Beta Asuswrt-Merlin 386.9 beta is now available for AC models

[ColinTaylor]

Looks like dnsmasq will attempt to retry the original packet size after a wait of 60 (not sure if it’s measured in seconds or minutes...

Seconds apparently. https://www.tutorialspoint.com/c_standard_library/c_function_difftime.htm

 

[SheikhSheikha]

Try just doing an electrical reset of your router (i.e. unplug the power for 5-10 secs, then plug it back in).

Thank you Erik for the response. I tried that again. (had already switched it off for 15 minutes). Unfortunately no improvement. I have done a full reset and clean install and now there is no longer a difference in signal strength on 2.4 main and 2.4Guest is zero. Apparently my routers needed that :-D

Have taken the opportunity to fully clean install all devices (not only the smart ones) and as it looks my network is ready for 2023.

 

[dave14305]

Seconds apparently. https://www.tutorialspoint.com/c_standard_library/c_function_difftime.htm

Yes, but then dnsmasq has its own time function dnsmasq_time for now that i still don’t understand. No wonder most people hate this code.

 

[SheikhSheikha]

try uploading the firmware again!

And clear the cache of your browser and unplug the power for a short while after the first reboot after you have flashed.

 

[Mogsy]

This is just a kernel warning that conntrack helpers (such as ct_sip or ct_ftp) will be removed in a future kernel version. This is irrelevant in the router's case since it won't get any kernel upgrade, unlike a server where you might be upgrading to a newer distro.

CT helpers can be disabled on the WAN -> NAT Passthrough page (in fact most of these should already be disabled by default, unless you possibly came from stock firmware and didn`t do a factory default reset, in which case you might be using the stock Asuswrt defaults).

You're right. Just tried to reproduce by launching the app and scan for ports/services nf_conntrack warning no longer pop ups.

 

[Tech9]

About dnsmasq message - I've seen it once to OpenDNS with no DoT. No ill effects though. Log very quiet when set on debug.

[ALPHA] 386.9_alpha1 for WiFi 5 (802.11ac) Models (GT-AC2900, RT-AC68U, and RT-AC86U)

Wished ASUS pushed their gpl for the ac5300. Never mind...;-) So, removed foot from mouth, applied update, works like a charm both with and without Nord VPN. Only missing piece was the IPv6 function. That never worked in NATIVE, but I don't use it due to Nord being required from time to time...

www.snbforums.com

Just got kicked out of the GUI when scrolling through system log. Forgot to increase the time after last test run. Is this thing fixable?

 

[heysoundude]

RT-AC66U B1 ... try update from 386.7_2 to 386.9_beta1 ...after reboot, still on 386.7_2

your browser is looking at the cached GUI page. Shift-F5 will clear that out of your way and show you what's happening now

 

[RMerlin]

Is there a chance this new beta will fix "Parental Controls, time scheduling".....every since latest alpha update time scheduling does not work properly....it often does not cut the intended device off from the net

Nobody has provided me with the requested info so far (what's on the line where the syntax error is). Since I cannot reproduce the issue, I cannot fix anything without additional information.

Looks like dnsmasq will attempt to retry the original packet size after a wait of 60 (not sure if it’s measured in seconds or minutes, but minutes would explain the log timestamps).

That would explain it then, it would mean it's not as permanent as the code comment above hints it to be. That makes it more problematic then to support large buffer sizes if you would end up being downgraded again every hour. That makes little sense to me, the downgrade should be permanent (i.e. until next restart).

About dnsmasq message - I've seen it once to OpenDNS with no DoT. No ill effects though. Log very quiet when set on debug.

It doesn't hurt in any way indeed (except for that particular query that might need to be resent?), it's dnsmasq auto-adjusting itself to a server that's unable to handle a larger reply without fragmentation. It might be annoying however if truly it reverts back to the large buffer size every hour or so, and generate a new LOG_WARN entry every time. I'm kinda tempted to move it to LOG_DEBUG level.

Just got kicked out of the GUI when scrolling through system log. Forgot to increase the time after last test run. Is this thing fixable?

Some pages are supposed to be excluded from auto logging off. I don't remember where or how the logic is implemented for this, I would have to try to track it down to confirm if the system log page is indeed excluded or not.

 

[RMerlin]

your browser is looking at the cached GUI page. Shift-F5 will clear that out of your way and show you what's happening now

The version string isn't cached. His upgrade failed, most likely reason being lack of free RAM.

 

[dave14305]

That makes little sense to me, the downgrade should be permanent (i.e. until next restart).

Yes, I don’t see how or why a DNS server would suddenly start accepting larger UDP packets out of the blue. I’m sure Simon has a good reason.

 

[Clark Griswald]

It's a pleasure reading the Forum Thoughts working out solutions.
beta working well in my environment.

 

[Tech9]

it's dnsmasq auto-adjusting itself

Up and running for 14h, same OpenDNS servers. Hit it with thousands of DNS queries with DoT on/off and nothing in logs.

Some pages are supposed to be excluded from auto logging off

Not a big deal, but once I was playing with DHCP manual reservations and custom icons and got logged out before hitting Apply.

 

[Master3395]

Run "top" over SSH and see what is using your CPU.

I have ran top now, and it seems to be dnsmasq that takes up a lot from time to time.
But i will keep checking.

But I also noticed this on the network map page
It shows no info below Ethernet ports.

 

[RMerlin]

Yes, I don’t see how or why a DNS server would suddenly start accepting larger UDP packets out of the blue. I’m sure Simon has a good reason.

Did some more reading. If I understand this correctly, a DNS resolver that supports EDNS0 should be able to provide the maximum supported size. I found a query that's larger than 1232 bytes:

dig +dnssec twitter.com txt

Running the queries from a Linux client, I get a value of 1232 from Quad9, and 1280 from my router (which uses my ISP's resolvers) in the EDNS0 header.

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags: do; udp: 1232

In both cases queries were successfully resolved.

I'm wondering if maybe it's getdns breaking EDNS0 here, causing dnsmasq to be expecting 1280 bytes but quad9 only allowing for 1232, without the two negotiating the buffer size. I'd have to setup a DOT server to test that theory. But at this point, I'm tempted to just reduce the max size to 1232, and call it a day. I could change the log priority of that warning, but I try to avoid patching components such as dnsmasq/openssl as much as possible, to make it easier to update them when needed.

 

[RMerlin]

But i also noticed this on the netwrokmap page
It shows no info below Ethernet-ports.

Open the browser console, and look for a Javascript error. What's the content of the line with the error?

 

[laracroftonline]

I have ran top now, and it seems to be dnsmasq that takesup a lot from time to time.
But i will keep checking.
View attachment 46820

But i also noticed this on the netwrokmap page
It shows no info below Ethernet-ports.
View attachment 46819

have you tried a forced refresh?

 

[Yota]

Some pages are supposed to be excluded from auto logging off. I don't remember where or how the logic is implemented for this, I would have to try to track it down to confirm if the system log page is indeed excluded or not.

The syslog page was never excluded, but the Traffic Monitor page was, and if you need a reference code, that's the place to start.

If you decide to fix that page, could you also consider this request?

[Feature Request] Make it easier to copy the MAC addresses of the Time Scheduling page

I don't know when it was broken, maybe since 386, I can't copy anything from this page (http://192.168.50.1/ParentalControl.asp ), I'm using Firefox on desktop. Some copyrighted sites usually use JavaScript to prohibit copying of text content, but I don't understand why routers are blocking it...

www.snbforums.com

Happy New Year

 

[panther7]

Did you use the Beta version firmware for the RT-AC68U to update your RT-AC66U B1?
https://sourceforge.net/projects/asuswrt-merlin/files/RT-AC66U_B1/Use the RT-AC68U firmware.txt/download

As a troubleshooting step try clearing the browser's cache.

Ofcourse, i used AC68U.

try uploading the firmware again!

Ofcourse, same issue.

---

UI is not cached, after ssh login have old ssh, `openssh version` is stil on 1.1.1q... is some log about "fail" flash?

 

[John Fitzgerald]

Ofcourse, i used AC68U.


Ofcourse, same issue.

---

UI is not cached, after ssh login have old ssh, `openssh version` is stil on 1.1.1q... is some log about "fail" flash?

Things to try:
electrical reset
unplug all usb devices
manual upload, "do not use app", be hardwired
disconnect from modem, the wire

If all fails then it's time for a full reset

 

[bennor]

UI is not cached, after ssh login have old ssh, `openssh version` is stil on 1.1.1q... is some log about "fail" flash?

Other suggestions (if you haven't tried them already).
If you have a USB drive(s) attached to the router, remove them before trying to update the firmware.
Do a hard reset then try to update the firmware after a basic setup post reset.
Disconnect all other Ethernet cables leaving just the one wired PC you use to update the firmware.
Try a different wired computer.
Try a different web browser.
Temporarily disable any browser extensions/add-ons.
Temporarily disable any security/anti virus programs on the computer.
On the computer, configure a fixed IP address in the same IP range as the router.
Disconnect any upstream devices from the router's WAN port.