[Preview] Asuswrt-Merlin 384.3 pre-release test builds

[XIII]

Hi @RMerlin , just curious to know where can we see the commits for 384.3. I don’t see the source in github.

Thanks.

Maybe this part of the release notes helps?

384.3 Beta (xx-xxx-xxxx)
   - NOTE: To reduce confusion following the version
           bump to 384, the current Github repository
           was renamed from asuswrt-merlin.382 to
           asuswrt-merlin.ng (for New Generation).
           It's recommended that you update your
           local repository if you're a developer,
           for example by running:

              git remote set-url origin \
                 [email protected]:RMerl/asuswrt-merlin.ng.git

 

[XIII]

New test builds should appear in the next few minutes on https://asuswrt.lostrealm.ca/test-builds containing security fixes.

ASUS patches root command execution flaws haunting over a dozen router models

 

[RMerlin]

Just out of curiousity: why is the third alpha named alpha2 (like the previoud one) instead of alpha3?

Because it's not an official release. I only bump alpha versions when I hit certain development milestones. For instance, I generally bump to a new alpha version once I finish merging a new GPL, or when I finish adding a new feature. Right now, the alpha 2 stage is focused on debugging new issues introduced by the GPL merge. Once that stage is done, I will bump the version.

 

[Gregory Phillips]

Since asus released new firmware today will these updates find their way into this build?

 

[scjr]

Since asus released new firmware today will these updates find their way into this build?

Hi Gregory,

They should, as soon as Asus gives Merlin the source code and barring anything is unusual in the code. Sometimes it takes Asus awhile to release it.

 

[Reinvented]

AC3100, seems to work okay. Fixed an issue that's been plaguing my router for a while which was not showing any traffic on wireless 2.4/5 in Traffic Analyzer. Memory usage is a lot lower too, which is nice!

QoS stats don't show using Adaptive QoS with manual or automatic setting with fq_codel.
QoS performance problems, experiencing a ton more bufferbloat on a fiber 150/150 connection. Quality has gone down, with lots of slowdowns.

Reverted back to stable release, and it's stable again.

Hoping that the new firmware that was released today will make it in. The patch notes say that it fixed a lot of these problems, which hopefully wouldn't be too hard to commit.

 

[applesnowleo]

The 5300 firmware is giving an error, says firmware is for the ac88 RT-AC5300_380.69_1-ga160108.

Nevermind... Downloaded again and it's fine.

 

[Gregory Phillips]

Memory usage is a lot lower too, which is nice!

Speaking of memory use I've noticed that if I have AiProtection enabled my ram use is 33% plus but if I turn it off it drops down to 21%.

 

[Sanna1967]

AiProtection enabled my ram use is 33% plus

Normal with the new TrendMicro engine

 

[PDinDetroit]

RT-AC3200 firmware:
-384.3_alpha2-g5b8da38

Config:
-SmartConnect with default rule
-Adaptive QOS-Automatic Bandwidth-fq_codel Queue Discipline-Cable(DOCSIS) Wan packet overhead-Web Surfing Mode

Notes (nits):
-Network Map->View List shows Tx/Rx rates as zeros for Wi-Fi clients and dashes for Wired clients, System Log->Wireless Log shows correct values
-Adaptive QoS->QoS Wan packet overhead was set to Cable (DOCSIS), but after Apply it goes back to "Select preset:" (but keeps 18 in the following unnamed field)
-Administration->Temperature Cooler Status graph does not adjust to show all temps, 39/40/50 degrees Celsius (2.4 GHz, 5 GHz, CPU)

No issues noted for any clients or network devices, all working as expected.

I have 100/10 Internet and getting 115/16: http://www.speedtest.net/my-result/d/15335e70-652c-4e82-909c-10207b6829ce

It helps that the house is 30 feet from the drop and having a powered cable booster block for all runs.

 

[eastavin]

Right now, the alpha 2 stage is focused on debugging new issues introduced by the GPL merge. Once that stage is done, I will bump the version.

I have been running the previous version and now this one RT-AC68U_384.3_alpha2-g5b8da38 since it came out. This is pretty darn good stuff for alpha software... just in case you did not know

Other people's alpha software crashes regularly. so far this stuff is just humming along like a finely tuned machine. Thank you for your fine work.

Edward

 

[Reinvented]

Speaking of memory use I've noticed that if I have AiProtection enabled my ram use is 33% plus but if I turn it off it drops down to 21%.

That's normal as another user pointed out. My statement was merely to draw attention to the fact that it doesn't shoot up in the span of 30 mins or so. It's creeping up slowly!

 

[skeal]

@RMerlin I'm getting this error in my logs since updating to newest alpha for ac3100.

TCP: time wait bucket table overflow

Is it something I've done wrong?

 

[M@rco]

Those with QoS issues, try removing the old Trend Micro signature files, then rebooting:

rm /jffs/signature/rule.trf
reboot

After reboot, wait at least 5 minutes before checking if it starts working.

@RMerlin, an update: on the latest alpha (RT-AC68U_384.3_alpha2-g5b8da38) I still can't use QoS. It just won't work:

I did a clean install this morning, factory reset before and afterwards with WPS button while powering on and a power cycle after the second factory reset. I haven't made any changes to the default config. I haven't restore any of my backups, I dind't restore any config files. I only went through QiS to set it up and I have USB3 Datatraveler plugged in the back.

I've setup QoS (Adaptive QoS, manual 20/200, fq_codel, Media Sreaming) and gave it at least 5 minutes to start. No charts. Toggled QoS off, waited, toggled it back on, waited for 5 minutes, no charts. Removed the old TM signature as suggested before, rebooted, waited for around 10 minutes. No charts.

For now, it's seems safe to say that my QoS is not working (yet) on the 384 branch. For what is worth, I have a RT-AC68U HW Rev E1. When you need more info, just let me know, I'll be happy to provide it, when you have time to debug this issue.

 

[M@rco]

@RMerlin I'm getting this error in my logs since updating to newest alpha for ac3100.

TCP: time wait bucket table overflow

Is it something I've done wrong?

I have this on occasions too. When I reboot it's gone. Not sure what causes it.

 

[Martineau]

I have this on occasions too. When I reboot it's gone. Not sure what causes it.

Note sure if this is still relevant?

TCP: time wait bucket table overflow

 

[skeal]

Note sure if this is still relevant?

TCP: time wait bucket table overflow

What should that tcp time out value be then? Any ideas?

 

[M@rco]

Note sure if this is still relevant?

TCP: time wait bucket table overflow

I'm sorry, but I don't understand your reply. I read the post you linked to several times trying to understand it, but I can't really see why this would apply. I did a clean install this morning and beside QiS setup and some trial (and mostly error) with QoS trying to provide feedback for @RMerlin nothing has changed in the (user part of the) configuration. As for wearing out NAND: this RT-AC68U has been purchased April 15th last year, so if NAND has already worn out, it seems quite fast...

From the SkyNet thread:

There's the infamous "flash wear" debate. (..) The write cycle count on modern nand is so high that its more likely you will get a new device or it will stop working from other causes before it ever became an issue.

Could there be another cause? As mentioned, I have seen it before, mostly after fresh installs (or maybe right after a reboot), but it's always gone after another reboot. Everything else seems to work fine, it's just those lines which keep filling syslog for some reason which is still unknown to me (which could perfectly be explained by a lack of knowledge on my behalf).

 

[Martineau]

I'm sorry, but I don't understand your reply. I read the post you linked to several times trying to understand it, but I can't really see why this would apply. .

Your OP

Not sure what causes it.

​

My intention was to allude to the fact that this error has been discussed before in the forum and was an attempt to provide a link to a post that may assist towards answering your query.

i.e. See the paragraph that contains "For the uninitiated, that particular message means .......etc."

Furthermore using the forum search function...
Possible cause

so there may be a misconfiguration issue?

 

[M@rco]

so there may be a misconfiguration issue?

I don't want to go off topic too far, as this doesn't seem an issue specific to this alpha release, but I wouldn't know where to start searching. I haven't seen it in a while (@skeal was actually the one who reported it a few posts back). I searched on earlier occassions, but couldn't really find a definitive cause and as it went just as fast as it came (after a reboot), I've never put much effort in finding out the cause.

I think my network is fairly simple: a router from my ISP in bridged mode, the RT-AC68U as main router in Wireless Router mode, we use two laptops, two iPads and three iPhones, an Apple TV, a Philips hue bridge, a security cam and the rest are IoT appliances (robot vacuum cleaner, thermostat, scale, AVR etc). The Google Chromecast which was connected until recently, but barely used, has been taken offline after the recently reported issues regarding possibly flooding your home network. All IP's are assigned through static leases, so there's nothing as far as I'm aware of on the clients side that I could have misconfigured. But thanks anyway for sharing, maybe @skeal can benefit from it and isolate a cause, and I will keep it in mind should it ever occur again. For a minute I though you meant my precious RT-AC68U was dying on me, when I read that first post you linked to and I read that horrifying part about NAND wearing out ...