Release ASUS RT-AX86U Firmware version 3.0.0.4.386_49447 (2022/06/20)

[Tech9]

Could try 5 and see what happens.

CONSOLE messages still there.

It doesn't bother me much because this router was purchased for experiments only, but it's annoying to see. The router was released >2 years ago and the users are still beta testers. Not a bad device for the price, bit I'm not impressed by Wi-Fi capabilities. The AX88U I had available before was better in my Wi-Fi environment and with my test clients. Anyway, waiting to see further firmware developments.

By the way... WTFast is available on AC86U, but not on AX86U? WTF? I want $50 back...

 

[OzarkEdge]

It's not Google Chrome QUIC, but Google QUIC:

That was understood. It's not active in Edge... yet.

In general, consumers are way behind in regulating excess.

OE

 

[OzarkEdge]

By the way... WTFast is available on AC86U, but not on AX86U? WTF? I want $50 back...

You got what you paid for.

OE

 

[Paliv]

You got what you paid for.

View attachment 42210

OE

WTFast is a gimmick anyway.

 

[Tech9]

With current gas prices this is a good thing, I guess:

Jun 26 14:55:14 kernel: CONSOLE: 197011.100 wl1: wlc_mbo_gas_event_cb Ignore GAS request

 

[SAL9K]

CONSOLE messages still there.

It doesn't bother me much because this router was purchased for experiments only, but it's annoying to see. The router was released >2 years ago and the users are still beta testers. Not a bad device for the price, bit I'm not impressed by Wi-Fi capabilities. The AX88U I had available before was better in my Wi-Fi environment and with my test clients. Anyway, waiting to see further firmware developments.

By the way... WTFast is available on AC86U, but not on AX86U? WTF? I want $50 back...

The ranker/charts show ax86u/ax88u neck to neck on 2.4G (slight edge to ax88u), and the ax86u clearly better on 5G, and multi-band throughput. Are these charts no longer reliable (due to the many f/w updates since the testing), or you're 2.4G heavy?

https://www.smallnetbuilder.com/too...-asus_rtax86u/4731-asus_rtax88u?testmethod=44

I do see, that the ax88u overall multi-band latency performance is considerably better (lower latency), but that may be because the ax86u is having to route ~50% more total data over its 2.5GbE WAN (1432 Mbps), versus the 941Mbps the ax88u can do with its 1GbE limit. A better latency comparison would have been to test the ax86u with the 1GbE WAN, effectively comparing latency with a fixed/capped 1GbE throughput on both routers. Latency vs throughput would be a telling metric (that isn’t given in SNB charts).

 

[SteverinoLA]

One other I notice aside for massive log crap in this update is that it maxes out the 1GB RAM in my 86U. I have a hard drive connected to USB. Previously, I had 500MB RAM unused unless I accessed something on the drive but not it clearly caches pro-actively. That's good, unused RAM is wasted RAM.

 

[tallytr]

I learned to totally ignore the log spam and solely focus on the router performance, the last update ran stable for a couple months, once I had to re-boot to get the 160MHz working again for one client (notebook), but only after I looked at the link speed, had no impact at all
on my work, I guess 80 MHz in practice works just fine, that's just me of course...

 

[Paliv]

I learned to totally ignore the log spam and solely focus on the router performance, the last update ran stable for a couple months, once I had to re-boot to get the 160MHz working again for one client (notebook), but only after I looked at the link speed, had no impact at all
on my work, I guess 80 MHz in practice works just fine, that's just me of course...

Good approach with the logs. I only have one 160 client, and it’s wired so it’s not worth the potential DFS issues to gain a couple hundred more Mbps over 940.

 

[Tech9]

Oops... large transfer to USB storage, httpd crashed:

Jun 27 09:14:07 kernel: CPU: 2 PID: 1818 Comm: httpd Tainted: P           O    4.1.52 #2
Jun 27 09:14:07 kernel: Hardware name: Broadcom-v8A (DT)
Jun 27 09:14:07 kernel: task: ffffffc032152180 ti: ffffffc02ca7c000 task.ti: ffffffc02ca7c000
Jun 27 09:14:07 kernel: PC is at 0xf7552dc8
Jun 27 09:14:07 kernel: LR is at 0xf7552a44
Jun 27 09:14:07 kernel: pc : [<00000000f7552dc8>] lr : [<00000000f7552a44>] pstate: a00e0010
Jun 27 09:14:07 kernel: sp : 00000000fff8d5e8
Jun 27 09:14:07 kernel: x12: 0000000000000000 
Jun 27 09:14:07 kernel: x11: 0000000000000000 x10: 000000000000746e 
Jun 27 09:14:07 kernel: x9 : 000000000021f354 x8 : 0000000000000000 
Jun 27 09:14:07 kernel: x7 : 00000000001f0f30 x6 : 00000000002267c0 
Jun 27 09:14:07 kernel: x5 : 0000000000000000 x4 : 0000000000000000 
Jun 27 09:14:07 kernel: x3 : 00000000fffca61c x2 : 00000000fffca61c 
Jun 27 09:14:07 kernel: x1 : 0000000000000000 x0 : 00000000fff8d664 
Jun 27 09:14:27 watchdog: restart httpd
Jun 27 09:14:27 rc_service: watchdog 1828:notify_rc stop_httpd
Jun 27 09:14:27 rc_service: watchdog 1828:notify_rc start_httpd

 

[SAL9K]

Oops... large transfer to USB storage, httpd crashed:

Jun 27 09:14:07 kernel: CPU: 2 PID: 1818 Comm: httpd Tainted: P           O    4.1.52 #2
Jun 27 09:14:07 kernel: Hardware name: Broadcom-v8A (DT)
Jun 27 09:14:07 kernel: task: ffffffc032152180 ti: ffffffc02ca7c000 task.ti: ffffffc02ca7c000
Jun 27 09:14:07 kernel: PC is at 0xf7552dc8
Jun 27 09:14:07 kernel: LR is at 0xf7552a44
Jun 27 09:14:07 kernel: pc : [<00000000f7552dc8>] lr : [<00000000f7552a44>] pstate: a00e0010
Jun 27 09:14:07 kernel: sp : 00000000fff8d5e8
Jun 27 09:14:07 kernel: x12: 0000000000000000
Jun 27 09:14:07 kernel: x11: 0000000000000000 x10: 000000000000746e
Jun 27 09:14:07 kernel: x9 : 000000000021f354 x8 : 0000000000000000
Jun 27 09:14:07 kernel: x7 : 00000000001f0f30 x6 : 00000000002267c0
Jun 27 09:14:07 kernel: x5 : 0000000000000000 x4 : 0000000000000000
Jun 27 09:14:07 kernel: x3 : 00000000fffca61c x2 : 00000000fffca61c
Jun 27 09:14:07 kernel: x1 : 0000000000000000 x0 : 00000000fff8d664
Jun 27 09:14:27 watchdog: restart httpd
Jun 27 09:14:27 rc_service: watchdog 1828:notify_rc stop_httpd
Jun 27 09:14:27 rc_service: watchdog 1828:notify_rc start_httpd

Excellent, stress testing! Are you running the stock 49447, or a Merlin build? Also, some info on how this error was produced and the drive configuration would be helpful.

 

[Tech9]

Are you running the stock 49447, or a Merlin build?

Yes, current Asuswrt 49447. This router is in double NAT behind my Netgate main router/firewall and I encountered some issues in Asuswrt-Merlin 386.7 release. Also, I wanted to play with AiMesh with available AC86U the way it is described by Asus and it works in Asuswrt only.

Also, some info on how this error was produced

Just a few GB file transfer to SanDisk Ultra Flair 32GB in USB 3.0 mode. I wanted to see the CPU utilization during the transfer and the GUI stopped responding. It recovered 20 seconds later leaving the log above. This firmware does not drop caches even with Samba currently disabled. Strange behavior for me. What the cache is going to be used for if Samba is disabled?

 

[SAL9K]

Good approach with the logs. I only have one 160 client, and it’s wired so it’s not worth the potential DFS issues to gain a couple hundred more Mbps over 940.

It kinda makes sense that Asus would intentionally make the logs more verbose after updating the wireless SDK on this 49447. Like others have suggested, they'll probably go back to a lower verbosity after the changes are beta tested on us, lol. Log writes shouldn't take up resources anyway.

I've been having much better 160MHz success with this 49447. On the previous 46061, with all settings set to auto, the router would rarely set itself up with 160MHz, and if it did, after making a 160->80 MHz transition, would never go back to using 160MHz until rebooted. With this new 49447, I've seen it go vice-versa 80<->160 MHz pretty regularly. It seems to be checking the channel noise and updating every 1hr or so. Btw, I always hard-reset after a f/w update.

 

[OzarkEdge]

I've been having much better 160MHz success with this 49447. On the previous 46061, with all settings set to auto, the router would rarely set itself up with 160MHz, and if it did, after making a 160->80 MHz transition, would never go back to using 160MHz until rebooted. With this new 49447, I've seen it go vice-versa 80<->160 MHz pretty regularly. It seems to be checking the channel noise and updating every 1hr or so. Btw, I always hard-reset after a f/w update.

Admittedly, with only one wired ax client, I did not test 160MHz much prior to this firmware. The following is holding 160MHz here for days:

5.0 auto, low DFS - 20/40/80/160MHz bw; ch Auto, excluding DFS channels (36-48,149-165)

I first tried:

5.0 fixed, low DFS - 160MHz bw; ch 36-48 (non-DFS)

And when the connection dropped to 80MHz, I changed to auto above. I probably should have waited longer to test it... and should try some other non-DFS control channels.

The not knowing what it should do sort of discourages testing. For instance, if set to the fixed settings above, is it suppose to return to 160MHz after dropping to 80 MHz, or just stick at 80MHz since it didn't stick at 160MHz(?).

OE

 

[bbunge]

Yes, current Asuswrt 49447. This router is in double NAT behind my Netgate main router/firewall and I encountered some issues in Asuswrt-Merlin 386.7 release. Also, I wanted to play with AiMesh with available AC86U the way it is described by Asus and it works in Asuswrt only.



Just a few GB file transfer to SanDisk Ultra Flair 32GB in USB 3.0 mode. I wanted to see the CPU utilization during the transfer and the GUI stopped responding. It recovered 20 seconds later leaving the log above. This firmware does not drop caches even with Samba currently disabled. Strange behavior for me. What the cache is going to be used for if Samba is disabled?

View attachment 42238

Well, so much for everything working well... Just tried to copy to/from the USB2 SanDisk thumbdrive formatted with FAT32. Faiiled on both and router GUI locked up. Power cycle in order I guess..
Also noted Merlin 386.7 RAM usage went up to the 90%+ range. Only good thing there is you can set a swap partition or file which should help with the high RAM usage.

Edit: Should add that I tested with both USB3 and USB2 thumb drives. Made no difference. Now I have media servers and AiCloud disabled and the thumb drive removed. I suspect that the RAM usage will go up and lock the GUI again. OK, so I am being pessimistic here with the Asus firmware. Just may go back to Merlin... Can't wait for the pokes and teases.

 

[SAL9K]

I normally don't use ax86u as a file server, but was curious. Attached a 2TB lacie drive, file system NTFS, and tried copying 2TB files. RAM usage also goes up,
~500MB to ~994MB (almost full usage), however file copies over at 90 MB/sec. I've tried writing several large files, and nothing locks up for me. Now, I just need to figure out how to safely unmount the usb drive, lol.

EDIT: There's a big "Eject USB disk" button, top-right of GUI!

 

[Tech9]

I can crash https again and again with transfers to USB stick. There is no better firmware @bbunge. Just different issues. Reports for dnsmasq crashing on Merlin. Also needs fixing, I guess.

 

[SAL9K]

Hmm, are you all doing factory resets after f/w flashes?

Also, there's a typo in the logs! This router is going in the trash!

Jun 27 13:01:32 Samba Server: smb daemon is stoped

 

[Tech9]

Yes, my test routers perhaps have more reset time combined than run time. Reset button in Asus routers is important part of the user experience.

 

[SAL9K]

Yes, my test routers perhaps have more reset time combined than run time. Reset button in Asus routers is important part of the user experience.

Yep, indeed. Good thing they've only been publishing new stock f/w every ~5mo on this one, lol.