QoS not working on latest version 386.3

[Tech9]

I am running AC88U with 386.3 Traditional QoS does not work infact it slows my connection down to 5 meg download

I currently play with an older RT-AC1900P with exactly the same BCM4709C0 1.4GHz CPU as in your RT-AC88U. With Traditional QoS enabled (it disables NAT acceleration) the router can do >200Mbps no problem. Something else is wrong with your router or firmware. Mine is running Asuswrt 386_43129.

 

[SomeWhereOverTheRainBow]

I currently play with an older RT-AC1900P with exactly the same BCM4709C0 1.4GHz CPU as in your RT-AC88U. With Traditional QoS enabled (it disables NAT acceleration) the router can do >200Mbps no problem. Something else is wrong with your router or firmware. Mine is running Asuswrt 386_43129.

Or ethernet cable. Make sure you are using an up-to-date cable as well.

 

[SomeWhereOverTheRainBow]

You Somewhere Over the Rainbow people created the most pointless configuration example.
I hope it won't be adopted and included in other pointless recommendations regularly floating around.

All the better to keep you tech9 people on your toes instead of somewhere over the rainbow.

 

[SomeWhereOverTheRainBow]

I am running AC88U with 386.3 Traditional QoS does not work infact it slows my connection down to 5 meg download

What does your connection look like without traditional QoS?

 

[john9527]

The router's disk_share code uses this function to translate non-alphanumeric characters. I can't think of a simple solution to emulating the same translation that doesn't involve writing a similar function into a script. In the long run it's probably better to just stick with a user name that consists of only alphanumeric characters.

If all you are worried about are dashes.....pipe the get thru sed
$(nvram get http_username | sed "s/%2D/-/g")

 

[ColinTaylor]

If all you are worried about are dashes.....pipe the get thru sed
$(nvram get http_username | sed "s/%2D/-/g")

Good point. I think he would need to do it in the other direction:

$(nvram get http_username | sed "s/-/%2D/g")

 

[john9527]

Good point. I think he would need to do it in the other direction:

$(nvram get http_username | sed "s/-/%2D/g")

I didn't have Merlin up to try it....my fork doesn't do any translation, so wasn't sure about which direction

 

[SomeWhereOverTheRainBow]

Good point. I think he would need to do it in the other direction:

$(nvram get http_username | sed "s/-/%2D/g")

That is basically how I modified for the one use case the other night. My problem arose from changing the admin name on administrator page a while back. I didn't realize when I changed it to include a dash, that the admin name on disk share would not update properly. It was a previous post provided by ColinTaylor that provided the key to remedying the issue.

 

[Tech9]

sed "s/%2D/-/g")

Aha... please, don't shoot!

 

[L&LD]

Do you have the up and down numbers in the correct boxes?

 

[slidermike]

Isn't traditional qos broken; like for a long time. Broken for a long time then fixed for some revisions and then broken again and since?
Isnt that a part of why folks were told to use adaptive qos if they needed qos?
Could be faulty or incomplete memory on my part.

Here is a thread that sort of covers it to some degree.

(RT-AC88U) Traditional QoS fully fixed (I believe)

EDITED: applies to 384.4.beta1 Looking closely at the output of the various tc -s ... commands when downloading/uploading files, I have discovered that the ingress (ffff:) qdisc on the eth0 interface behaves very strangely: downloaded/ingress traffic DOES appear on the ingress (ffff:) qdisc...

www.snbforums.com

 

[Volt]

No. It would be nice to mention the router model at least.

Regardless of the model.

 

[Jumpstarter]

Regardless of the model.

Basically heinz 57 model.

 

[Jumpstarter]

No. Tested already. QoS works properly in double NAT.

It even works in a triple nat with the bug zapper on.

 

[wub901]

What does your connection look like without traditional QoS?

It works fine if the qos is disabled I reset the router back to stock and re uploaded the merlin firmware now if I enable traditional qos it still downloads at max speed and not what I have set it to.

 

[macster2075]

I've been playing around with Traditional Qos and the rules and I see the rules are a hit or miss. Sometimes it works, but most of the times it completely bypasses the rules altogether and I get the full speeds when doing a speedtest. I even set the device to the lowest download/upload and set the priority of the rule to match it,..but again, it's a hit or miss.

Are there any known issues with QoS and rules? I started to test this after watching a recent Youtube video about a guy having issues with it with stock Asus firmware... I am seeing the same issues as he.

 

[L&LD]

As you found out, issues abound.

 

[macster2075]

oh ok.. just wanted to know so I don't keep trying it out haha. Be nice if it worked as I would like to have each device have it's own min/max speed. The way I have it now is using Bandwidth Limiter, but I have to set the speeds manually... it works, but it's more "work" lol

 

[L&LD]

Why do you need to be in such control? Do they (the client devices) misbehave when they're set to freely use the bandwidth they (obviously) need?

 

[macster2075]

Why do you need to be in such control? Do they (the client devices) misbehave when they're set to freely use the bandwidth they (obviously) need?

Lots of Youtube/Instagram uploading and also to have full control of the Guest network so they get a dedicated bandwidth.
Before I started using Bandwidth Limiter, every time someone uploaded something to any cloud service, it would bog down the rest of the bandwidth..not anymore.