Release Asuswrt-Merlin 386.12 is now available for AC models

[Tech9]

The signature version has nothing to do with it

What is it? I see it on stock Asuswrt now. It wasn't like this before. I had an RT-AC66U B1 not long ago on the same stock firmware and it was working fine for months. Now I got one RT-AC1900P to play with and dcd is crashing constantly. Something was updated/changed automatically. ASD again?

 

[RMerlin]

What is it? I see it on stock Asuswrt now. It wasn't like this before. I had an RT-AC66U B1 not long ago on the same stock firmware and it was working fine for months. Now I got one RT-AC1900P to play with and dcd is crashing constantly. Something was updated/changed automatically. ASD again?

No, it's not something that was pushed to routers that triggered it. The initial diagnosis was a certificate that had to be replaced (either expired or failing minimal key strenght requirement would be my guess), and replacing it did resolve the crashes both for me and the engineer working on this issue. So why it's still crashing for some, I have no idea. What the details are, only Trend Micro knows (and to be honest, I'm not sure even themselves know, seeing how highly unstable dcd has shown to be over the years).

 

[KrypteX]

No, it's not something that was pushed to routers that triggered it. The initial diagnosis was a certificate that had to be replaced (either expired or failing minimal key strenght requirement would be my guess), and replacing it did resolve the crashes both for me and the engineer working on this issue. So why it's still crashing for some, I have no idea. What the details are, only Trend Micro knows (and to be honest, I'm not sure even themselves know, seeing how highly unstable dcd has shown to be over the years).

Very strange, indeed. Do you know whether this occurs on all AC devices or only on some of them? Also, has this ever been an issue on AX routers? (haven't been following AX model reports lately, hence the question)

 

[Tech9]

Hopefully they fix it soon because I’ve noticed the crashes impact the overall experience. It’s not something in logs only. CPU cores jump up and down and whatever is running there slows down too.

 

[langesgnu]

installed the testing firmware on my AC88U today and tried it now for 2 hours, the process is still crashing approximately every 2 minutes, while in between I had also 20 minutes where it did not crash

 

[RMerlin]

Do you know whether this occurs on all AC devices or only on some of them?

So far it seems specific to these older AC models.

installed the testing firmware on my AC88U today and tried it now for 2 hours, the process is still crashing approximately every 2 minutes, while in between I had also 20 minutes where it did not crash

I currently have it running in verbose logging mode on my test RT-AC66U_B1. What I noticed is that it seems to go quiet at some point, at which point it usually crashed in the past. Seems like in some cases it flat out quits with a crash, or in some cases (like mine) the process seem to just silently get stuck.

 

[RMerlin]

Asus are going to try a different approach at working around the issue. I'll try to get at least an RT-AC68U test build ready before heading to bed, and will compile the other ones tomorrow. For now I'll remove the current test builds to avoid any confusion.

 

[KrypteX]

Asus are going to try a different approach at working around the issue. I'll try to get at least an RT-AC68U test build ready before heading to bed, and will compile the other ones tomorrow.

Thank you! Has anyone tried the new test build RT-AC68U_386.12_1-g0251ee445a.trx ?

 

[RMerlin]

I uploaded the other models.

 

[RocketJSquirrel]

Thank you! Has anyone tried the new test build RT-AC68U_386.12_1-g0251ee445a.trx ?

I've been running it for just over an hour and it seems fine. No crashes or mysterious log entries.

 

[RMerlin]

Thanks.

When testing these new builds, also please confirm that the Trend Micro features (Adaptive QoS, Traffic Analyzer, etc...) are still working properly.

 

[Calkulin]

I updated a AC88U about 4 hours ago and no crashing anymore, so I just updated a AC68U to test that one. I'll play around with the TM features to see if I notice anything

 

[Tireless Hyena]

Ok, I have this problem with 4 ac68u on 4 remote sites in version 386.7_2.
What can I do, knowing that I have 3 Openvpn clients on each (with no NVRAM problems despite a large number of DHCP reservations directly done via nvram tools) and that since release 386.11 "CHANGED: Reduce max OpenVPN clients to 2 for RT-AC68U and DSL-AC68U due to lack of NVRAM on these two models"?
Does anyone have any ideas?
Thanks in advance

 

[RMerlin]

Does anyone have any ideas?

Either upgrade the router, or manually configure your VPN clients through scripting. Which is probably not worth the effort for a 11 years old router, to be honest...

 

[Squall Leonhart]

This log output has stopped, i have not installed the new test firmware.

I blame October.

 

[Woody]

I uploaded the other models.

@RMerlin this is astonishing..

I have been watching this conversation for some time, because on an RT-AC86U, from 386.10, I have been getting (initially) an issue where at a certain moment (after about two weeks of uptime) the router would refuse any further connections on 2.4GHz. The "fix" was to power cycle the router (literally pull the power out). [I know that others have said to cycle 2.4Ghz on/off and/or a reboot from the UI - but that just didn't work for me]

I started to get more insight into this issue, as I migrated through 386.11 and 386.12 - because I discovered that this was not just a 2.4Ghz issue. It was as if, at the point at which 2.4Ghz wouldn't accept any new connections (it had about 30 max - and existing connections remained fine), I realised, that it wasn't accepting any new connections - via ethernet or Wifi. So I kind of got to thinking that maybe it wasn't a WiFi issue, just simply that the router wouldn't accept new connections OR would not provide a DHCP lease (I couldn't tell from the logs which).

So I upgrade all the way to 386.12 (when it was released) and the situation got progressively worse - to the point at which the router becomes completely and utterly unusable - doesn't accept wifi or any ethernet connections (windows marks ethernet as down - so I couldn't even connect to look at the logs). It has been this way for months (since 386.12 was released) - necessitating a full power-cycle every 6-7 days or so.

To be clear, this is stock, no tools, no scripts, no VPN. The ONLY config change I made, was on the DHCP side, to block just one IOT device from having internet access.

So today, I install RT-AC86U_386.12_1-g0251ee445a_ubi.w - because even being alpha, anything is better than what I have been wrestling with - right?

Obviously, it is early days - we'll know if this is more stable in a week or two, but why I call this astonishing is this:

I am in a small village in the countryside in the UK. The theoretical max we have in the village is 80Mbps up/20Mbps down FTTC - and my personal max (down the copper phone line approx 0.5 KM from the FTTC street cabinet) is 67Mbps - so any extra bandwidth I can wring out of this link, is always a bonus. Imagine my surprise with this alpha build when I find that:

1. My download speed has increased to 74Mbps.

2. Free memory has gone from 96MB to 226MB.

3. I haven't seen CPU go much over 10% on either core - and it is mostly tinkling along at around 3%

4. Temp is kind of unchanged (in a 20C environment) at around 86C

I just want to say a BIG thank you - loving your work!

 

[KrypteX]

@Woody The improvements you are seeing with the latest 386.12_1 alpha are probably due to one of these updates/fixes from https://github.com/RMerl/asuswrt-merlin.ng/blob/386.12_x/Changelog-386.txt

386.12_2 (xx-xxx-2023)
- UPDATED: openssl to 1.1.1w.
- UPDATED: curl to 8.4.0.
- FIXED: WPS not working on SDK6/SDK7 devices (affecting RT-AC68U and RT-AC88U/3100/5300)
- FIXED: dcd constantly crashing (updated Trend Micro components)

 

[langesgnu]

Asus are going to try a different approach at working around the issue. I'll try to get at least an RT-AC68U test build ready before heading to bed, and will compile the other ones tomorrow. For now I'll remove the current test builds to avoid any confusion.

Installed the updated testing firmware yesterday and had the AC88U now run on it until noon. The dcd process did not crash anymore, but asd did crash immediately after enabling the TrendMicro things again.

Furthermore, the whole system is getting terribly slow for wifi clients when QoS / TrendMicro is enabled, but this something 386.12 always had. After pulling out of the privacy agreement, the performance is again normal and I can properly work on my wifi clients.

 

[Calkulin]

I've had ASD crash once on all 6 routers I updated following a reboot and 1 router did crash DCD 4 times back to back but hasn't done it again since

 

[RocketJSquirrel]

When testing these new builds, also please confirm that the Trend Micro features (Adaptive QoS, Traffic Analyzer, etc...) are still working properly.

Uptime now 22 hours without incident. dcd remains well behaved.

EDIT: I found this from 10 minutes ago in my log:
Nov 8 09:36:43 kernel: asd/5924: potentially unexpected fatal signal 11.

The TM features appear to be working. I can turn them off and on and they respond. I'm not sure how to test them for functionality, though.