Duma R2 Daily Internet Outage

[rymcc86]

Hello, 

On a daily basis I'm having to reboot my R2, it's typically around the same time. The router appears to lock up, the WLAN and LAN loses internet connectivity and it takes a hard reset of the router to restore service.

I've also attempted to reach the router when the issue occurs using powershell to release/renew the IP configuration - the router is unresponsive.
 

My current setup is like below;
Ethernet LAN
[PC/XBOX/UC DEVICES]-----[Cisco PoE Switch]-----[NETDUMA R2]-----[Virgin SuperHub 3]

Ethernet LAN
[iPhone/iPad/Laptop/UC Devices]-----[NETDUMA R2]-----[Virgin SuperHub 3]

It feels a lot like a memory leak on the router but I've not noticed a gradually increasing CPU and RAM nor any odd or repeating log entries which would accompany such an issue. 


Can you advise?

[Netduma Fraser]

Interesting that it seems to be happening around the same time each day. With the Hub have you set it to modem mode or router mode? I have the same hub and I find it works better with the hub in router mode - the modem mode in my opinion is a bit finicky

[rymcc86]

That's interesting Fraser, I'll try the Hub in Router mode and disable the WiFi. I've always known it as best practice not to run two routers adjacent to each other, either run one or the other.

I've also noticed a ton of these warnings in the logs;
kern.warn kernel: [ 116.689456] HTB: quantum of class 10008 is big. Consider r2q change

I noticed an article when searching for answers on the Netgear forum regarding their devices running Dumaos and it was suggested as an attempt to compromise the device, is that possible? I assume the default switching position of the device when it stops responding is all ports disabled? 

https://community.netgear.com/t5/Hardware-VPN-Firewalls-and/HTB-quantum-of-class-11024-is-big-Losing-internet/td-p/1538066

 

[Netduma Fraser]

Make sure you put the R2 WAN IP found on the System Information page into the DMZ on the hub and you won't have any issues.

That's absolutely fine, no need to worry about that. The log is primarily aimed at developers so shows normal background processes that can appear like issues to a user, you can safely ignore it.

[rymcc86]

Hey Fraser, 

I changed into router mode, configured the virgin hub the two have been working fine together. It actually seems to be performing better with the virgin hub acting as the WAN router. 

Then today, I got another freeze. So we've improved, we're down to once a week which is liveable but it does feel a lot like a memory leak.

Happy to pull the logs if required, though I'll need a pointer as I don't see where I can download a copy on this device.


Best,

Ryan

[rymcc86]

Just now, rymcc86 said:

Hey Fraser, 

I changed into router mode, configured the virgin hub the two have been working fine together. It actually seems to be performing better with the virgin hub acting as the WAN router. 

Then today, I got another freeze. So we've improved, we're down to once a week which is liveable but it does feel a lot like a memory leak.

Happy to pull the logs if required, though I'll need a pointer as I don't see where I can download a copy on this device.


Best,

Ryan

Found log file download button on a hunch, see attached.

log-1623613689015.txt

[Netduma Fraser]

Great to hear! As far as I can tell the WAN IP of the R2 changed so double check the IP shown on the System Information page now (you may need to update the DMZ on the hub with this IP) then set it as static/reserved on the hub and see if it lasts longer than a week please.

[rymcc86]

Interestingly I got a second failure 15mins ago while I was looking at your message. 

I now have a static reservation on the Virgin Hub, the router IP is now in the DMZ on the Virgin Hub and on the Duma I've configured the static IP rather than DHCP.

I'll monitor it now. Next time it freezes is there any way to capture the pre-freeze/reboot logs? I can capture them following the issue but that's only giving us half the picture

[Netduma Fraser]

Great to hear that, let us know how you get on! Only if you have access to the interface at the time unfortunately, if it's the issue I'm thinking of then we're already looking into it and collecting as much data as we need

[rymcc86]

Hey Fraser, 

I've started to notice the symptoms which usually lead up to the router freezing/crashing, but the device has not locked up yet.

I dont see any smoking gun in the device metrics beyond an 80% CPU spike upon login but that drops off pretty quickly.



I've uploaded my latest pre-crash logs.

 

 

Best,

Ryan

log-1624523389847.txt

[Netduma Liam]

Awesome, thank you.

Do you know roughly the time that you started noticing the symptoms?

[rymcc86]

Not specifically Liam no, it starts with random web pages not loading. No particular pattern. 

It's been going on for ~2-3 days, usually the complete crash/freeze happens shortly after.

[Netduma Liam]

12 minutes ago, rymcc86 said:

Not specifically Liam no, it starts with random web pages not loading. No particular pattern. 

It's been going on for ~2-3 days, usually the complete crash/freeze happens shortly after.

Right OK thanks, do you notice the base of the router getting particularly hot during this time?

[rymcc86]

No it feels within operating params but if there is a method to connect via SSH and glean this info I'm more than happy to do it.

[Netduma Liam]

I'm afraid not, you won't be able to access the R2 via SSH.

As you mentioned you typically see this behaviour before a crash, could you please click the three-dot menu in the top right of the interface and go to Advanced -> Rapp performance. Click the advanced tab and hover your cursor just above the Rapp column. You should see a 'Show kernel values' box appear, tick that and send over a screenshot please!  

[rymcc86]

[Netduma Fraser]

The kernel gap value is higher than it should be there. A few users have found correlation between the heat of the router and issues like this. Make sure the router is elevated and has good airflow, if possible pointing a small fan towards it could help. I'm confident that will reduce the frequency of the issue

[rymcc86]

Come on, that can't be it surely? The temps are well within acceptable levels for tech equipment, the area where it lives is out in the open, cool and well ventilated (just on top of a cabinet near the TV).

And if it is, and this thing is that sensitive to heat surely that's a manufacturing defect?

 

What are the long term plans to remediate this issue? The ability to schedule a regular reboot would be valuable here.

[Netduma Fraser]

That's just a recommendation based on what others with similar issues have experienced. It's just something to be wary of while we work on the issue. We're actively working with a user and gaining lots of useful information from them so it's only a matter of time before we nail down the issue. Hoping to have a new version available very soon. That is on our roadmap as well