DNS broken after upgrade to 4.9.0

[marcgise]

On Flint 3, I upgraded the firmware to 4.9.0, and I found that DNS is not working. With Chatgpt I tried to solve it, and after disabling Ad Guard, everything work again. But if I enable it, DNS is not working. For instance, DDNS is not working. My public IP is not binded to glddns if Ad Guard is enabled.

I have two other Flint 3 connected to the main router as AP , and these 2 routers appears connected to GoodCloud. But tue main router appears offline.

[marcgise]

I finally solve it. I was getting the following error:

Sun May 24 23:41:34 2026 user.notice AdGuardHome[21578]: 2026/05/24 21:41:34.407119 [error] dnsproxy: exchange failed upstream=9.9.9.9:53 question=";checkip.dyndns.com.\tIN\t AAAA" duration=417.877µs err="exchanging with 9.9.9.9:53 over udp: write udp 79.112.130.222:36892->9.9.9.9:53: write: invalid argument"

I do not use IPV6, and I always had it deactivated in the GUI. What I did to solve it was: first enable it, and then disable it again.

I don’t understand why, but magically it worked…

[will.qiu]

Hi,

There seem to be a few strange points here:

1. By default, the router itself (for example services like GL DDNS) should not resolve DNS through AdGuard Home unless the WAN DNS configuration was manually changed to 127.0.0.1.
2. Based on the error messages, the issue also appears to be related to IPv4 UDP rather than IPv6.

In any case, we’re glad to hear that everything is now working properly.

We will continue monitoring to see whether other users encounter similar issues.

[knee_mountain523]

Although my problem is different, I also experienced a problem with the DNS. I've already reported it to you. It seems like disabling and re-enabling things resolves them. Pretty annoying that I spent several hours doing trial and error to figure it out!

[marcgise]

Again DDNS doesn’t work. In the logs appears something like that:

Thu May 28 12:14:20 2026 user.notice AdGuardHome[26607]: 2026/05/28 10:14:20.187040 [error] dnsproxy: exchange failed upstream=9.9.9.9:53 question=";ifconfig.me.\tIN\t A" duration=964.046µs err="exchanging with 9.9.9.9:53 over udp: write udp 79.112.43.27:52937->9.9.9.9:53: write: invalid argument"

This time, if I have AdGuard enabled, DDNS doesn’t work:

If I disable AdGuard, it works. But after enabling it again, DDNS doesn’t work.

I want to highlight that I have not changed nothing in the router.

[marcgise]

I’m facing some errors when updating the packages. These errors doesn’t occur in the 4.8.4 version. Doesn’t know there is something related with…:

Downloading https://fw.gl-inet.com/releases/qsdk_v12.5/kmod-4.7/be9300-ipq53xx/Packages.gz
*** Failed to download the package list from https://fw.gl-inet.com/releases/qsdk_v12.5/kmod-4.7/be9300-ipq53xx/Packages.gz

Downloading https://fw.gl-inet.com/releases/qsdk_v12.5/packages-4.x/ipq53xx/be9300/glinet/Packages.gz
*** Failed to download the package list from https://fw.gl-inet.com/releases/qsdk_v12.5/packages-4.x/ipq53xx/be9300/glinet/Packages.gz

Downloading https://fw.gl-inet.com/releases/qsdk_v12.5/packages-4.x/ipq53xx/be9300/packages/Packages.gz
*** Failed to download the package list from https://fw.gl-inet.com/releases/qsdk_v12.5/packages-4.x/ipq53xx/be9300/packages/Packages.gz

Updating database.
Database update completed.

[will.qiu]

Hi,

These errors appear to be related to the router's own DNS configuration.

The next time the issue occurs, please SSH into the router and run the following commands:

cat /etc/resolv.conf
nslookup google.com

If DNS resolution fails, please try the following and see if this help:

1. Go to LuCI → Network → Interfaces and edit the corresponding WAN interface.

2.Manually change the WAN's DNS settings to another available public DNS server.

---

In addition, under the current network environment, connectivity to the Quad9 upstream DNS configured in AdGuard Home appears to be unreliable.

Please try switching to another public DNS provider and see whether it helps, for example:

• Google: 8.8.8.8
• Cloudflare: 1.1.1.1

[marcgise]

Dear,

The error only appears when I have enables AdGuard. If I disable it, everything works ok. The problem is that I did not change anything before and after upgrading to 4.9.0 version. And it has stopped working on its own.

I have another Flint 3 in other home that still has the 4.8.4 version, configured like the other broken Flint 3, and everything is going ok. This is the reason why I still have not upgraded it to 4.9.0

If I go to DDNS page:

And click on DDNS Test, I always get this error:

And Log says:

Fri May 29 15:27:35 2026 user.notice AdGuardHome[12415]: 2026/05/29 13:27:35.249081 [error] dnsproxy: exchange failed upstream=1.1.1.1:53 question=";ifconfig.me.\tIN\t A" duration=497.293µs err="exchanging with 1.1.1.1:53 over udp: write udp 79.112.43.27:50121->1.1.1.1:53: write: invalid argument"
Fri May 29 15:27:36 2026 user.err ddns-scripts[27747]: glddns: cURL Error: '6'

Of course nslookup to google fails…

Also if I go to AdGuard Home and change the Upstream DNS, when I click on “Test upstreams” button also fails:

The same test, in the other router Flint 3 with 4.8.3 version, this tests works without any problem

[will.qiu]

It appears that AdGuard Home is not functioning properly for some reason.

Could you please temporarily disable it, then follow the guide below and share your device with us via GoodCloud so that we can investigate remotely?

Technical Support via GoodCloud - GL.iNet Router Docs 4

Please also send us the MAC address and the router password via private message so that we can access the device.

[will.qiu]

Thank you for providing remote access.

We have checked the device and it appears to be the same issue discussed in the following thread:

Issues with enabled Wireguard Client connection and AdGuard

Hi, I have a GL-BE9300 router. I have the issue, that if I connect to my parents network with Wireguard, I can’t setup the AdGuard DNS anymore (see last picture) If I disable the Wireguard connection, AdGuard has no issues with the DNS… Here is my setup: [Screenshot from 2026-03-27 14-10-42] [Screenshot from 2026-03-27 14-25-18] [Screenshot from 2026-03-27 14-12-21] [Screenshot from 2026-03-27 14-13-47] Am I missing something?

We have already applied the workaround on your device and verified that the issue has been resolved. Could you please check again on your side?

[marcgise]

I'm out of the house right now. I’m trying to connect to my vpn and I see that router is down. It seems that it’s not connected to Internet… I will try again tomorrow when I go home, or if I can connect

[marcgise]

I don’t know what you changed on my routers when you connected remotely, but all of them are unaccesaible at all. I’m really angry, because nothing is working. I lost my internet connectivity for all the home. Can’t connect to any routers via wifi. Also I can’t connect using cable ethernet. I don’t know what happened, but I’m really angry about this situation…… I think the only solution is to make a hard reset…

[will.qiu]

We sincerely apologize for the inconvenience caused.

The only change we made remotely was adjusting the following AdGuard Home setting. This modification should not affect the normal operation of other router functions, and it has been tested and verified both locally and by other users.

@will.qiu

Please try SSH-ing into the router and run the following commands to force AdGuard Home to send DNS queries through the WAN only:

sed -i 's/explict_vpn/nonevpn/g' /etc/init.d/adguardhome
/etc/init.d/adguardhome restart

# To restore:
cp -r /rom/etc/init.d/adguardhome /etc/init.d/adguardhome
/etc/init.d/adguard

As shown in the screenshots we provided, when we finished the remote session, the router was still connected to the internet normally and had successfully restored connectivity to the AdGuard Home upstream DNS servers, the opkg repositories, and DDNS services after making change.

@will.qiu

Could you please clarify the following?

1. Did you discover the issue as soon as you returned home?
2. What was the status of the router's LED indicators when the issue occurred?
3. Was the device still shown as online in GoodCloud at that time?
4. What exactly happened when devices could not connect to Wi-Fi? Were they unable to connect at all, able to connect but unable to obtain an IP address, or able to connect but unable to access the internet?
5. Did rebooting the router restore normal operation?

[marcgise]

Immediately after applying the changes, I was able to access it. I checked the same you said, and I logout the router. I promise you that I did not changed nothing. After that, I notice that the 3 routers appeared as disconnected in GoodCloud.

I was away from home the weekend, so I could not check nothing until yesterday at night when I came back to home.

The LED of the main router was blinking everytime. I tried connecting via Wi-Fi, but none of the three networks (2.4, 5, and 6 GHz) appeared. I also tried connecting via Ethernet cable, but the laptop wasn't receiving the router's IP address.

I tried restarting the router, but the same thing happened again. Can’t connect to the router. I tried pressing the reset button three or four times, but it didn't do much. Finally, after unplugging the router and waiting a while, I was able to press the reset button and it reset to factory settings, allowing me to configure the router from scratch.

The problem now is I don't have a backup of the router. When I updated from version 4.8.3 to 4.9.0, I did make a backup, which I imagine was saved on the router itself. But when I restored the factory settings, I imagine it was deleted. I have to reconfigure the VPN and everything else. I honestly don't know why this happened.

[will.qiu]

Thanks for your update.

We managed to review the GoodCloud logs and found that the issue may be related to memory usage. Before we fixed the connectivity issue, the available memory was already running very low, with only about 25 MB remaining.

Therefore, it is possible that after we restored network connectivity for those services, they resumed normal operation and consumed additional memory, eventually causing the device to run out of memory and become unresponsive.

However, this is only a hypothesis. Since there are currently no relevant logs available and the device has already been reset to factory settings, it may not be possible to further investigate or confirm the exact cause.

---

Router backups are typically not stored on the router itself. When you click Export, the backup file is automatically downloaded to your local device through the browser.

If you have exported a backup before, please check your browser's download history or the Downloads folder on your device.

How to Export/Import Backup Configuration Files via LuCI

Introduction When using a GL.iNet router, you may need to back up your current system configurations or restore previous settings after a reset or firmware upgrade. This guide will walk you through how to export and import your router's backup configuration files via LuCI. Part 1: How to Export Backup Configuration Files via LuCI Step 1: Access the LuCI Advanced Settings Log into your GL.iNet router's web Admin Panel (the default IP address is usually 192.168.8.1). On the left-hand menu, navi…

[marcgise]

Thanks for your response. I have not any backup. When I was asked about creating a backup before upgrading to 4.9.0, I click “yes” but I don’t remember to be asked about downloading any backup file. This is why I thought the backup was stored on the router itself. OK, no problem about that.

I performed a factory reset because I needed to work today, and otherwise I wouldn't have had internet access. I apologize if this has prevented you from accessing the history to try and find the cause of all this.

If the problem was the memory left in the device, I don’t know why the router is full…

Anyway, thanks for your help

[will.qiu]

Thank you for your update.

@marcgise

When I was asked about creating a backup before upgrading to 4.9.0, I click “yes” but I don’t remember to be asked about downloading any backup file. This is why I thought the backup was stored on the router itself. OK, no problem about that.

The option you selected is used to determine whether the router configuration should be preserved during the firmware upgrade. If it is not checked, the upgrade will be performed without retaining the existing configuration.

Therefore, it is not an option for creating a backup file.

Unfortunately, if the device has since been reset to factory defaults and no backup file is available, all settings may need to be configured manually again.

We sincerely apologize for the inconvenience.

@marcgise

If the problem was the memory left in the device, I don’t know why the router is full…

Once you have restored your previous configuration, you can SSH into the router and run the following commands to check which processes are consuming memory:

opkg update && opkg install htop
htop

# Then use mouse to click MEM% to sort processes by memory usage

Based on our previous experience, this issue can sometimes occur when AdGuard Home is configured with a large number of filter lists.

It is also worth noting that Tailscale and the new DPI feature introduced in v4.9.0 can consume a amount of memory as well.

[marcgise]

Sorry, your’re wright about the button to keep the configuration.

As far as I remember, the only extra package I installed on the 3 routers was Dawn (from openwrt documentation), and another package to see on lucy the mounted devices.