Beta Asuswrt-Merlin 388.1 Beta is available for select models

[capncybo]

I understand this release is still a Beta & as such (support & more-over support of plugins) is not provided.
But I'll just share this for the group.
I had dirty flashed my RT-AX86U from the previous version to the current.
Mostly uneventful, even removed the previous wireguard plugin to test the new integrated wireguard.
But as I tinkered & observed how it would disable my HW Flow Cache, It thought...
Hmmm without flow cache acceleration then I probably can uninstall the FlexQOS plug-in also & retry CAKE.
Perhaps for even less buffer-bloat than my previous waveform A/A rating.
Note: My WAN IP connection is Shaw 300/15 & I can often achieve about 325/15
But as I tested Cake & then tried to re-enable FlexQOS...
And yesss I even tried a couple-reboots.
My Download Speed was severely limited.
It was about the same as my upload speed of around 15Mbps.
Please understand my testing window was severely limited as this was my households main connection.
With 4 other people in the family... I have to act quickly re network disruptions.
Hence I quickly disabled QOS & seemed to have avoided the Family LYNCH-Mob.
Anyways, I'd like to tinker more but...
It's difficult when it's not really a test environment.

 

[L&LD]

Did you possibly transpose the upload and download?

 

[Jaime Alvarez]

Dirty upgrade to 388.1_beta1 from 386.7_2 on my AX86U.

First beta version I've ever installed. Everything seems normal so far.

WiFi performs better than with 386.7_2. There is a corner of my apartment where I had a very weak signal, and there my phone would connect and disconnect all the time. Now I can stay on the WiFi continuously at that same spot.

In general everything seems to work fine.

 

[capncybo]

Did you possibly transpose the upload and download?

Oh for F___'s sake that's totally what I did. And In-fact... I've also done this to myself at least once before.
When most of us discuss network speeds I'm rather used to putting Download speed first & Upload second.
I wonder if anyone else does this too?
And Obviously we should always slow down & read but since I've at least done this myself twice...
Would It be too much to (ask/suggest)...
@RMerlin Could those fields two be flipped or re-ordered?
Or do you think it would likely cause more user problems than it would prevent.
And Obviously... I'm seriously overdue for some reading glasses
Thanks again, both of you.

 

[kernol]

Oh for F___'s sake that's totally what I did. And In-fact... I've also done this to myself at least once before.
When most of us discuss network speeds I'm rather used to putting Download speed first & Upload second.
I wonder if anyone else does this too?
And Obviously we should always slow down & read but since I've at least done this myself twice...
Would It be too much to (ask/suggest)...
@RMerlin Could those fields two be flipped or re-ordered?
Or do you think it would likely cause more user problems than it would prevent.
And Obviously... I'm seriously overdue for some reading glasses
Thanks again, both of you.

Stick to vanilla Cake-Qos on your RT-AX86U like I do - and set your speeds at 15 up and 300 down [don't throttle them - Cake will do that for you]. Throwing FlexQos into the mix will not add anything = more likely detract. I get A+ bufferbloat and near full speeds from this test URL https://www.waveform.com/tools/bufferbloat

 

[RMerlin]

@RMerlin Could those fields two be flipped or re-ordered?

No. Because if 0.005% of users are making the mistake, swapping it means the other 99.995% of users who are used to the current order that has existed for 10+ years will now start making mistakes.

 

[capncybo]

Stick to vanilla Cake-Qos on your RT-AX86U like I do - and set your speeds at 15 up and 300 down [don't throttle them - Cake will do that for you]. Throwing FlexQos into the mix will not add anything = more likely detract. I get A+ bufferbloat and near full speeds from this test URL https://www.waveform.com/tools/bufferbloat

While I do appreciate the advice.
The entire reason I went the FlexQOS route previously was to prevent (Hardware Acceleration's) Flow Cache from being disabled. (I kinda thought that was FlexQOS niche)
The rational was... I didn't want to kneecap any higher-speed LAN to LAN data-transfers. (In otherwards limit their Max to say 350-500)
And to be honest FlexQOS did seem to work fairly well IMO.
BUT...
Recently, If I enable Wireguard & am going to loose (Hardware Acceleration's) Flow Cache anyways... I thought, Let me see if CAKE can do better.
And perhaps, It might...

 

[capncybo]

No. Because if 0.005% of users are making the mistake, swapping it means the other 99.995% of users who are used to the current order that has existed for 10+ years will now start making mistakes.

LOL, okay Boss (I kinda thought you might say that).
This mistake was definitely mine... I'll try to be more careful in the future.

 

[Nemesis8]

Hi to all, @RMerlin your do a great work, thanks for that, i try the firmware and all work fine but have only one problem i try wireguard server and is too fast, but when i try acces to smb or media servers i cant, but if try remotedesktop i can and work great, you know the problem?

 

[ZebMcKayhan]

What I don't know is how/if you can add a MAC Ipset to Policy Rules in VPN Director,

Nope, mac is not known to ip (as in "ip rule..."), it would require fw marks which @RMerlin already stated not possible since ai-protect is using all fw-mark bits so possible conflict.
Policy routing with dynamic ipv6 and SLAAC with privacy extension is a pain (read challenging). I dont know any other way then involve the mac-address which means firewall. But if fw-marks could not be used, perhaps TOS can?

ASUSWRT-Merlin RT-AC86U 386.7_0 Wed Jun 22 18:49:26 UTC 2022
admin@RT-AC86U-D7D8:/tmp/home/root# ip -6 rule add from all tos 0x04 lookup 121 prio 9800
admin@RT-AC86U-D7D8:/tmp/home/root# ip -6 rule
0:      from all lookup local
9800:   from all tos 0x04 lookup 121
9900:   from aaff:a37f:fa75:1::/64 fwmark 0x8000 lookup main
9910:   from all to aaff:a37f:fa75:1::1/48 lookup main
9911:   from aaff:a37f:fa75:1::1/64 lookup 121
9921:   from aaff:a37f:fa75:6::1/64 lookup 122
9991:   from all fwmark 0x1000/0x1000 lookup 121
32766:  from all lookup main
admin@RT-AC86U-D7D8:/tmp/home/root#

seems possible but perhaps some other conflict with other parts of the firmware?

Then ofcource the next problem. To handle devices that randomizes mac address in a non-persistant way (if they exist)...

 

[skeal]

@RMerlin Is this a page formatting error? My long description messes with the columns and stuff.

 

[RMerlin]

@RMerlin Is this a page formatting error? My long description messes with the columns and stuff.

No, it`s that issue noted in post #2:

VPN Director layout issues when a WireGuard client is running (Fixed in beta 2)

There's an extra </tr> tag in the HTML code when a WG client is enabled.

 

[samuraijmb]

Hey,
I updated my AX88U to 388.1-beta and all has gone smoothly.
I signed up to SurfShark and configured Wireguard for a couple of clients. They were all working fine for a few days and then my phone and a Windows 11 PC in the house lost internet connection. After restarting the Wireguard client it all connected again.

I had read with Asus Official firmware that there was a bug fixed related to Surf Shark and Wireguard and I wondered if that is something that is in the Merlin build?

Surf Shark Wireguard Issue

I have gone back to OpenVPN Client for now.

Cheers

 

[SR-G]

Hello,
Cool, a new firmware update !
But ... about the selected models, does it mean that some model like the "RT-AC88U" (that i own since a couple of years) are not going to be supported anymore ??! Or is there a chance for them to also receive that update in the future ?

 

[skeal]

Hello,
Cool, a new firmware update !
But ... about the selected models, does it mean that some model like the "RT-AC88U" (that i own since a couple of years) are not going to be supported anymore ??! Or is there a chance for them to also receive that update in the future ?

The AC routers will stay on the 386 code branch and won't be seeing Wireguard.

 

[snah001]

Will there in the future be support for the AX53U?
Just curious since I bought it a week ago as extender.

 

[JVC]

I hope a Kill Switch is implemented for WireGuard.

 

[skeal]

I hope a Kill Switch is implemented for WireGuard.

This has been talked about at length. Search this thread for references of kill switch. Both spelled as one word and two words.

 

[zer0bitz]

I just updated to this beta firmware and everything went succesfully. I got wireguard working as a server and client. Only issue im having is that DDNS wont update. I keep getting this error in my system log. I tried to manually disable ddns and start it again but that didnt work. What should I do?

Nov 11 14:40:00 RT-AX56U-CCD8 Let's_Encrypt: Err, DDNS update failed.

 

[S.claus]

Hello. I've installed the latest beta and so far it's been a smooth experience besides one minor inconvenience.
How do you route IPv4 and IPv6 over the Wireguard protocol at the same time?