R7800 Constant-Short WAN Interruptions

  • ATTENTION! As of November 1, 2020, you are not able to reply to threads 6 months after the thread is opened if there are more than 500 posts in the thread.
    Threads will not be locked, so posts may still be edited by their authors.
    Just start a new thread on the topic to post if you get an error message when trying to reply to a thread.

Warlord1981

Occasional Visitor
Hello everyone,

I own a R7800 and until recently I had my ISP's modem/router Arris TG2492S in RoutedWithNAT mode and everything was ok. (Yes I was behind double NAT)

Ever since I configured the Arris in bridged mode I have been constantly having short (5 sec max) Internet Interruptions on my R7800 approximately every ~1 hour. During that time, sometimes the WAN-Internet Led on R7800 turns red and then goes back to white after a few seconds, other times it just stays white (since the whole interruption is not long).

This could reflect on disconnections from online game servers, or 5-10 sec freezes on my IPTV.

I checked my R7800 Logs from Netgear's Administration menu and here is a part:

[Internet connected] IP address: my-public-ip, Sunday, February 21, 2021 19:07:35
[Internet connected] IP address: my-public-ip, Sunday, February 21, 2021 18:14:14
[Internet connected] IP address: my-public-ip, Sunday, February 21, 2021 17:20:54
[Internet connected] IP address: my-public-ip, Sunday, February 21, 2021 16:27:34
[Internet connected] IP address: my-public-ip, Sunday, February 21, 2021 15:34:14
[Internet connected] IP address: my-public-ip, Sunday, February 21, 2021 14:40:54

It seems that Internet on R7800 is cut and reconnected every 53mins and 20 secs.

My R7800 has the latest Voxel fw & Kamoj-addon installed.

My ISP provides me with dynamic IP. Not static. Though I think it has been quite some time that I have the same. (even though I have restarted Arris modem/router several times the last few days)

Can you please advise? I read on official Netgear's forum that some modems (specifically the ones having Intel Puma chipsets) have issues when in bridged mode, where they need to only connect to the device which was initially connected to configure the bridge mode, otherwise every one hour they will look for this device, a process that can cause a short interruption. However I am sure that when i did the bridge config i only had my R7800 connected. (and my PC on R7800 to access the Arris' GUI and make the config).

I have the following config on Arris, based on my ISP's bridged mode instructions:


1.PNG


2.PNG

3.PNG
 
Last edited:

HELLO_wORLD

Very Senior Member
The key is to find if the disconnection comes from the modem or the router WAN interface...

Several ideas for that:
  • run a ping (ping 8.8.8.8 for ex.) from the router when it is about to disconnect, and see if latency changes or you have lost packets during the disconnected period. If not, then the R7800 never loses connection and it is its LAN side that is disconnected a few seconds. However, since your WAN LED changes color, I suspect that the ping will show the disconnection.
  • put directly a device (PC) behind the modem (left in bridge mode) instead of the router, and if it disconnects the same way, then it definitely comes from the modem.
If it is the bridge mode that is defect on the modem, just remove the bridge mode, disable modem’s WiFi, put the router in the modem’s DMZ, and removes as many options - rules as possible from the modem (firewall, nat, UPnP...), and you will control these points on your router.

It would be interesting to see how your R7800 is set up too... like advanced -> setup -> internet setup, WAN setup, LAN setup...
 

Warlord1981

Occasional Visitor
The key is to find if the disconnection comes from the modem or the router WAN interface...

Several ideas for that:
  • run a ping (ping 8.8.8.8 for ex.) from the router when it is about to disconnect, and see if latency changes or you have lost packets during the disconnected period. If not, then the R7800 never loses connection and it is its LAN side that is disconnected a few seconds. However, since your WAN LED changes color, I suspect that the ping will show the disconnection.
  • put directly a device (PC) behind the modem (left in bridge mode) instead of the router, and if it disconnects the same way, then it definitely comes from the modem.
If it is the bridge mode that is defect on the modem, just remove the bridge mode, disable modem’s WiFi, put the router in the modem’s DMZ, and removes as many options - rules as possible from the modem (firewall, nat, UPnP...), and you will control these points on your router.

It would be interesting to see how your R7800 is set up too... like advanced -> setup -> internet setup, WAN setup, LAN setup...
Thank you for your answer!

Will do 2 steps. In the meanwhile please find below requested info regarding R7800 Setup:

internet.PNG


wan.PNG



lan.PNG
 

kamoj

Very Senior Member
Interesting...
I made a check on my system with this little script:
(SNBforums doesn't allow me to use the word "substr", so if you want to run it in your router,
replace "sxubstr" by "substr" in the code below.)
Bash:
awk '/Internet connected/ {print $0" "sxubstr($NF,1,2)*3600+sxubstr($NF,4,2)*60+sxubstr($NF,7,2)}' </var/log/messages | while IFS= read -r S; do 
current="$(echo "$S"|awk '{print $NF}')"
[ -n "$prev" ] && delta="$(echo "$prev $current"|awk '{print ($NF-$1)}')"
prev="$current"
echo "$S delta=$delta" | sed -r 's/(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)/xxx.xxx.xxx.xxx/g'
done
and got this result:
Code:
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 01:04:20 3860 delta=
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 03:41:51 13311 delta=9451
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 06:19:21 22761 delta=9450
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 08:56:50 32210 delta=9449
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 11:34:20 41660 delta=9450
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 14:11:50 51110 delta=9450
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 16:49:21 60561 delta=9451
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 19:26:50 70010 delta=9449
So 9450 seconds (2h 37 min 30 sec) between "Internet connected".
I have no real disconnect however, iptv, vpn, ssh etc are all working ok, what I know...
 
Last edited:

Warlord1981

Occasional Visitor
Interesting...
I made a check on my system with this little script:
(SNBforums doesn't allow me to use the word "substr", so if you want to run it in your router,
replace "sxubstr" by "substr" in the code below.)
Bash:
awk '/Internet connected/ {print $0" "sxubstr($NF,1,2)*3600+sxubstr($NF,4,2)*60+sxubstr($NF,7,2)}' </var/log/messages | while IFS= read -r S; docurrent="$(echo "$S"|awk '{print $NF}')"
[ -n "$prev" ] && delta="$(echo "$prev $current"|awk '{print ($NF-$1)}')"
prev="$current"
echo "$S delta=$delta" | sed -r 's/(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)/xxx.xxx.xxx.xxx/g'
done
and got this result:
Code:
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 01:04:20 3860 delta=
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 03:41:51 13311 delta=9451
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 06:19:21 22761 delta=9450
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 08:56:50 32210 delta=9449
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 11:34:20 41660 delta=9450
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 14:11:50 51110 delta=9450
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 16:49:21 60561 delta=9451
[Internet connected] IP address: xxx.xxx.xxx.xxx, Monday, February 22, 2021 19:26:50 70010 delta=9449
So 9450 seconds (2h 37 min 30 sec) between "Internet connected".
I have no real disconnect however, iptv, vpn, ssh etc are all working ok, what I know...

Hmm so what does that mean? Your ISP checks every 2h 37 min 30 sec to re-provision new IP ? Do you have static or dynamic IP from your ISP ?

And what settings do you have on Internet Setup of Netgear ? Do we need to manually input the IP our modem reports when it is bridged or we leave it to "Get dynamically from ISP" ?
 
Last edited:

HELLO_wORLD

Very Senior Member
Hmm so what does that mean? Your ISP checks every 2h 37 min 30 sec to re-provision new IP ? Do you have static or dynamic IP from your ISP ?

And what settings do you have on Internet Setup of Netgear ? Do we need to manually input the IP our modem reports when it is bridged or we leave it to "Get dynamically from ISP" ?
Not an expert with bridge, but I believe the modem is delegating its IP attribution to the router. It is then up to R7800 to get it (DHCP).

I think it fails to renew the lease on time, hence the mini cut.

One more test idea:
About 30 minutes after a connection, use:
ps | grep -F udhcpc
It should return a line containing something like this on its 5th column:
udhcpc -b -i brwan .....

Copy the 5th column of this line (starting with udhcpc -b etc...)

Then do:
killall udhcpc
[paste what you copied and enter] udhcpc -b -i brwan .....

Then just wait and see if you have the disconnection as usual or almost one hour after you ran these commands.
 

Warlord1981

Occasional Visitor
Not an expert with bridge, but I believe the modem is delegating its IP attribution to the router. It is then up to R7800 to get it (DHCP).

I think it fails to renew the lease on time, hence the mini cut.

One more test idea:
About 30 minutes after a connection, use:
ps | grep -F udhcpc
It should return a line containing something like this on its 5th column:
udhcpc -b -i brwan .....

Copy the 5th column of this line (starting with udhcpc -b etc...)

Then do:
killall udhcpc
[paste what you copied and enter] udhcpc -b -i brwan .....

Then just wait and see if you have the disconnection as usual or almost one hour after you ran these commands.
Dear @HELLO_wORLD thank you as always for your help! Will try that for sure. For now I have enabled today static IP on my Internet line from my ISP and inputed manually my-static-IP/Subnet/Gateway on Internet Setup>Internet IP Address of R7800 and have no disconnections until now. Almost 3 hours have passed. Is this an acceptable config? Or we should always leave it dynamically (that's when i get the disconnections)?

What I have since last night though is some sort of congestion/packet-loss in all traffic. People can hear me with cuts on Discord, Speedtest is freezing randomly for a couple of seconds, fifa21 freezes for a couple of seconds during an online session. I am trying to remember what I changed, but really there is no change i made on R7800 config since Sunday. When I connect directly to modem I have no issues. It's driving me nuts. I am even considering reseting everything in R7800 and starting over. I ran multiple tests in packetlosstest.com and i always get 147/149 received with 0% Packet Loss. However http://www.dslreports.com/speedtest is very unstable. Bufferbloat goes from A+ to A-B-C
 
Last edited:

HELLO_wORLD

Very Senior Member
Dear @HELLO_wORLD thank you as always for your help! Will try that for sure. For now I have enabled today static IP on my Internet line from my ISP and inputed manually my-static-IP/Subnet/Gateway on Internet Setup>Internet IP Address of R7800 and have no disconnections until now. Almost 3 hours have passed. Is this an acceptable config? Or we should always leave it dynamically (that's when i get the disconnections)?

What I have since last night though is some sort of congestion/packet-loss in all traffic. People can hear me with cuts on Discord, Speedtest is freezing randomly for a couple of seconds, fifa21 freezes for a couple of seconds during an online session. I am trying to remember what I changed, but really there is no change i made on R7800 config since Sunday. When I connect directly to modem I have no issues. It's driving me nuts. I am even considering reseting everything in R7800 and starting over. I ran multiple tests in packetlosstest.com and i always get 147/149 received with 0% Packet Loss. However http://www.dslreports.com/speedtest is very unstable. Bufferbloat goes from A+ to A-B-C
Ok, no need to try anymore ;)
That confirms what I suspected, and using static IP gets rid of DHCP (WAN side) and its interruptions. So that is solved.

For the other problem, it appears to be something in the R7800 (unless it is related to the bridge mode on the modem side for some weird reason...)
On the R7800, you can use the command top to check if the load is fine and nothing is using the CPU or MEM abnormally.

You can also change the congestion control algorithm (I personally use yeah):
Code:
echo "yeah westwood reno" >/proc/sys/net/ipv4/tcp_allowed_congestion_control
echo yeah >/proc/sys/net/ipv4/tcp_congestion_control
If it is better, there is away to make it permanent (the code I gave you is only lasting until next reboot).

To finish, when you put the R7800 behind the modem, modem in NAT mode, not bridge mode, and keep same settings on your R7800 (except WAN ip of course that needs to be in subnet of modem LAN), do you experience the same troubles?
If not, it would point towards the modem bridge mode.
 

Warlord1981

Occasional Visitor
Ok, no need to try anymore ;)
That confirms what I suspected, and using static IP gets rid of DHCP (WAN side) and its interruptions. So that is solved.

For the other problem, it appears to be something in the R7800 (unless it is related to the bridge mode on the modem side for some weird reason...)
On the R7800, you can use the command top to check if the load is fine and nothing is using the CPU or MEM abnormally.

You can also change the congestion control algorithm (I personally use yeah):
Code:
echo "yeah westwood reno" >/proc/sys/net/ipv4/tcp_allowed_congestion_control
echo yeah >/proc/sys/net/ipv4/tcp_congestion_control
If it is better, there is away to make it permanent (the code I gave you is only lasting until next reboot).

To finish, when you put the R7800 behind the modem, modem in NAT mode, not bridge mode, and keep same settings on your R7800 (except WAN ip of course that needs to be in subnet of modem LAN), do you experience the same troubles?
If not, it would point towards the modem bridge mode.
It has been more than 2 weeks since I am running the modem in bridge mode, but never experienced this. Only the short hourly disconnections. This problem started last night (before setting up static-ip).

Have already used top as well as netdata from Kamoj addon and there are no spikes in cpu/mem when the issue occurs.

I will try yeah congestion control from Kamoj addon (it has a bunch of congestion where you can select) and report back. Thanks :)

By the way the command above (ps | grep -F udhcpc) what does it do?
 

kamoj

Very Senior Member
I would say that this change of congestion control does not even survive until next reboot.
Netgear/DNI has some code overwriting these settings with those in sysctl.cfg now and then,
so you better change the config file - that survives until new firmware.
(No, I don't know what triggers the overwrite. It's the same as with iptables -
suddenly all own values are lost!)
You can also change the congestion control algorithm (I personally use yeah):
Code:
echo "yeah westwood reno" >/proc/sys/net/ipv4/tcp_allowed_congestion_control
echo yeah >/proc/sys/net/ipv4/tcp_congestion_control
If it is better, there is away to make it permanent (the code I gave you is only lasting until next reboot).
..
 

HELLO_wORLD

Very Senior Member
By the way the command above (ps | grep -F udhcpc) what does it do?
It just shows udhcpc running process if any, and how it was called (command line).

udhcpc is the DHCP client used by Netgear in R7800 and other models.
 

HELLO_wORLD

Very Senior Member
I would say that this change of congestion control does not even survive until next reboot.
Netgear/DNI has some code overwriting these settings with those in sysctl.cfg now and then,
so you better change the config file - that survives until new firmware.
(No, I don't know what triggers the overwrite. It's the same as with iptables -
suddenly all own values are lost!)
Exactly.
I am using the sysctl.cfg way myself.
I did suggest the proc way to @Warlord1981 because it is temporary (for him to test) and not permanently change things in his config. A test first and make it permanent if successful.

But with you addon, it is way easier for him :):cool:
 

Warlord1981

Occasional Visitor
@HELLO_wORLD Unfortunately changing congestion control does not help. I even uninstalled completely Kamoj-addon and didn't help either. Now only option to factory reset R7800?
 

HELLO_wORLD

Very Senior Member
@HELLO_wORLD Unfortunately changing congestion control does not help. I even uninstalled completely Kamoj-addon and didn't help either. Now only option to factory reset R7800?
Well, comes a point when a factory reset is faster than trying many many things...
Before though, are you sure you don’t have QoS, traffic meter or something like that turned on? They should be off.

Just write down the settings that are important (WAN, LAN, etc...) before doing a reset ; better check from advanced tab than basic.

Also, I suggest doing aegis unset before the reset, then put back your settings, then aegis up.
 

Warlord1981

Occasional Visitor
Well, comes a point when a factory reset is faster than trying many many things...
Before though, are you sure you don’t have QoS, traffic meter or something like that turned on? They should be off.

Just write down the settings that are important (WAN, LAN, etc...) before doing a reset ; better check from advanced tab than basic.

Also, I suggest doing aegis unset before the reset, then put back your settings, then aegis up.
I Just reset modem to RoutedWithNAT and i have no such behaviour. Which means there is something wrong with the bridge of that particular router-modem combo?
 

HELLO_wORLD

Very Senior Member
I Just reset modem to RoutedWithNAT and i have no such behaviour. Which means there is something wrong with the bridge of that particular router-modem combo?
Looks like it, BUT you said you had it in bridge and it was not doing that for a while, right?

Now, a simple way to have this working is without bridge and R7800 in modem’s DMZ (static IP). If all is working fine for you, no real need to dig further.
 

Warlord1981

Occasional Visitor
Looks like it, BUT you said you had it in bridge and it was not doing that for a while, right?

Now, a simple way to have this working is without bridge and R7800 in modem’s DMZ (static IP). If all is working fine for you, no real need to dig further.

I reset again the modem and re-created the bridge and seems for now the problem is gone.. Need to install kamoj-addon back and restart R7800.
 

Similar threads

Latest threads

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top