XR300 Is the worst piece of networking equipment I've owned.

[Admiral]

Well the title says it.

 

Back when I bought it I was promised premium features and cutting edge OS with constant updates (at the time 3.0 DumaOS was promised within a year).

It's been over 3 years of promises now, If only I knew that "new update that will fix your issues and give you promised features is right around the corner" would in reality mean that my hardware will become obsolete before I get any updates.

Current state of affairs is that what I have is godawfully buggy piece of kit that needs restarting every few days to even work properly. 
On top of that multiple "premium features" break the router functionality altogether. QoS does not work as intended and results in all clients getting worse experience than with it enabled.
Network is limited to max 100Mbps is Traffic Monitoring function is enabled. Took me full day to figure out what was limiting connection. Now keeping it disabled.
UI having multiple bugs, missing strings and such  (question marks as a replacement, restarting sometimes helps, related to translations). 

IPV6 does not work properly and causes issues. Changing 90% of the settings results in dropped connection to apply the update (seems like a restart on the background) that makes debugging a slow painful hell.

There are definitely more issues that simply don't come to mind at this moment. And please don't come with your have you tried resetting, reinstalling firmware type of suggestions. I've done all that multiple times. QoS configuations have all been according to documentation and other guides and even experimentation on my own without resolve.

 

It is nothing more than old and forgotten unmaintained piece of firmware that's the root of all these issues. You have moved the goalpost on the new firmware so many times that nothing you can say today will make us XR300 owners ever believe your words.

 

I'll try to flash OpenWrt on it to at least get some use out of it before it ends in a landfill. Other XR300 owners I suggest you try the same.

[Netduma Fraser]

Totally understand your frustration, I've never heard of a product becoming end of life just 2 years after release which is what happened here and of course that means that less resources then get put into it. It's something we've planned for to get it up to a higher firmware level but with lack of support and also the architectural challenge of it, it's looking unlikely.

I'm happy to help troubleshoot your issues if you wish to but understand if you don't. 

[Admiral]

So do I understand correctly that at current state no new firmware is planned on XR300 and Netgear has dropped support and funding?

Regarding troubleshooting, sure, let's start with something that I forgot to mention on my original post, but is a source of much frustration in competitive gaming.

This issue has been present since 2020 and multiple people have reported it happening. Same with me, no issues when I skip the XR300 and connect directly to ISP.
I get disconnected from activision/blizzard servers every 2 hours or while playing warzone.

I found these threads, none of them have actual solutions in them, I have tried the steps suggested in these.

 

and these too:
https://community.netgear.com/t5/Nighthawk-Pro-Gaming-Routers/XR300-Disconnects-From-Online-Games-Exactly-at-2-hours/td-p/2209559
https://community.netgear.com/t5/Nighthawk-Pro-Gaming-Routers/XR300-Disconnects-every-2-hours/td-p/2069703

You have gone as far as suggesting completely disabling firewalls, which in my honest opinion is pretty nuts security-wise especially considering it has been clearly established that it's router related issue.


Geo-filtering was disabled, because suprise suprise IT SIMPLY DOES NOT WORK.

Setting the radius and disabling ping assist - Game still connects to any server it wants ourside the radius.
Setting minimum radius and allowing only ping assist to just get best connection - Game still connects to any server it wants with ping higher than selected value.

Tried for following CoD games: MW2, Warzone 2, MW3 Beta. All same results.

[Netduma Fraser]

I'm saying it's still planned on our end but with what I've mentioned it may not go ahead, there is still some hope. What platform are you playing on specifically? Could you grab the log from the System Information page when you get the disconnect please?

[Admiral]

@Netduma Fraser Fuuck please help me, I can't even work at home properly.

I don't even care at the games or QoS atm but please help me get this thing into a state where I at least have stable connection at speeds my ISP provides.

I have 500/500 connection. As soon as these is any noticeable traffic on one device the CPU 0 shoots up to 100% and starts throttling everything. For example about 80 Mbps/s traffic on my Truenas server is all it takes for the whole network to halt, all other devices get throttled very spotty 5-60 Mbps/s max (speedtests) with very frequent drops in video streaming and video calls etc where it's obvious router is struggling.

I have DISABLED ALL the features that I can which could use CPU resources on the router. QoS, Geo-filter, Traffic meter, site blocking, all USB funtions, IPv6, VPN, UPnP.

Just finished another restart sequence like this without any resolve:
 

Quote

• Turn off and unplug modem.
• Turn off router and computers.
• Plug in and turn on modem. Wait 2 minutes for it to connect.
• Turn on the router and wait 2 minutes for it to connect.
• Turn on computers and rest of network.

There is nothing interesting in the logs after the restart, just Firmware initialized, NTP time sync and DHCP ip assignment notices.

There are around 5 devices online. 2 mobiles, 2 PC-s, 1 NAS. 2 Wired, 3 wireless. No quest networks.

[Netduma Fraser]

Does the CPU hitting 100% and throttling only happen when using local traffic on the NAS or does it happen if you're downloading from the internet as well for example? Also is the NAS directly attached via USB to the router?

[Admiral]

Not sure about LAN traffic, but WAN definitely. Both download and upload. There is Nextcloud service for example I use to transfer big files between people. Both incoming and outgoing.

NO USB shenanigans, It's a bare metal server running TrueNas Scale OS attached via 1gig link.

[Netduma Fraser]

Interesting, if you re-enable QoS and set the download/upload sliders to 70%, then initiate one of those transfers/downloads, does the CPU hit 100 and cause the same issues?

[Admiral]

It des not matter, no matter the % set to bufferbloat or device %. 

100Mbits of total traffic is max it can handle and 80-100range maxes out CPU0 at 95-100%

Therefore no excess is being shared because there seems to be a cap that the router itself can handle.

 

 

 

[Netduma Fraser]

If you set the XR to AP mode (only if you have a modem/router that can provide WiFi) and try again does it still occur?

[Admiral]

19 hours ago, Netduma Fraser said:

If you set the XR to AP mode (only if you have a modem/router that can provide WiFi) and try again does it still occur?

If I had another router, believe me I would be using it.

I have XR WAN port connected to ISP fiber ONT. When I connect any one device to ONT directly. There are no issues with it. Also I had router from ISP too before I got the Netgear one and this also did not have any issues like it.

 

Also I really can't see how this would help me resolve the issue.

 

 

[Netduma Fraser]

AP mode disables DumaOS, if the issue wasn't then present then it would be an issue for Netgear specifically to look into. 

Do you still have access to the interface when the issue happens? If so could you get the log from the System Information page please?

[Admiral]

I am sure it's router. I can skip the router and get full speeds no issue.

Yes I have. There is nothing of importance in the logs. Did restart the router. Gave it a download load , maxes out at around 80-100mbit and one CPU core is maxed, and this is all there is in the logs:

[DHCP IP: (192.168.1.93)] to MAC address XX:XX:XX:XX:XX:XX, Tuesday, Oct 31,2023 16:40:22
[DHCP IP: (192.168.1.92)] to MAC address XX:XX:XX:XX:XX:XX, Tuesday, Oct 31,2023 16:36:24
[LAN access from remote] from XXXXXXXXX:48975 to NAS-SERVER-IP, Tuesday, Oct 31,2023 16:36:23
[LAN access from remote] from XXXXXXXXX:52664 to NAS-SERVER-IP, Tuesday, Oct 31,2023 16:36:12
[LAN access from remote] from XXXXXXXXX:34574 to NAS-SERVER-IP, Tuesday, Oct 31,2023 16:36:10
[LAN access from remote] from XXXXXXXXX:53693 to NAS-SERVER-IP, Tuesday, Oct 31,2023 16:36:09
[Internet connected] IP address: MY-WAN-IP, Tuesday, Oct 31,2023 16:36:08
[Time synchronized with NTP server] Tuesday, Oct 31,2023 16:36:08

 

 

[Netduma Fraser]

I know it's the router, which aspect of the router is the important part, could you download the full logs so I can have a look at it further please?

[Admiral]

This was the whole log. I did a restart to ONT and Router just before that. I wanted to clear everything, then replicate the issue and get the log.

Only line that's missing was [Initialized, firmware version: V1.0.3.78_10.3.59] that was present on the downloaded file but wasn't copied here

[Netduma Fraser]

Did you always have this issue or did it happen after upgrading to a specific firmware?

[Admiral]

On 10/31/2023 at 8:04 PM, Netduma Fraser said:

Did you always have this issue or did it happen after upgrading to a specific firmware?

Can not tell for sure because I built the NAS when I was already on latest firmware.

 

So as of now I have nailed down the issue somewhat. I can tell that the issue is related to DHCP with some degree of certainty.

I can resolve the issue temprarily by getting my NAS ip via DHCP.

So my DHCP range starts from 1.150 and I usually want my wired clients to have static IP that that is outside of that range. I had set my NAS ip to .1.80 and also had IP reservation for it's MAC.

This CPU overload issue gets temporarily resolved if I set the NAS to get IP via DHCP.  Why temporarily? Well dhcp does not want to give out IP-s that are outside of it's range. It will do it when changing from static to DHCP probably because the IP was already assigned to it. So at some point when either clients reset or some threshold is reached it will be given new ip.

 

 

Does this information help you figure out what might be causing it and maybe provide a hotfix update?

Is there any way to increase log level so it would output more information? Currently there are no information in logs about clients with static ip-s or potential issues with these, nor which processes cause excessive CPU usage.

 

[Admiral]

To clarify. DHCP does not need to give it new IP in order for the problem to come back, but when it occurs again switching from DHCP to Static and then back to DHCP fixes the issue immediately without needing to reset any devices. IP remains the same in the whole process.

[Netduma Fraser]

If you disable DHCP entirely does that resolve it at all? If it's a DHCP issue that would be something for Netgear to resolve and my hopes would not be high for a hotfix to be honest.