almost to the minute), I loose internet connection for about 15 seconds, 2 times in a row. So loose internet for 15 seconds it comes back for about 15 seconds, loose it again, then it's rock solid for the next 36 hours. When it happens in the middle of the night, I don't care or notice. My problem is when it happens during the day and I'm on a Teams call for work and I get cut off from the call until my network comes back.
2. Logs from latest event:
   

   Sep 11 05:52:49 kernel: eth1 (Int switch port: 1) (Logical Port: 1) (phyId: 2) Link DOWN.
   Sep 11 05:52:54 kernel: eth1 (Int switch port: 1) (Logical Port: 1) (phyId: 2) Link Up at 1000 mbps full duplex
   Sep 11 05:54:58 kernel: eth1 (Int switch port: 1) (Logical Port: 1) (phyId: 2) Link DOWN.
   Sep 11 05:55:06 kernel: eth1 (Int switch port: 1) (Logical Port: 1) (phyId: 2) Link Up at 1000 mbps full duplex
   Sep 11 05:55:09 kernel: eth1 (Int switch port: 1) (Logical Port: 1) (phyId: 2) Link DOWN.
   Sep 11 05:55:15 kernel: eth1 (Int switch port: 1) (Logical Port: 1) (phyId: 2) Link Up at 1000 mbps full duplex
   Sep 11 05:55:58 WAN_Connection: WAN(0) link up.
   Sep 11 05:55:58 rc_service: wanduck 1540:notify_rc restart_wan_if 0
   Sep 11 05:55:58 ovpn-server1[4409]: event_wait : Interrupted system call (fd=-1,code=4)
   Sep 11 05:55:58 ovpn-server1[4409]: Closing TUN/TAP interface
   Sep 11 05:55:58 ovpn-server1[4409]: /usr/sbin/ip addr del dev tun21 10.8.0.1/24
   Sep 11 05:55:58 ovpn-server1[4409]: ovpn-down 1 server tun21 1500 0 10.8.0.1 255.255.255.0 init
   Sep 11 05:55:58 ovpn-server1[4409]: PLUGIN_CLOSE: /usr/lib/openvpn-plugin-auth-pam.so
   Sep 11 05:55:58 ovpn-server1[4409]: PLUGIN AUTH-PAM: Error signaling background process to exit: Connection refused (errno=111)
   Sep 11 05:55:58 ovpn-server1[4409]: SIGTERM[hard,] received, process exiting
   Sep 11 05:55:58 dnsmasq[2334]: read /etc/hosts - 24 names
   Sep 11 05:55:58 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:55:58 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:55:58 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:55:58 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:55:58 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:55:58 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:56:00 dnsmasq[2334]: read /etc/hosts - 24 names
   Sep 11 05:56:00 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:56:00 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:56:00 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:56:00 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:56:00 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:56:00 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:56:01 dnsmasq[2334]: read /etc/hosts - 24 names
   Sep 11 05:56:01 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:56:01 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:56:01 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:56:01 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:56:01 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:56:01 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:56:01 wan: finish adding multi routes
   Sep 11 05:56:01 openvpn-routing: Applying all killswitches
   Sep 11 05:56:01 miniupnpd[4387]: shutting down MiniUPnPd
   Sep 11 05:56:06 miniupnpd[19428]: HTTP listening on port 34561
   Sep 11 05:56:06 miniupnpd[19428]: Listening for NAT-PMP/PCP traffic on port 5351
   Sep 11 05:56:06 ddns: IP address, server and hostname have not changed since the last update.
   Sep 11 05:56:06 rc_service: udhcpc_wan 19200:notify_rc start_vpnserver1
   Sep 11 05:56:06 dhcp_client: bound XXX.XXX.XXX.96/255.255.255.0 via XXX.XXX.XXX.1 for 3600 seconds.
   Sep 11 05:56:06 ovpn-server1[19449]: OpenVPN 2.6.12 arm-buildroot-linux-gnueabi [SSL (OpenSSL)] [LZO] [LZ4] [EPOLL] [MH/PKTINFO] [AEAD]
   Sep 11 05:56:06 ovpn-server1[19449]: library versions: OpenSSL 1.1.1w  11 Sep 2023, LZO 2.10
   Sep 11 05:56:06 ovpn-server1[19450]: NOTE: the current --script-security setting may allow this configuration to call user-defined scripts
   Sep 11 05:56:06 ovpn-server1[19450]: PLUGIN AUTH-PAM: initialization succeeded (fg)
   Sep 11 05:56:06 ovpn-server1[19450]: PLUGIN_INIT: POST /usr/lib/openvpn-plugin-auth-pam.so '[/usr/lib/openvpn-plugin-auth-pam.so] [openvpn]' intercepted=PLUGIN_AUTH_USER_PASS_VERIFY
   Sep 11 05:56:06 ovpn-server1[19450]: Diffie-Hellman initialized with 2048 bit key
   Sep 11 05:56:06 ovpn-server1[19450]: TUN/TAP device tun21 opened
   Sep 11 05:56:06 ovpn-server1[19450]: TUN/TAP TX queue length set to 1000
   Sep 11 05:56:06 ovpn-server1[19450]: /usr/sbin/ip link set dev tun21 up mtu 1500
   Sep 11 05:56:06 vpnserver1[19451]: PLUGIN AUTH-PAM: BACKGROUND: initialization succeeded
   Sep 11 05:56:06 ovpn-server1[19450]: /usr/sbin/ip link set dev tun21 up
   Sep 11 05:56:06 ovpn-server1[19450]: /usr/sbin/ip addr add dev tun21 10.8.0.1/24
   Sep 11 05:56:06 ovpn-server1[19450]: ovpn-up 1 server tun21 1500 0 10.8.0.1 255.255.255.0 init
   Sep 11 05:56:06 ovpn-server1[19450]: Socket Buffers: R=[524288->524288] S=[524288->524288]
   Sep 11 05:56:06 ovpn-server1[19450]: UDPv4 link local (bound): [AF_INET][undef]:2031
   Sep 11 05:56:06 ovpn-server1[19450]: UDPv4 link remote: [AF_UNSPEC]
   Sep 11 05:56:06 ovpn-server1[19450]: MULTI: multi_init called, r=256 v=256
   Sep 11 05:56:06 ovpn-server1[19450]: IFCONFIG POOL IPv4: base=10.8.0.2 size=253
   Sep 11 05:56:06 ovpn-server1[19450]: Initialization Sequence Completed
   Sep 11 05:57:01 WAN_Connection: WAN was restored.
   Sep 11 05:57:01 dnsmasq[2334]: read /etc/hosts - 24 names
   Sep 11 05:57:01 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:57:01 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:57:01 dnsmasq[2334]: using only locally-known addresses for Home
   Sep 11 05:57:01 dnsmasq[2334]: using nameserver 1.1.1.1#53
   Sep 11 05:57:01 dnsmasq[2334]: using nameserver 149.112.121.10#53
   Sep 11 05:57:01 dnsmasq[2334]: using only locally-known addresses for Home

Troubleshooting steps done so far:

1. Contacted my ISP. They have sent technicians out 2 times. First time they replaced the entire outdoor Coax line from the pole to my house. Next time I wanted technician to replace my modem but they determined nothing was wrong and that my equipment was at fault. They say the logs from the modem show a port disconnect and that means it must be something wrong with my ethernet cable or my router.
2. I replaced my ethernet cable from the modem to the router
3. I completely replaced the Coax cable from the outside of my house to my modem. It's a brand new, straight cable RG6 cable from outside to my modem. No splitters. But even the cable company says my signal quality is fine.
4. I have factory reset my modem and router
5. I have changed the DHCP setting in Asus to Normal/Aggressive/Continuous. All the same.
6. Used the 'dual wan' function and changed my WAN port to the 2.5G port
7. Used the 'dual wan' function and changed my WAN port to the LAN port 1
8. Changed from port 1 on my modem to the 2.5G port (Port 4) on the XB7 modem
9. Changed to port 2 on my modem

My current thinking:

1. I think it might have something to do with how the Asus router is handling the DHCP renewal request from my cable modem. My cable modem reports (when logged into the webpage of the modem) that is has a DHCP lease time of 3 days = 72 hours. According to DHCP specs, "The client typically sends the DHCP Request when half (50%) of the lease time has elapsed". 50% of 72 hours is 36 hours. This can't be a coincidence that at the 36 hour mark is when my Asus logs show that the port goes down.
2. My next idea is going to be to try: https://github.com/lynxthecat/maintain-wan-lease
3. Or, to use this idea, https://www.snbforums.com/threads/wan-connect-and-disconnect-using-script.11232/page-2#post-185023 and create a script that pre-emptively reboots the WAN connection at the 24 hour mark sometime in the middle of the night when it doesn't bother me.
4. Or, put my ISP modem back in Normal mode, disable everything I can (which they make limited), and put the Asus router in a DMZ. Still not sure how this will work with my VPN server running on Asus along with DDNS for that server.

At this point I'm pretty much out of ideas. I've searched this forum and I've seen similar posts that come close to my issue, but a lot of them mention the common DHCP not functioning error in the logs. I'm wondering if in my version of firmware that I'm running if this message has been suppressed and that's why I'm not seeing it, or if I have a different issue.

Current WAN page settings:

 

[Nick24]

Wanted to add some more interesting information I found. There seems to be evidence to suggest this might be a common problem with modems running the Intel's Puma 7 chipset. I found this interesting post here: https://forum.fizz.ca/en/discussion/2641803/ongoing-bridge-mode-wan-problem

So my idea is that if my current provider has a garbage modem, then I need to find another provider. In my area, only Coax internet gives me speeds over 120 which I need. So I took a look at the sub brand to Videotron which is Fizz. And I searched for the model of their modem 'CODA 4680 from Hiltron' and Asus wan disconnect, and I found the thread above. So the common element here is both the Hiltron and XB7 modems are running on Intel Puma chipsets, as well as being the same Coax ISP provider (Fizz and Videotron share the same lines).

Given that I don't have much choice but to stick to this provider and their choice of modem, what can I do?

 

[Analog-1]

From your logs your modem or ISP is what's dropping not the router. Also the Puma 7 chipset was known for packet loss and high ping rates under load not disconnects. Are you able to post your upstream and downstream signal levels and SNR's ?

 

[Nick24]

From your logs your modem or ISP is what's dropping not the router. Also the Puma 7 chipset was known for packet loss and high ping rates under load not disconnects. Are you able to post your upstream and downstream signal levels and SNR's ?

This is what I have access to from the modem GUI:

Better quality link for downstream:

 

[ColinTaylor]

I would agree with @Analog-1. This looks like the modem rebooting. Notice this line:

Sep 11 05:56:06 dhcp_client: bound XXX.XXX.XXX.96/255.255.255.0 via XXX.XXX.XXX.1 for 3600 seconds.

The lease time is only 60 minutes. That's usually because a) "XXX.XXX.XXX.96" is a temporary private IP address used during reboots, or b) there's some engineering work being carried out in your area.

Does your modem not have a system log?

 

[Analog-1]

Your signal levels at the time you took them look good. As ColinTaylor said a very short lease time on a docsis connection means work being done in your area.

 

[Nick24]

I would agree with @Analog-1. This looks like the modem rebooting. Notice this line:

Sep 11 05:56:06 dhcp_client: bound XXX.XXX.XXX.96/255.255.255.0 via XXX.XXX.XXX.1 for 3600 seconds.

The lease time is only 60 minutes. That's usually because a) "XXX.XXX.XXX.96" is a temporary private IP address used during reboots, or b) there's some engineering work being carried out in your area.

Does your modem not have a system log?

The modem doesn't reboot. The uptime on the modem page is how I see it. The logs don't really work on the modem page.

Also, the modem takes at least 2 minutes to boot and my internet doesn't go down for that long.

 

[Nick24]

This is the current status page from the modem and you can see the uptime is longer than the last time I lost internet which was yesterday:

The DHCP lease time is coming up on the 36 hour mark and I'm patiently waiting to catch it live in 38 min.

 

[Analog-1]

The DHCP lease time is coming up on the 36 hour mark and I'm patiently waiting to catch it live in 38 min.

DHCP leases have a half life. That is when things work properly a 36 hour lease time would renew at the 18 hour mark.

 

[ColinTaylor]

The DHCP lease time is coming up on the 36 hour mark and I'm patiently waiting to catch it live in 38 min.

If you log into the Asus router, go to the Network Map page and click on the Internet "globe" icon; what is the DHCP lease information shown in the Internet status tab on the righthand side?

 

[Nick24]

DHCP leases have a half life. That is when things work properly a 36 hour lease time would renew at the 18 hour mark.

Sorry I should have clarified. The DHCP lease time in the modem is 72 hours, and I experience my WAN port loss at exactly the 36 hour mark. Which is the halfway mark.

 

[Nick24]

If you log into the Asus router, go to the Network Map page and click on the Internet "globe" icon; what is the DHCP lease information shown in the Internet status tab on the righthand side?

 

[Nick24]

And I was right. The event just occurred:

I had a ping going at the same time to google:

In the modem, the lease time has reset:

The log mechanism in the modem is broken and doesn't generate anything:

 

[Justinh]

Can you supply your own modem?

 

[Nick24]

Can you supply your own modem?

No

And they refuse to replace it too since they say nothing is wrong. In the logs when they view it on their end they say it's the port disconnecting. So they tell me it's either a bad cable or my equipment is bad. However if you search online, I'm not the only one with similar issues.

 

[Analog-1]

Do something to the modem. (Over volt ) pop it then call and say your service is out. Now they will replace. If your issue continues then it's a service issue.

 

[Tech9]

Or, put my ISP modem back in Normal mode, disable everything I can (which they make limited), and put the Asus router in a DMZ. Still not sure how this will work with my VPN server running on Asus along with DDNS for that server.

This may fix the issue. Your VPN server will be fine.

And they refuse to replace it too since they say nothing is wrong.

Most likely there is nothing wrong with the ISP Gateway.

 

[Magister]

I have Videotron too, with the 400/50 plan. What's your version of the XB7 ? I noticed mine does not have the same wording in the channel bonding

Mine is :

HW Version: 2.0
Vendor: Technicolor
BOOT Version: S1TC-3.63.20.104
Core Version: 1.0
Model: CGM4331COM
Product Type: XB7
Flash Part: 4096 MB
Download Version: Prod_22.2_d31 & Prod_22.2

 

[Magister]

Also yes, try to put the modem in normal mode (no bridge), disable wifi on the xb7 ; set as DMZ or forward the OpenVPN port to your asus router (you need their horrible helix application to do this).

Doing this, your XB7 will handle the 3 days DHCP from the ISP, then your asus will have a by default 2 days DHCP with a 10.0.0.x address.

BTW this is how I setup mine, my helix TVbox is wired connected to the XB7, as well as the asus, I set a static 10.0.0.128 to my asus router, and from their application I forward openvpn/wireguard port to 10.0.0.128. Yes it doubled NAT, no I don't care I don't game and it has 0 impact.

 

[Nick24]

I have Videotron too, with the 400/50 plan. What's your version of the XB7 ? I noticed mine does not have the same wording in the channel bonding

Mine is :

HW Version: 2.0
Vendor: Technicolor
BOOT Version: S1TC-3.63.20.104
Core Version: 1.0
Model: CGM4331COM
Product Type: XB7
Flash Part: 4096 MB
Download Version: Prod_22.2_d31 & Prod_22.2