[Beta 384/NG] Asuswrt-Merlin 384.3 Beta is now available

[tommyv2]

New AC86U owner here, using this firmware. I have all wireless disabled (using Ubiquiti because reasons), but everything is working well! Thanks, Merlin!

 

[SMS786]

Just like the previous alpha, a very solid and reliable build thus far on my 68u.

Thank you @RMerlin!

 

[NGRhodes]

Okay, an update:

I tried Asus stock firmware (RT-AC68U_3.0.0.4_384_20308-gead790e), which has the exact same issue: QoS is not starting. No matter what I try, stats aren't loaded, traffic sorting containers remain empty. It's just not functioning.

I really don't get it anymore, as I have seen several people with an RT-AC68U with firmware from the 382 and 384 branch reporting QoS is just working fine.

So I reverted to 380.69_2. QoS works practically immediately. I did nothing different (always follow the exact same procedure when doing an update - or - downgrade in this case), but after going through QiS, I enabled Qos and bam: it just works. No sure what causes this, but something has changed. At least for me something has changed.

@RMerlin: Could this be hardware revision (E1 in my case) specificic? For now I'll stick with 380.69_2 which works great as well, I can only hope that SkyNet, AB-Solution and dnscrypt will stay working.

I got new a new AC66U_B1 this weekend and upgraded (from the stock 380) to 3.0.0.4.384.20308 and Adaptive QoS, app analysis all working fine. I did notice some display issues with layout, but clearing my browser cache solved.

Just wondering if its a browser your suffering ?

 

[M@rco]

I got new a new AC66U_B1 this weekend and upgraded (from the stock 380) to 3.0.0.4.384.20308 and Adaptive QoS, app analysis all working fine. I did notice some display issues with layout, but clearing my browser cache solved.

Just wondering if its a browser your suffering ?

Thanks for thinking with me. Thought of that too, when trouble started, tried three browsers, all the same. The fact that tc gives no output proves it's 'simply' not working. What makes is it hard (and terribly frustrating) is that it apparently works for most, if not all, RT-AC68U users.

 

[skeal]

Thanks for thinking with me. Thought of that too, when trouble started, tried three browsers, all the same. The fact that tc gives no output proves it's 'simply' not working. What makes is it hard (and terribly frustrating) is that it apparently works for most, if not all, RT-AC68U users.

This is probably a stupid question but have you tried:

services restart_qos

If so what does it basically say in the logs that it is already running or....?

 

[RMerlin]

@RMerlin: Could this be hardware revision (E1 in my case) specificic? For now I'll stick with 380.69_2 which works great as well, I can only hope that SkyNet, AB-Solution and dnscrypt will stay working.

Doubt it, but the whole Trend Micro code being closed source, it's impossible for me to tell for sure.

 

[h3lmut]

runs very well on my ac56u. qos is working fine. thank you !!

 

[M@rco]

This is probably a stupid question but have you tried:

services restart_qos

If so what does it basically say in the logs that it is already running or....?

I haven't tried it manually, it's done automatically when toggling on/off. It doesn't say a thing. For now I'm back to 380.69_2 which runs smooth like butter. About to install FreshJr's script, I'm all good. Thanks though for trying to help me figuring this out.

 

[tyreman]

Working okay here

 

[ewokuk]

For a phone, that is not too bad.

Anyway, Merline mentioned it somewhere else. Asus is releasing too many versions of each model too soon. They are also increasing number of closed source binaries. Which makes it harder to move into any new codebase as it leaves some users stranded in a old codebase in case you move on. And also makes it harder to track down any new bug due to the closed source nature of some parts. So he requested users to be patient about incorporating any new versions.

This is why I think I am favouring the R7800. I believe it is more open source anyway as it is fully supported by openwrt/lede and the ac86 doesnt seem to be supported at all by it. I like the interface on my AC68u and use Merlin firmware on that so the AC86 was the logical upgrade but I don't like the direction asus seem to be going in with more closed parts, and I really want wireguard support which might never come to the ac86. Having said that, I expect the interface in openwrt/LEDE is probably not nearly as good.

 

[JohnD5000]

Special characters have been causing random issues for years, be it in SSIDs, device names or passwords. Just don't try anything fancy and stick to alphanumeric.

Is a space " " considered a special character? What about dash "-" and underscore "_"? Thanks

 

[drboss]

Thanks for fast answer, but @RMerlin why all version 380.xx work fine with USB modem at every time after upgrade to 382 or 384 (on clear, after factory reset too) they stop works. Any difference in dmesg is no section for inicialize USB modem, no ttyUSB0,1,2 devices in /dev and in SYSLOG 0x12d1 product=0x1506 invalid for parameter 'vendor'. When I downgrade to 380.69... all work fine.

@RMerlin I now that You do very big work for as, but do You thinks to correct problems with dual WAN if second WAN is USB LTE modem?
After version 380.69_2 I have two problems (this version work fine) but in 382+ I receive Syslog message like: `0x12d1 product=0x1506' invalid for parameter `vendor'
1. the Huwai E3372 isn't in data, after start AC68 I need add the: insmod option vendor=0x12d1 product=0x1506, some time this work with service restart_wan
2. but they not work all the time, sometime I need do reset usb modem (by usbreset 12d1:1506, or at^reset, , but certain time they result of change mount point of second usb
3. many time when I go over inicialization of modem, they start work fine, but after few minut I have problem with routing (ex. chain FUPNP not found)

 

[routers]

Off-topic, sorry, I just want to say that I'm grateful that support for the RT-AC56U has survived 'till this point (and beyond), by both ASUS and Merlin. May it be supported forever My RT-AC86U arrives tomorrow.

Edit: I wonder if it's possible that in the future bashrc could be implemented like nanorc? Commands like ALIAS and EXPORT would be very useful (to me at least).

 

[martywd]

Edit: I wonder if it's possible that in the future bashrc could be implemented like nanorc? Commands like ALIAS and EXPORT would be very useful (to me at least).

Custom aliases can be added to the '/jffs/configs/profile.add' file. Not sure, but maybe exports too?

.

 

[hifiwifi]

Solid on RT-AC1900P. Had VPN issue with Alpha 3. Haven't tried to replicate yet.

 

[john9527]

Custom aliases can be added to the '/jffs/configs/profile.add' file. Not sure, but maybe exports too?

.

Yes, exports too....

 

[RMerlin]

Is a space " " considered a special character? What about dash "-" and underscore "_"? Thanks

Those should be fine, as they carry no special meaning in Javascript. Most dangerous characters are quotes, double quotes, dollar signs and percents.

Thanks for fast answer, but @RMerlin why all version 380.xx work fine with USB modem at every time after upgrade to 382 or 384 (on clear, after factory reset too) they stop works. Any difference in dmesg is no section for inicialize USB modem, no ttyUSB0,1,2 devices in /dev and in SYSLOG 0x12d1 product=0x1506 invalid for parameter 'vendor'. When I downgrade to 380.69... all work fine.

I don't know. I don't touch the USB modem code at all, which means I've never even tried to study it either.

Off-topic, sorry, I just want to say that I'm grateful that support for the RT-AC56U has survived 'till this point (and beyond), by both ASUS and Merlin. May it be supported forever My RT-AC86U arrives tomorrow.

Edit: I wonder if it's possible that in the future bashrc could be implemented like nanorc? Commands like ALIAS and EXPORT would be very useful (to me at least).

Asuswrt uses dash, not bash.

 

[feranick]

Asuswrt uses dash, not bash.

As previously suggested, dash has support for a ".profile" file, so the original poster suggestion to include it in /jffs/configs may still have merit... When saved in /tmp/home/root it doesn't survive a reboot. Yet, I create one all the time (copy and paste) after reboot. Not ideal.

 

[Jack Yaz]

As previously suggested, dash has support for a ".profile" file, so the original poster suggestion to include it in /jffs/configs may still have merit... When saved in /tmp/home/root it doesn't survive a reboot. Yet, I create one all the time (copy and paste) after reboot. Not ideal.

Profile.add in /jffs/configs

 

[RMerlin]

As previously suggested, dash has support for a ".profile" file, so the original poster suggestion to include it in /jffs/configs may still have merit... When saved in /tmp/home/root it doesn't survive a reboot. Yet, I create one all the time (copy and paste) after reboot. Not ideal.

profile.add is already supported.