[Alpha] Pre-release test builds available (both 380 and 382 branches)

[RMerlin]

Do you mean turning off just Mu-Mimo, or also inmplicit and explicit beam forming (since they are three different options in the professional menu)?

We were specifically discussing the uselessness of MU-MIMO. Beamforming works, but it can also introduce compatibility issues depending on your clients.

Merlin, are you going to be integrating on top of that newer code base, or wait until the dust settles a bit.

I can't answer that until I've actually seen the source code, which is not available yet.

 

[bradbort]

We were specifically discussing the uselessness of MU-MIMO. Beamforming works, but it can also introduce compatibility issues depending on your clients.



I can't answer that until I've actually seen the source code, which is not available yet.

Thanks.

What would the symptoms of MU-MIMO misbehaving be? I'm asking because I have a situation where the router needs to be on the first floor and one of my prime clients is in the basement. When I have MU-Mimo off, the gui map is telling me that my TX speed is good, but my RX speed is much reduced. When I enable it, both are good....I have about 15 other clients, of mixed speed capabilities, on my system.

 

[RMerlin]

What would the symptoms of MU-MIMO misbehaving be?

That will depend on the clients, it's not always the same thing. Samsung mobile devices for instance downgrade their connection to single-stream. Other devices will see their speed drop significantly (due to the BCM4366's problems). And others might fail to connect or stay connected.

 

[bradbort]

We were specifically discussing the uselessness of MU-MIMO. Beamforming works, but it can also introduce compatibility issues depending on your clients.



I can't answer that until I've actually seen the source code, which is not available yet.

I was poking around some more to see what the effects of turning the Mu=Mimo and beamforming on and off is having on my system.

This is probably a really basic question, but I'm finding that when I look at the speed (TX/RX) that my router says my basement PC is getting (in the list function of the GUI) it's way overstating what it is really getting.

For example, it currently says I'm getting 975 RX and TX. When I run a speed test from my internet procider, however, I'm only getting about 3-4. When I test with other PC's that are closer to the router, I get what I expect, about 150. When I swap out the wirless connection on my basement pc for a powerline adapter, my internet speed goes up to about 80M, not as good as the client's closer to the router, but fast enough to be useable.

So, my question is, what is the speed in the device map for the GUI on the router actually meaning? Why is it so far off what I'm achieving in the case of my one somewhat distant PC client? I'm wondering if I have a dodgy wifi adapter in the basement.

Thanks for any insights you can give me.

 

[qisdtech]

Is anyone having issues making the Dual-Wan feature of the RT-AC86U work? I have two ISP's and simply want it to failover to my backup provider when the primary is down. I've performed the steps and have my Primary ISP in the WAN port and the backup in LAN 1 port. I have searched the forums and few people report that the dual-wan implementation has been broken since .382 firmware was released. I have tried 382.1, 382.1.1 alpha, as well as the current stock firmware and the beta AIMESH one and they all work the same way with respect to dual wan, as in they don't

 

[skeal]

Without QOS classifying vpn traffic as anything (anything at all). QOS is basically useless if you use a vpn for video streaming.

Ok I found a solution to my problem with QOS not classifying traffic. I run static ip I have used the router as a modem by playing with iptv settings. I reverted back to a bridged modem and pow I have vpn classified traffic in QOS stats. Not sure if this cured it so I'm testing back and forth between modem and no modem I'll get back to you.

 

[Pir8pete]

@skeal and to others doing this

Yes it's off topic!!!

I just felt like picking on you as it was right in front of me today and i keep seeing many guys do this;
Do you manually update your sig. every time a new fw version comes out? LOL
That means you guys would be going into your account settings to change it every time!!! WOW must be bored, I see alot of guys put this crap in their sig. to show off or whatever and wondered.
Why not make it some what generic like mine to save all the trouble.
Or do you guys have some sort of script linked to the router that auto updates to sig. highly unlikely but I have to ask bahaha

 

[skeal]

@skeal and to others doing this

Yes it's off topic!!!

I just felt like picking on you as it was right in front of me today and i keep seeing many guys do this;
Do you manually update your sig. every time a new fw version comes out? LOL
That means you guys would be going into your account settings to change it every time!!! WOW must be bored, I see alot of guys put this crap in their sig. to show off or whatever and wondered.
Why not make it some what generic like mine to save all the trouble.
Or do you guys have some sort of script linked to the router that auto updates to sig. highly unlikely but I have to ask bahaha

Sorry but I don't see what this has to do with the test builds problems being discussed. You're entitled to your opinion I guess no matter how off topic it is.
Regards

 

[routers]

I just want to add that my RT-AC56U has been incredibly stable with Merlin/380.68x and I look forward to buy the RT-AC86U in the coming months. I commit to donate $10 in early december and you all should too if you haven't already, in my opinion. Thanks Merlin.

 

[RMerlin]

So, my question is, what is the speed in the device map for the GUI on the router actually meaning?

It's the link rate. On top of this, you have to add the encryption overhead from WPA2-AES. You have to take into account that wifi is a shared network, which means the available bandwidth is split among every devices on that same channel (and that includes your neighbours).

General rule of thumb is to expect roughly half the link rate as the actual throughput, provided there's no one else transmitting at the same time on that channel (or on an interfering channel).

 

[MidKnight]

Running 382.1_1 with the AC88U (had the same issue with 382.1 also), after a few days speeds drop slower and slower that by the time a week of uptime is up, things are coming to a crawl. What I've noticed is the CPU status of CORE 1 slowly gets higher and higher over the days, until it comes to a point that when theres traffic going through the router its on 100% until the traffic drops off.

Once router is rebooted, speeds are back to normal and CPU status varies as normal.

anyone having this type of issue?

 

[pattiri]

Running 382.1_1 with the AC88U (had the same issue with 382.1 also), after a few days speeds drop slower and slower that by the time a week of uptime is up, things are coming to a crawl. What I've noticed is the CPU status of CORE 1 slowly gets higher and higher over the days, until it comes to a point that when theres traffic going through the router its on 100% until the traffic drops off.

Once router is rebooted, speeds are back to normal and CPU status varies as normal.

anyone having this type of issue?

I haven't keep mine for that long but I think all the 382 releases eats more CPU and RAM. I have 50 mbit down speed. When torrenting I used to see CPU usage for core 1 max 45% with 380, with 382 sometimes it goes up to 90% and min 55-60%. I have doubts if router can handle 100 mbit with release 382. (HW acceleration is off)

 

[MidKnight]

seems odd that's for a few days after a reboot its working excellent, but slowly dies... I have HW acceleration set to AUTO, I'm going to try to set to disable and see what happens.

I normally run parental schedule and speed limits in my settings, but I turn all that off a few days ago to see if anything there was causing the issue, but seems not.

 

[skeal]

I'm using 382.1_1 and my AC3100 doesn't miss a beat. No speed decay here!

 

[MacG32]

Anyone know what this error message is from and why? https://pastebin.com/CC3MHUPf Too large to post here.

 

[Gingernut]

Anyone know what this error message is from and why? https://pastebin.com/CC3MHUPf Too large to post here.

I think it's an out of memory panic.

I was getting something similar on my 86U and Transmission kept being shut down, I ended up making a swap file on my attached drive and haven't seen that warning since.

 

[Vexira]

That will depend on the clients, it's not always the same thing. Samsung mobile devices for instance downgrade their connection to single-stream. Other devices will see their speed drop significantly (due to the BCM4366's problems). And others might fail to connect or stay connected.

Wait quirousity could it also involve connection issues to devices in the form of apps having delays eg Skype having delays into receiving messages or a Windows phone dropping connection then re connecting, I just noticed this on my 87u in April mode with mu mimo on the 5g band.

 

[bradbort]

It's the link rate. On top of this, you have to add the encryption overhead from WPA2-AES. You have to take into account that wifi is a shared network, which means the available bandwidth is split among every devices on that same channel (and that includes your neighbours).

General rule of thumb is to expect roughly half the link rate as the actual throughput, provided there's no one else transmitting at the same time on that channel (or on an interfering channel).

Thanks. I actually found that my usb3 ports were misbehaving and throttling back the speed to something like 1-5% of the link speed. Eventually, I gave up and just bought an internal wifi adapter card (asus's), and everything is right with the world again.

I wonder if windows10 creator edition did something to the usb 3/wifi. At any rate, this is a bit off topic, but thanks for filling me in on the difference between link speed and realized speed. The more you know...

 

[RMerlin]

I uploaded new builds these past few days.

382:

c8fdf58 nvram: split dhcp_staticlist into 10 nvrams on HND platform to work around the 1000 chars limitation
e3f4935 samba35: fix incomplete a606baa2b7743d7ae2fc5e482802bf78c2582de7 merge
dee9cea Updated documentation
120677f samba35: only fallback to anonymous if authentication was not requested (CVE-2017-12150)
a606baa samba35: Prevent client short SMB1 write from writing server memory to file. (CVE-2017-12163)
d1a7664 samba35: Chain code can return uninitialized memory when talloc buffer is grown. (fixes CVE-2017-15275)
57140f4 samba36: only fallback to anonymous if authentication was not requested (CVE-2017-12150)
f25d797 samba36: Prevent client short SMB1 write from writing server memory to file. (CVE-2017-12163)
c850234 samba36: Chain code can return uninitialized memory when talloc buffer is grown. (fixes CVE-2017-15275)
775e78a webui: re-sync AiProtection page with Asus's GPL
d507d26 webui: accept ":" as valid character for validator.safeName()
5802d1e webui: remove extra line leftover from copy/paste (close #49)

380:

f4da758 samba35: fix incomplete 300792d3ab171669978b8f44f8df8c57f698e269 merge
ffb1308 Updated documentation
a57deab webui: replace remaining is_string() calls (which no longer exists) with validator.isString() method
dc23968 webui: accept ":" as valid character for validator.safeName()
f785f26 webui: make 'sign in' a button instead of <div> (fixes #1488)
09e22f2 samba35: Chain code can return uninitialized memory when talloc buffer is grown. (fixes CVE-2017-15275)
300792d samba35: Prevent client short SMB1 write from writing server memory to file. (CVE-2017-12163)
af91fd1 samba35: only fallback to anonymous if authentication was not requested (CVE-2017-12150)
50c8dd3 samba36: only fallback to anonymous if authentication was not requested (CVE-2017-12150)
97e9269 samba36: Prevent client short SMB1 write from writing server memory to file. (CVE-2017-12163)
749c5dd samba36: Chain code can return uninitialized memory when talloc buffer is grown. (fixes CVE-2017-15275)
1231157 webui: update OUI database

 

[bradbort]

Thanks for the new build! I just uploaded it, and things went smoothly. Everything is initially very stable and fast.