Release Asuswrt-Merlin 3006.102.6 is now available

[bpsmicro]

The other thing I'm noticing is on the System Status page Core 2 & Core 4 are pretty much pegged at 100% almost all the time now, and Core 1 & Core 3 bumping between 50% and 100%. And that's with the router not really doing anything. I have no "extra" options enabled, just firewall rules.
CPU temp is hovering around 60C.

 

[rickysuperdog]

BE92U dirty upgrade with no issues for 4 days. 4 cores of CPU usage normal and temperature normal. VPN server on, Firewall off, QOS disabled. Using separate channels, manual WIFI channels parameters setting and MLO was enabled.

 

[bpsmicro]

BE92U dirty upgrade with no issues for 4 days. 4 cores of CPU usage normal and temperature normal. VPN server on, Firewall off, QOS disabled. Using separate channels, manual WIFI channels parameters setting and MLO was enabled.

Interesting. Other things I have that might be different is I have syslogging on (to a local machine), SNMP enabled (get only), a handful of ports open for incoming, 31 "manual IP" settings, and AIMesh with two nodes.
This morning CPU cores are back to low usage, temp just under 60. It appears to have rebooted at least once overnight.
Today I'm going to turn off various settings to see if I can narrow it down, but it's like troubleshooting a car. Sometimes a problem can vanish for hours or even days and you *think* it's solved, and then it comes back. :-(
I'll also re-update to 102.6, since it appears this specific problem isn't "new". I'm getting pretty quick at re-doing my customizations from a factory reset.

 

[AlbumMan]

Dirty flashed RT-AX86U Pro main router 48 hours ago from 102.5 directly to 102.6 skipping alphas/betas this time. The RT-AX3000 node running 388.10 connected and the combined 30-something IoT devices connected as they were and cruising along fine. Checked logs this morning that have been clean {they are set to notice/notice} and saw this a few minutes ago. From other posts I've read, I think it can be ignored.

Nov 29 06:23:07 kernel: CPU: 3 PID: 2418 Comm: asd Tainted: P O 4.19.183 #1
Nov 29 06:23:07 kernel: Hardware name: RTAX86U_PRO (DT)
Nov 29 06:23:07 kernel: pstate: 20000010 (nzCv q A32 LE aif)
Nov 29 06:23:07 kernel: pc : 00000000f7351894
Nov 29 06:23:07 kernel: lr : 00000000f7353450
Nov 29 06:23:07 kernel: sp : 00000000ffd36af0
Nov 29 06:23:07 kernel: x12: 00000000f7366174
Nov 29 06:23:07 kernel: x11: 00000000f7354b48 x10: 00000000f7351b88
Nov 29 06:23:07 kernel: x9 : 00000000f73544f8 x8 : 0000000000000000
Nov 29 06:23:07 kernel: x7 : 00000000002ae580 x6 : 00000000002ae580
Nov 29 06:23:07 kernel: x5 : 00000000002b0a80 x4 : 00000000002af178
Nov 29 06:23:07 kernel: x3 : 0000000000006b63 x2 : 00000000f7000b88
Nov 29 06:23:07 kernel: x1 : 0000000000000400 x0 : 0000000000000000

 

[GWTechTalk]

Just confirming all issues remediated with a factory reset and manual reconfigure. I bet a lot of these users that are not having issues with dirty flashes are likely using very simple “auto” configs. For me, dirty flashes seem to break built in static leases and UI stuff. Not major but I’m going to have to look into YazDHCP because dirty or not it’s a separate unaffected module. At least with my custom firewall script rules that’s easy to drop back on and execute. I feel bad for users using this device as an enterprise solution because it’s dangerous to restore from an old version config file. You never know what changes under the hood. Only ASUS really does most times. So with each version that dirty flashes causes issue the quickest and most likely method to be issue free is factory reset and manual reconfigure. Still loving the Merlin builds but it’s got its quirks.

 

[lgkahn]

not for me anyway... i did dirty flash on my main router from alpha and betas..(it was a fresh setup with reset on 105.0 ...
i have more than a class c ie 254 mask and over 200 devices about 110-120 (as best i can count manually) static leases. i do not have a firewall however as i have an upstream mikrotik router and i have 13 public ips with an extensive firewall.. the asus router maps one of the public to my privates.

 

[Bob575]

Filthy upgrade from 3006.102.5 to 3006.102.6 on all supported devices.
No Hits, No Runs, No Errors, Nobody Left on Base.
Even my wonky doorbell camera likes the IoT network.
Many Thanks @RMerlin

 

[dineshgyl]

Did a dirty update from 3006.102.5 to 3006.102.6 - Everything seems working except lighttpd.

I was working on 3006.102.5

/tmp/home/root# cat lighttpd.conf
server.port = 20000
server.document-root = "/tmp/home/root"
server.indexfiles = ( "index.html", "index.htm")

# Enable MIME type assignment
mimetype.assign = (
    ".html" => "text/html",
    ".htm"  => "text/html"
)

/tmp/home/root#  /usr/sbin/lighttpd -f lighttpd.conf
/tmp/home/root# echo $?
255

Seems lighttpd was changed

/tmp/home/root# ls -l /usr/sbin/lighttpd
-rwxr-xr-x    1 asusadmi root        212716 Nov 25 11:30 /usr/sbin/lighttpd

/tmp/home/root# lighttpd -v
lighttpd/1.4.39 (ssl) - a light and fast webserver
Build-Date: Nov 25 2025 11:18:08

 

[Mutzli]

Just confirming all issues remediated with a factory reset and manual reconfigure. I bet a lot of these users that are not having issues with dirty flashes are likely using very simple “auto” configs. For me, dirty flashes seem to break built in static leases and UI stuff. Not major but I’m going to have to look into YazDHCP because dirty or not it’s a separate unaffected module. At least with my custom firewall script rules that’s easy to drop back on and execute. I feel bad for users using this device as an enterprise solution because it’s dangerous to restore from an old version config file. You never know what changes under the hood. Only ASUS really does most times. So with each version that dirty flashes causes issue the quickest and most likely method to be issue free is factory reset and manual reconfigure. Still loving the Merlin builds but it’s got its quirks.

Before flashing I always run this command:

sync; echo 3 > /proc/sys/vm/drop_caches

to clear the caches. I’ve been doing this for a couple of years now before flashing a new firmware. In the past, when performing dirty flashes, I occasionally ran into issues. After I started clearing the caches first, those problems disappeared and the process has worked reliably every time since.

 

[CaptainLockness]

Before flashing I always run this command:

sync; echo 3 > /proc/sys/vm/drop_caches

to clear the caches. I’ve been doing this for a couple of years now before flashing a new firmware. In the past, when performing dirty flashes, I occasionally ran into issues. After I started clearing the caches first, those problems disappeared and the process has worked reliably every time since.

This can be useful, bookmarked it for future updates.

 

[Natty]

Before flashing I always run this command:

sync; echo 3 > /proc/sys/vm/drop_caches

to clear the caches. I’ve been doing this for a couple of years now before flashing a new firmware. In the past, when performing dirty flashes, I occasionally ran into issues. After I started clearing the caches first, those problems disappeared and the process has worked reliably every time since.

Would power cycling achieve the same result, or are the caches non-volatile?

 

[RMerlin]

Did a dirty update from 3006.102.5 to 3006.102.6 - Everything seems working except lighttpd.

You cannot use lighttpd yourself. It's a customized version that's specifically for use with AiCloud and Covchili. Custom config files will be ignored/overwritten.

 

[Alexander Savkin]

Hello everyone!
This question may have been asked before, but I missed it...
My GT-AX6000 currently has firmware version 3004.388.8_4. If I upgrade to version 3006.102.6, will I be able to revert to 3004.388.8_4 if I have problems? As I understand it, a factory reset is required.
Is a reset required when upgrading from 3004 to 3006, or is there a chance everything will work without a reset?
Thanks!

 

[GWTechTalk]

Before flashing I always run this command:

sync; echo 3 > /proc/sys/vm/drop_caches

to clear the caches. I’ve been doing this for a couple of years now before flashing a new firmware. In the past, when performing dirty flashes, I occasionally ran into issues. After I started clearing the caches first, those problems disappeared and the process has worked reliably every time since.

Thanks Ill start doing that. However, issue returned this morning. Windows 11 most up to date devices are not consistently being shown when connected to 6Ghz bands. I can go in and force a refresh and they come back for a bit but then they disappear again after a while. Not sure what's causing the issue but Ill like have to roll back a couple versions to check the behavior. I don't see anything in the debug log to indicate what's happening.

 

[NoLight]

Hello everyone!
This question may have been asked before, but I missed it...
My GT-AX6000 currently has firmware version 3004.388.8_4. If I upgrade to version 3006.102.6, will I be able to revert to 3004.388.8_4 if I have problems? As I understand it, a factory reset is required.
Is a reset required when upgrading from 3004 to 3006, or is there a chance everything will work without a reset?
Thanks!

Always do a factory reset clearing all settings followed by a manual config when jumping between major versions or going from stock to Merlin

 

[DrTurk]

Dirty update from 3006.102.5 to 3006.102.6 without any noticeable problems. Thanks RMerlin.

 

[octopus]

3 days uptime working just fine.
Discovered missing pictures.

 

[heywire9]

Unable to flash my RT-BE96U to latest 3006.102.6 firmware. Currently on 3006.102.5, I flash it multiple times and each time, I would go and check and my firmware still shows as 3006.102.5 even after manually rebooting. Anyone has any suggestions?

 

[heywire9]

Unable to flash my RT-BE96U to latest 3006.102.6 firmware. Currently on 3006.102.5, I flash it multiple times and each time, I would go and check and my firmware still shows as 3006.102.5 even after manually rebooting. Anyone has any suggestions?

After my 5 time, the firmware finally took and is now displayed correctly as 3006.102.6. Wanted to put this here in case others has the same issue. But his time I did not manually reboot. Just waited and once the router came back up. I just reconnected to the wifi and checked the version.

 

[TheLev]

Before flashing I always run this command:

sync; echo 3 > /proc/sys/vm/drop_caches

to clear the caches. I’ve been doing this for a couple of years now before flashing a new firmware. In the past, when performing dirty flashes, I occasionally ran into issues. After I started clearing the caches first, those problems disappeared and the process has worked reliably every time since.

Is there any way to (or reason not to) have the firmware updater do this automatically before doing anything else?