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

[ExtremeFiretop]

Why did they only post that under Windows 7, 8, 8.1 and NOT 10? The other 2 W10 builds have it (DCH & RS5)?

Why select an OS at all when it's not related to the router?

Just select the "Bios & Firmware" tab

 

[maxbraketorque]

Here are a few observations after running 386.1b4b overnight on my AC86U:

• ...
• There is another OVPN warning about cipher usage: <WARNING: 'cipher' is present in local config but missing in remote config, local='cipher AES-128-CBC'>. However, both the server and the client have exactly the same cipher configuration. I guess there is something new with v2.5.0 vs v2.4.9 with cipher settings. I'm trying to figure it out now, but so far, its not totally clear what I should change on the client or server side.

Disabling cipher fall back on the server side running 384.19 + OVPN v2.4.9 appears to have eliminated this warning. So apparently OVPN 2.5.0 no longer supports the -cipher option.

 

[John Fitzgerald]

Why select an OS at all when it's not related to the router?

Just select the "Bios & Firmware" tab

View attachment 29377

For uniformity sake they should make it consistent. Sorry I'm not a multi billion dollar conglomerate.

 

[ExtremeFiretop]

For uniformity sake they should make it consistent. Sorry I'm not a multi billion dollar conglomerate.

Fortunately the "Bios & Firmware" tab is uniform to what I've seen, and it's okay neither am I

 

[bradbort]

I also was able to downgrade to the latest AX88U asus stock with no issues.

So what are next steps for AX88U beta? Waiting for new Asus release?
I am on beta4 and reading posts it sounds I will might have an issue with the next upgrade, perhaps even requiring rescue tool.

 

[roboots21]

Since I've moved to beta 1. (And did a factory default)

The schedule are not working anymore. My kids used to be cut out at the exact time I put in. But now internet is available 24/7 what ever I put in.

Anybody else is having issues? I've tried beta 2, 3 and now 4 !

Thank you

The time scheduling interface changes a bit, and I had a similar issue on update to betas. I removed all my time scheduling and re-added it all using the new interface and all is working now (beta1 through beta4).

 

[FelixGV]

The time scheduling interface changes a bit, and I had a similar issue on update to betas. I removed all my time scheduling and re-added it all using the new interface and all is working now (beta1 through beta4).

The aiprotection time scheduling only block web access but not other internet services: netflix,

The time scheduling interface changes a bit, and I had a similar issue on update to betas. I removed all my time scheduling and re-added it all using the new interface and all is working now (beta1 through beta4).

The aiprotection time scheduling only blocks access to web pages from browsers but it doesn't block other services such as netflix or spotify.

 

[FelixGV]

If you want to control internet access, you only can do what I do now: turn off the router when you don't want your kids use the internet. Aiprotection is just a parcial parental access control.

 

[RMerlin]

look like PT time scheduling has been fixed in V 3.0.0.4.386.41634 . AC86U. Can you look into it?

No I can't.

 

[aex.perez]

Kinda interesting, maybe water under the bridge at this point with beta 4/4b available...

With the performance issues on the AX88U with Beta 4 and Asus 386 RC10, I wanted to do some bandwidth testing with Beta 3, with and without TrendMicro related features before updating to B4b on the AX88U and B4 on the AC5300 Aimesh nodes.

AX88U HW 1.1 - AT&T Fiber 1Gb - Skynet, scribe/uiscribe, ntpMerlin, spdMerlin, conmon, scMerlin
Manually selected Wan Port/Preferred server to minimize variences, same server used for Windows Speedtest App
Tested with minimal devices/traffic over the LAN/Network. CPU's in the 0%-2% range before each run. Fan/lots of air flow from mini tower fan on high/active cooling on the AX88 as well, 55c-to 57c, 25.5c/78f ambient on each run. 10 spdMerlin Ookla Speedtests samples/runs for each of the scenarios, 2-3 minutes in between each, to get an average, feature on, then off, by itself and in concert with the other features, took awhile to do to make sure to try minimize self-induced variances. Ran a few with Asus provided Speedtest as a control, no variance to spdMerlin results.

In all cases, Windows 10 with Speedtest App connected 1GbE to the AX88U, ~940 Up/~940 Down with little variance

About equal to ISP provisioned down and close to what's seen on PC, short on upload by a 1/3, with or without TrendMicro Privacy on its own (no features enabled).
386.1 B3 - spdMerlin 4.1.1
Trend Micro Privacy withdrawn
~910 Down/~600 Up

386.1 B3 - spdMerlin 4.1.1
AIProtection Enabled, TrendMicro Privacy Accepted, AiProtection Disabled
Same ~910 Down/~600 Up

With TrendMicro feature, a different story...
386.1 B3 - spdMerlin 4.1.1
Several scenarios tested, sames results. Enable Either "QoS - WAN/LAN Bandwidth Monitor", "Web History", "Two-Way IPS", or "Infected Device Prevention and Blocking" individually or all together
~450 Down/ ~450 up* Note+/- 50Mbps)
Oddly consistent regardless of what Trend Micro feature(s) is/are enabled, CPU penalty?

Did not test if the drop happens with Adaptive QoS, Game boost or Malicious Sites Blocking together or individually. I don't use those features and the testing I did was pretty late at night to minimize traffic variables as much as I could, and I was beginning to nod off.

After this I'm wondering whether we are starting to see the limits of the CPU's on the router when it comes to bandwidth tests at higher provisioned internet bandwidth?

Once I introduced any Trend Micro enabled feature, bandwdith test results dropped in half. Wondering because all the times during these tests, the PC consistently had no impact in terms of its bandwidth results. The other thing that has me curious is the disparity of the down/up rates which had no variance on the PC, but the router reports a 1/3 less bandwidth than is provisioned on the upload test, 1000 synchronus. Results flipped when a TrendMicro Privacy required feature was enabled. With a TrendMicro feature enabled the upload would consistently out perform the download by ~50Mbps to ~100Mbps.

Worthy of note, applying priority to routers connection on the AT&T BGW210 had no bearing on any of the the results, of course the router is the only device on it so really didn't expect it to either, but had to rule it out. Waited the two hours for it to disable to complete testing performance and to make sure there wasn't any change in the results.

How this would apply to any QoS mechanism that would configure itself based on the Speedtest results vs manually what you are provisioned for vs Router CPU impact is for another time an place/thread...

While not completely scientific I'll agree, its as close as I can get to it with what I have at hand...

Next up Merlin 386.1 B4b, Beta 3 upgrade to 386.1 B4b on the AX88U and 4B on AImesh nodes (Ethernet backhaul) 2xAC5300, wait for it things to clear up and stabilize, then repeat this again tonight when the wife isn't binging on her cloud DVR recorded telenovelas and compare results...

 

[ExtremeFiretop]

The aiprotection time scheduling only block web access but not other internet services: netflix,

The aiprotection time scheduling only blocks access to web pages from browsers but it doesn't block other services such as netflix or spotify.

Blocks Netflix and Spotify fine for me here on Beta 4B on the AX88U. Just tested it myself on 2 TV's and a Google Nest smart speaker, none of them have browsers, all of them were blocked on their native app.

 

[maxbraketorque]

One other comment on 386.1b4b is that the built-in speed test option doesn't show full internet download speed when running it from my AC86U. I only got 584 mbps download, but I have gigabit service. If I run the Ookla speed test app from my wired Mac, I get 930 mbps which is typical for what I've seen over the past six months. I don't think the AC86U has the horsepower to measure full speed, at least with how I have it configured. I've seen some posts suggesting that fully disabling all Trend Micro related features and then opting out of the terms of service then allows the router to measure full speed, but I'm not going to try that.

 

[Makaveli]

Why select an OS at all when it's not related to the router?

Just select the "Bios & Firmware" tab

View attachment 29377

This has always been how the asus site is. When I go for bios updates to my x570 motherboard I just go straight to Bios & Firmware tab.

 

[dev_null]

The aiprotection time scheduling only block web access but not other internet services: netflix,

The aiprotection time scheduling only blocks access to web pages from browsers but it doesn't block other services such as netflix or spotify.

Blocks Netflix and Spotify fine for me here on Beta 4B on the AX88U. Just tested it myself on 2 TV's and a Google Nest smart speaker, none of them have browsers, all of them were blocked on their native app.

IIRC it has been reported (elsewhere) that existing connectivity is allowed to continue when the rules kick in, at least until there was a change in that connectivity.

If that is the case, then a streaming movie or game may continue for a period of time, depending on caching but should eventually be terminated.

New streams would not initiate if that's the case.

I can't find the reference, but seem to recall that issue from the days when I restricted the kids' devices at bedtime.

 

[PavelBab]

If you intend to use extensions, then you need to create a config file that defines those extensions.
[/QUOTE]

I have installed the following modules:
input-core, hid, usbhid
Installation is provided by init-start script.
What else can I do for it?

 

[ExtremeFiretop]

IIRC it has been reported (elsewhere) that existing connectivity is allowed to continue when the rules kick in, at least until there was a change in that connectivity.

If that is the case, then a streaming movie or game may continue for a period of time, depending on caching but should eventually be terminated.

New streams would not initiate if that's the case.

I can't find the reference, but seem to recall that issue from the days when I restricted the kids' devices at bedtime.

That would explain why it blocked everything fine for me. I didn't have any active streams in my test, I blocked and then started the TV's on Netflix and the speaker on Spotify.
All of which were successfully blocked using parental controls.

I don't have kids, so it's not an option I would ever use, but figured since we are testing firmware I may was well try to replicate the issue and couldn't.

 

[roboots21]

The aiprotection time scheduling only block web access but not other internet services: netflix,

The aiprotection time scheduling only blocks access to web pages from browsers but it doesn't block other services such as netflix or spotify.

When I set Time Scheduling to on and set time filters, it turns off all access to anything to the nodes I specified. I use it for RokuTV as well as consoles (PS/XBox), to which it disables them accordingly when I set it, even including access on Chromebooks. Although, I've never checked on the Chromebook to see if it specifically blocks Netflix or the like, so perhaps I'll check that just to verify and see if this is true.

 

[webdriver]

If you want to control internet access, you only can do what I do now: turn off the router when you don't want your kids use the internet. Aiprotection is just a parcial parental access control.

agree ... sometimes even can't block browsing so have to put "block" for certain device.
streaming is totally unable to block.

 

[ExtremeFiretop]

agree ... sometimes even can't block browsing so have to put "block" for certain device.
streaming is totally unable to block.

Maybe an issue specific to the RT-AC86U? But has discussed above, could not replace on my AX88U. Did you do a factory reset on Beta 4?

 

[aex.perez]

Kinda interesting, maybe water under the bridge at this point with beta 4/4b available...

With the performance issues on the AX88U with Beta 4 and Asus 386 RC10, I wanted to do some bandwidth testing with Beta 3, with and without TrendMicro related features before updating to B4b on the AX88U and B4 on the AC5300 Aimesh nodes.

AX88U HW 1.1 - AT&T Fiber 1Gb - Skynet, scribe/uiscribe, ntpMerlin, spdMerlin, conmon, scMerlin
Manually selected Wan Port/Preferred server to minimize variences, same server used for Windows Speedtest App
Tested with minimal devices/traffic over the LAN/Network. CPU's in the 0%-2% range before each run. Fan/lots of air flow from mini tower fan on high/active cooling on the AX88 as well, 55c-to 57c, 25.5c/78f ambient on each run. 10 spdMerlin Ookla Speedtests samples/runs for each of the scenarios, 2-3 minutes in between each, to get an average, feature on, then off, by itself and in concert with the other features, took awhile to do to make sure to try minimize self-induced variances. Ran a few with Asus provided Speedtest as a control, no variance to spdMerlin results.

In all cases, Windows 10 with Speedtest App connected 1GbE to the AX88U, ~940 Up/~940 Down with little variance

About equal to ISP provisioned down and close to what's seen on PC, short on upload by a 1/3, with or without TrendMicro Privacy on its own (no features enabled).
386.1 B3 - spdMerlin 4.1.1
Trend Micro Privacy withdrawn
~910 Down/~600 Up

386.1 B3 - spdMerlin 4.1.1
AIProtection Enabled, TrendMicro Privacy Accepted, AiProtection Disabled
Same ~910 Down/~600 Up

With TrendMicro feature, a different story...
386.1 B3 - spdMerlin 4.1.1
Several scenarios tested, sames results. Enable Either "QoS - WAN/LAN Bandwidth Monitor", "Web History", "Two-Way IPS", or "Infected Device Prevention and Blocking" individually or all together
~450 Down/ ~450 up* Note +/- 50Mbps)
Oddly consistent regardless of what Trend Micro feature(s) is/are enabled, CPU penalty?

Did not test if the drop happens with Adaptive QoS, Game boost or Malicious Sites Blocking together or individually. I don't use those features and the testing I did was pretty late at night to minimize traffic variables as much as I could, and I was beginning to nod off.

After this I'm wondering whether we are starting to see the limits of the CPU's on the router when it comes to bandwidth tests at higher provisioned internet bandwidth?

Once I introduced any Trend Micro enabled feature, bandwdith test results dropped in half. Wondering because all the times during these tests, the PC consistently had no impact in terms of its bandwidth results. The other thing that has me curious is the disparity of the down/up rates which had no variance on the PC, but the router reports a 1/3 less bandwidth than is provisioned on the upload test, 1000 synchronus. Results flipped when a TrendMicro Privacy required feature was enabled. With a TrendMicro feature enabled the upload would consistently out perform the download by ~50Mbps to ~100Mbps.

Worthy of note, applying priority to routers connection on the AT&T BGW210 had no bearing on any of the the results, of course the router is the only device on it so really didn't expect it to either, but had to rule it out. Waited the two hours for it to disable to complete testing performance and to make sure there wasn't any change in the results.

How this would apply to any QoS mechanism that would configure itself based on the Speedtest results vs manually what you are provisioned for vs Router CPU impact is for another time an place/thread...

While not completely scientific I'll agree, its as close as I can get to it with what I have at hand...

Next up Merlin 386.1 B4b, Beta 3 upgrade to 386.1 B4b on the AX88U and 4B on AImesh nodes (Ethernet backhaul) 2xAC5300, wait for it things to clear up and stabilize, then repeat this again tonight when the wife isn't binging on her cloud DVR recorded telenovelas and compare results...

Ok, took advantage of the wife going for a walk. Upgraded to B4b on the AX88U and B4 on the AC5300's - no issues. spdMerlin results so far, no TrendMicro related features enabled, Privacy withdrawn, same results with PC (~930Mbps in both directions) Router (~900Mbps down, ~600Mbps upload). More testing tonight.