Beta Asuswrt-Merlin 386.1 Beta is now available

[Deleted member 72428]

Ditto...With quality 1GE WAN throughput enabling QoS would be redundant

Apologies for hijacking the thread, this will be my last comment on the subject.

This here again is misinformation and those genuinely looking for solutions will be misguided. What you posted is a simple ping from a speed test which highlights your lack of understanding on bufferbloat. A speed test in an unloaded environment is not an accurate measure of bufferbloat. If you'd rather do the test manually then start 5 different large downloads and then continuously ping a server (google, etc) multiple times a second and see the variance in your ping to that server. Dslreports test can open up to 32 streams both up and down if you really want to test the robustness of your connection. What you want to see is a consistent and low ping throughout, without spikes. This is one of the goals of AQM/SQM algorithms, to prioritize time sensitive packets vs others, establish fairness by taking control of the queue under load, etc. This is far more complex than a simple bandwidth issue as some folks here make it out to be. While there are many variables that affect bufferbloat including ISP, modem, router, neighborhood load (if using cable), etc the things you can control will definitely help if you're looking for consistently low ping under load. If consistently low ping without spikes is not important and outside of your use cases then you need not to concern yourself with it... just please refrain from giving others advice that are looking for solutions to a real problem in their use.

For those that want to read more and inform themselves, there are actual research studies on this topic. They prove the efficacy of fq-codel, cake, and other AQM/SQM solutions for this problem irrespective of bandwidth. There are countless threads on AT&T forums, Dslreports, Reddit, SNB, etc including people with fiber connections. To make a blanket statement of just hey you got a gig connection you don't need to worry is a tad ignorant sorry to say.

Again, apologies for off topic.

 

[Slawek P]

Hi @ASUSWRT_2020, on Merlin 386 beta 3 the problem of slow Ookla tests while TrendMicro enabled remains.
Enabling any TrendMicro stuff like AiProtection or WebTraffic monitor slows down Ookla tests on AX88U router three times (for 1GB+ connections).
Disabling these features does not help, only withdrawing from Trend Micro privacy agreement restores Ookla speeds to 1GBs
TrendMicro changes somehow CPU load distribution on AX to be less favourable

 

[Soul_]

Yes this is my problem. When the AX mode is enabled my 3 google speakers and 1 LG sound bar doesn't see any SSID. When I disabled AX on the 2.4Ghz range my LG sound bar and google speakers work

So if I enable smart connect they wouldn't see any SSID unless I disable AX and that's on everything

When you say HE frames are you saying that's what disabled and not AX ?

I cant get google or LG to provide upgraded drivers I doubt they ever will

Correct. HE Frames are quite necessary for AX to operate. And in reality that toggle is for HE Frames, not AX. Funny thing is my LG C9 works just fine with HE Frames, even though it is not an AX device. I mean I have an old Samsung laser printer that only runs on 2.4GHz, and even that thing is running fine with smart connect and "AX" enabled.

Well in your case, I suppose there isn't a direct UI option to Enable smart connect, and disable HE Frames only for 2.4GHz. Maybe next black friday day swap that soundbar out

 

[cybrnook]

Hi @ASUSWRT_2020

I think you meant the asus beta, not merlin beta?

 

[RMerlin]

Hi @ASUSWRT_2020, on Merlin 386 beta 3 the problem of slow Ookla tests while TrendMicro enabled remains.

Asus has nothing to do with the development of my firmware.

 

[Frankflash]

Apologies for hijacking the thread, this will be my last comment on the subject.

This here again is misinformation and those genuinely looking for solutions will be misguided. What you posted is a simple ping from a speed test which highlights your lack of understanding on bufferbloat. A speed test in an unloaded environment is not an accurate measure of bufferbloat. If you'd rather do the test manually then start 5 different large downloads and then continuously ping a server (google, etc) multiple times a second and see the variance in your ping to that server. Dslreports test can open up to 32 streams both up and down if you really want to test the robustness of your connection. What you want to see is a consistent and low ping throughout, without spikes. This is one of the goals of AQM/SQM algorithms, to prioritize time sensitive packets vs others, establish fairness by taking control of the queue under load, etc. This is far more complex than a simple bandwidth issue as some folks here make it out to be. While there are many variables that affect bufferbloat including ISP, modem, router, neighborhood load (if using cable), etc the things you can control will definitely help if you're looking for consistently low ping under load. If consistently low ping without spikes is not important and outside of your use cases then you need not to concern yourself with it... just please refrain from giving others advice that are looking for solutions to a real problem in their use.

For those that want to read more and inform themselves, there are actual research studies on this topic. They prove the efficacy of fq-codel, cake, and other AQM/SQM solutions for this problem irrespective of bandwidth. There are countless threads on AT&T forums, Dslreports, Reddit, SNB, etc including people with fiber connections. To make a blanket statement of just hey you got a gig connection you don't need to worry is a tad ignorant sorry to say.

Again, apologies for off topic.

netgear for their ax models they have a way better Qos than asus has they have bufferbloat stuff added in so why cant asus make qos better !!!

 

[RMerlin]

Folks, enough with the QoS talks. Any further post will be deleted.

 

[Vexira]

is anyone else getting this error * Reminder: The System time zone is different from your locale setting. even with everything set correctly with daylight savings time ?

 

[Sanna1967]

is anyone else getting this error * Reminder: The System time zone is different from your locale setting. even with everything set correctly with daylight savings time ?

yes

 

[Neil62]

Folks, enough with the QoS talks. Any further post will be deleted.

Thank You

 

[Slawek P]

Asus has nothing to do with the development of my firmware.

Well if they do not fix it on their side it will not work on Merlin either, right? I reported this problem based on pure Asus 386 beta firmware before, clearly latest revisions did not change anything.

 

[beerglass007]

Correct. HE Frames are quite necessary for AX to operate. And in reality that toggle is for HE Frames, not AX. Funny thing is my LG C9 works just fine with HE Frames, even though it is not an AX device. I mean I have an old Samsung laser printer that only runs on 2.4GHz, and even that thing is running fine with smart connect and "AX" enabled.

Well in your case, I suppose there isn't a direct UI option to Enable smart connect, and disable HE Frames only for 2.4GHz. Maybe next black friday day swap that soundbar out

hahaha

I have a C9 but don't use the wifi because it wasn't great. The built in Ethernet was 100mb so now use a USB dongle with 1GB Ethernet and works lovely with my NAS

Regarding the wifi problem I found it I disable HE frames and then connect the google speakers they work on 5Ghz. Then go back to the settings and enable it again and they stay connected for good. Even a reboot of the speaker or ASUS Router and they reconnect. Its only when you first set the speaker up it doesn't work. Because I was setting up a new Asus device I used a new SSID and Password which meant setting up all the end devices again and the google speakers don't like the HE frame enabled when doing this....very odd but I have stuff working now

I don't know if its beta code related because I never used the stock firmware for more than 1 hour

Also on the beta code does the wifi scanner work ? I can't get anything to work in any browser ?

 

[Slawek P]

The same problem on both builds.

I think you meant the asus beta, not merlin beta?

 

[Soul_]

hahaha

I have a C9 but don't use the wifi because it wasn't great. The built in Ethernet was 100mb so now use a USB dongle with 1GB Ethernet and works lovely with my NAS

Regarding the wifi problem I found it I disable HE frames and then connect the google speakers they work on 5Ghz. Then go back to the settings and enable it again and they stay connected for good. Even a reboot of the speaker or ASUS Router and they reconnect. Its only when you first set the speaker up it doesn't work. Because I was setting up a new Asus device I used a new SSID and Password which meant setting up all the end devices again and the google speakers don't like the HE frame enabled when doing this....very odd but I have stuff working now

I don't know if its beta code related because I never used the stock firmware for more than 1 hour

Also on the beta code does the wifi scanner work ? I can't get anything to work in any browser ?

That is good news!! 2 less reasons to not use smart connect now . LOL!

I think google may have updated the firmware to support the HE Frames in the meantime. The first time your tried to connect them, they didn't have it, but now they do. Is your soundbar still not working on HE Frames?

 

[beerglass007]

That is good news!! 2 less reasons to not use smart connect now . LOL!

I think google may have updated the firmware to support the HE Frames in the meantime. The first time your tried to connect them, they didn't have it, but now they do. Is your soundbar still not working on HE Frames?

Will check my sound bar tonight

 

[aex.perez]

Hi @ASUSWRT_2020, on Merlin 386 beta 3 the problem of slow Ookla tests while TrendMicro enabled remains.
Enabling any TrendMicro stuff like AiProtection or WebTraffic monitor slows down Ookla tests on AX88U router three times (for 1GB+ connections).
Disabling these features does not help, only withdrawing from Trend Micro privacy agreement restores Ookla speeds to 1GBs
TrendMicro changes somehow CPU load distribution on AX to be less favourable

Just tested this on my AX88U, same result, also on Merlin 386 Beta 3. Turning every related feature (lists them when removing from privacy agreement) off didn't get the Ookla results back. Only after withdrawing from the privacy agreement do the Ookla results return on the router close to the 1Gb fiber I'm on. Oddly enough the PC I use as my benchmark, connected on 1GbE still gets 930-950Mbps up/down. On the Router, before the withdrawl of the privacy agreement 400Mbps +/-50Mbps down and 650Mbps +/- 50Mbps up. After the withdrawl, almost the same as the PC down but 720Mbps +/- 20Mbps up - in all cases pings and jitter with little to no varience and a dozen or so samples over a few hours across the three scenarios.

CPU cores utlization in the graph looked a little different, core utilization more balanced but that could be influenced by all sorts of things...

 

[beerglass007]

Found another problem tonight. Streaming .MKV files from a NAS is fine but streaming .MP4 files pause and freeze

My machine is 1M away from AX86u and only does this on 5Ghz and not 2.4Ghz

The stream is 600Kb/s basic nothing very odd

 

[jeff3820]

Found another problem tonight. Streaming .MKV files from a NAS is fine but streaming .MP4 files pause and freeze

My machine is 1M away from AX86u and only does this on 5Ghz and not 2.4Ghz

The stream is 600Kb/s basic nothing very odd

Just curious...which NAS for you have and what clients are having issues. I have a Synology NAS and streaming to iOS devices with the Synology dsvideo app stutters. Streaming to Synology dsvideo Roku app is fine. Streaming to a web browser is fine on either device.

Make sure WMM is turned on in the Wireless professional tab for both 2.4GHz and 5GHz.

 

[Unisoft]

Found another problem tonight. Streaming .MKV files from a NAS is fine but streaming .MP4 files pause and freeze

My machine is 1M away from AX86u and only does this on 5Ghz and not 2.4Ghz

The stream is 600Kb/s basic nothing very odd

Got a Synology and was getting this but only when streaming from one AX86U to another over WAN/Internet using HTTPS.

The receiving router needed ANOTHER factory reset and setting up by hand (I had already done this so something must have got corrupted during first setup), and I used Merlin's BETA4 in the test builds folder where QOS works properly.

All is fine on my AX86U's now.

Really do make sure you have factory reset and enter settings by hand again - in my case for one router it was twice!
On lan connection it was fine on 5GHZ using ROKU Emby. my MP4 files are high profile level 4.2 and 11mbps average bit rate.

 

[RMerlin]

Beta 4 is now available. I want people to re-focus on testing it thoroughly, and this thread has derailed, so I'm starting a separate thread for it, and will be locking this one.

Beta - Asuswrt-Merlin 386.1 Beta (stage 2) is now available

Continuation of the first thread which covered beta 1 through 3. Jan 25th: Beta 5 is now available. Changes since Beta 4: 3790c8fe9a Updated documentation fd1cb1eacd dnsmasq: update to 2.84 final 6ecc2e1df5 dnsmasq: update to 2.84test3 (openssl/2.84test3-40-gee342e7) 7328ccf7d8 dnsmasq: update...

www.snbforums.com