QoS not working on latest version 386.3

[Tech9]

I've been using Manual setting because I've heard QoS works better if you set 10% less of your actual upload/download speeds

Correct. Use Manual Setting.

 

[Tech9]

Lan-to-Lan will work. It actually allows one subnet to see the other

Sounds like misconfigured Access Point to me with NAT and DHCP left enabled.

 

[SomeWhereOverTheRainBow]

Sounds like misconfigured Access Point to me with NAT and DHCP left enabled.

I couldn't agree more. I would rather Lan-to-Wan.

 

[Tech9]

I couldn't agree more.

Good night! You're so annoying.

 

[SomeWhereOverTheRainBow]

Good night! You're so annoying.

More cheese with that wine please! Good Night man!

 

[Volt]

A bit off-topic, but have you ever seen the following messages in the log when using the Adaptive QOS? I get them on the stock firmware and thinking of switching to Merlin, because looks like they make Internet connection unstable when occur.

These messages look something like this:

ay  7 08:03:46 kernel: ------------[ cut here ]------------
May  7 08:03:46 kernel: WARNING: CPU: 3 PID: 6200 at net/sched/sch_htb.c:568 htb_qlen_notify+0x78/0x80()
May  7 08:03:46 kernel: Modules linked in: tdts_udbfw(O) init_addr(          (null) -           (null)), core_addr(ffffffbffc15b000 - ffffffbffc160500)
May  7 08:03:46 kernel:  tdts_udb(PO) init_addr(          (null) -           (null)), core_addr(ffffffbffc357000 - ffffffbffc37a054)
.......
.......
May  7 08:03:47 kernel: CPU: 3 PID: 6200 Comm: tc Tainted: P        W  O    4.1.52 #2
May  7 08:03:47 kernel: Hardware name: Broadcom-v8A (DT)
May  7 08:03:47 kernel: Call trace:
.......
.......

Resetting the router does not help.

 

[ColinTaylor]

Lan-to-Lan will work. It actually allows one subnet to see the other without defining any rules to Static-Routers. It is one way you can cascade routers, but not a very popular way. Some routers do not fully support it. ergo no QoS support.

How To Cascade Routers To Isolate Your IoT Devices – MBReviews

In this article we're going to explore some of the most common methods of cascading routers in order to isolate your smart devices from the rest of the

www.mbreviews.com

You would have to disable DHCP tho or make them non-overlapping.

The LAN to LAN setup described in that link is not using separate subnets. Both routers are on the same subnet.

 

[SomeWhereOverTheRainBow]

The LAN to LAN setup described in that link is not using separate subnets. Both routers are on the same subnet.

You are correct ColinTaylor. It was used mearly to show how a proper lan-to-lan configuration could look like. Any miss-configuration being contrary to the example.
Two ways to LAN-to-LAN
Turn dhcp off,
or
set one dhcp range on Primary Router to be something like 2 to 150
set dhcp range on Secondary Router to be something like 151 to 254

The easier and most acceptable (play nice) approach is to disable dhcp as the guide instructs.

 

[ColinTaylor]

or
set one dhcp range on Primary Router to be something like 2 to 150
set dhcp range on Secondary Router to be something like 151 to 254

That's not a valid setup. If both routers are part of the same subnet you can't have two active DHCP servers. It doesn't matter whether each server has different IP ranges or not.

 

[SomeWhereOverTheRainBow]

That's not a valid setup. If both routers are part of the same subnet you can't have two active DHCP servers. It doesn't matter whether each server has different IP ranges or not.

It can be done it is just not advisable because the first dhcp that responds decides the address.

Example

Router 1 IP -192.168.1.1

IPrange 192.168.1.2-192.168.1.150

Router 2 IP- 192.168.1.151

IPrange 192.168.1.152-192.168.254

The disadvantage is virtually no control over which DHCP gets priority since it depends on which one responds first or which one the device is connected to.

 

[ColinTaylor]

It can be done it is just not advisable because the first dhcp that responds decides the address.

I would say it's not simply "not advisable" but just plain wrong IMHO. Not only for the reason you just stated but also because it would break local DNS name resolution (dnsmasq automatically registers clients in its DNS) and might cause problems with lease renewal. I can't think why someone would deliberately misconfigure their network like this.

 

[SomeWhereOverTheRainBow]

I would say it's not simply "not advisable" but just plain wrong IMHO. Not only for the reason you just stated but also because it would break local DNS name resolution (dnsmasq automatically registers clients in its DNS) and might cause problems with lease renewal. I can't think why someone would deliberately misconfigure their network like this.

I am in full accordance with you. It is just one example that can go wrong in a Lan-to-Lan configuration. That is why the dhcp off is the preferred method (only method by your book).

 

[SomeWhereOverTheRainBow]

I would say it's not simply "not advisable" but just plain wrong IMHO. Not only for the reason you just stated but also because it would break local DNS name resolution (dnsmasq automatically registers clients in its DNS) and might cause problems with lease renewal. I can't think why someone would deliberately misconfigure their network like this.

@ColinTaylor

I could really use your skills in this post

http://www.snbforums.com/threads/asuswrt-merlin-386-3-is-now-available.73815/post-703447

Can you think of away a %2D can be replaced when using

nvram set acc_num="1"
nvram set acc_list="$(nvram get http_username)>$(nvram get http_passwd)"
nvram set acc_webdavproxy="$(nvram get http_username)>1"
nvram commit
reboot

If the user has a dash in the $(nvram get http_username) value, then this fails because nvram get acc_list substitutes dashes with a %2D . can you think of how this issue can be remedied so that the username properly reflects the dash with a %2D so the above commands will work properly.

edited to represent @ColinTaylor error correction %2D

 

[ColinTaylor]

@ColinTaylor

I could really use your skills in this post

http://www.snbforums.com/threads/asuswrt-merlin-386-3-is-now-available.73815/post-703447

Can you think of away a %2 can be replaced when using

nvram set acc_num="1"
nvram set acc_list="$(nvram get http_username)>$(nvram get http_passwd)"
nvram set acc_webdavproxy="$(nvram get http_username)>1"
nvram commit
reboot

If the user has a dash in the $(nvram get http_username) value, then this fails because nvram get acc_list substitutes dashes with a %2 . can you think of how this issue can be remedied so that the username properly reflects the dash with a %2 so the above commands will work properly.

The dash symbol is actually represented in hex by %2D not %2.

Is this a recurring problem or just a one-off fix?

 

[SomeWhereOverTheRainBow]

The dash symbol is actually represented in hex by %2D not %2.

Is this a recurring problem or just a one-off fix?

if you adjust the admin name on the administrator page, it creates a reoccurring problem from what i have seen.

 

[Tech9]

A bit off-topic, but have you ever seen the following messages in the log when using the Adaptive QOS?

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

 

[Tech9]

It can be done

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.

 

[wub901]

@macster2075 as @Tech9 has pointed out via his own testing, if your configuration is a dual wan, then the issue might reside with dual wan configuration. look at his previous post for more deets.

http://www.snbforums.com/threads/qos-not-working-on-latest-version-386-3.73901/post-703389

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

 

[wub901]

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

 

[ColinTaylor]

if you adjust the admin name on the administrator page, it creates a reoccurring problem from what i have seen.

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.