What's new

FlexQoS FlexQoS 1.2.4 - Flexible QoS Enhancement Script for Adaptive QoS

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

First notes after updating:
-my tab is still on 4th position even after closing/reloading the browser tab. Do we need to restart FlexQoS?
-personally I used to click the update check regularly when I opened the tab, now as it's out of sight maybe I won't anymore. Don't know if it's good or bad this way. Just my thoughts that I preferred the old position for version and check more. ;)

Best wishes,
chris
Trying out the beta.

Those tweaks you made, make it look very clean.

My flex tab is also in 4th position. Maybe a router restart will fix this but I'll hold off for now.

I appreciate your hardwork!!!

UPDATE: the router reboot fixed the the Flex into the 3 tab!!!
 
Last edited:
is the rt-ac5300 an hnd router to use cake or the new functions ?
 
I hope I didn't offend you with my suggestion. All your efforts are VERY-MUCH appreciated.
No offense taken. I meant in terms of the effectiveness of the Trend Micro classification engine.
 
I'm trying out the latest RMerlin 386.2 beta1. Also, this is with a fulltime VPN and FlexQoS (beta) enabled :

Screenshot (5).png
 
No offense taken. I meant in terms of the effectiveness of the Trend Micro classification engine.
Good... cause even "Partial" Qos/Network-shaping/Bandwidth-optimizing/Network-fairness" is better than "None"... Totally Understood!!!
 
So I usually use your recommended priority list because of the changed Learn from home category usage. So do I need to change it now? And to what if you have any suggestions? Thank you
There is no need to change anything. There are 2 changes I made:
  1. Renamed the label from "Game Transferring" to "Learn-From-Home". No other functional difference.
  2. Modified the default "Game Downloads" iptables rule to send traffic to the "File Transferring" priority instead of "Game Transferring".
All the original uselessness of Learn-From-Home still applies now. The category can be used for whatever traffic you like, because nothing will ever reach it as long as it is lower priority than Streaming and Web Surfing. But to minimize confusion, I decided I want the name to match the QoS tab when using the Customize list.
 
First notes after updating:
-my tab is still on 4th position even after closing/reloading the browser tab. Do we need to restart FlexQoS?
-personally I used to click the update check regularly when I opened the tab, now as it's out of sight maybe I won't anymore. Don't know if it's good or bad this way. Just my thoughts that I preferred the old position for version and check more. ;)

Best wishes,
chris
True, the tab will not move if it's already present. So a reboot or disable/enable at the command line should also do it.

I was bothered by all the buttons at the top of the page. I wanted to be able to hide the things you don't need to use every day. Eventually, updates will slow down and the button will gather dust.
 
is the rt-ac5300 an hnd router to use cake or the new functions ?
It is not an HND model. Run uname -r and if it is 2.6.36, then the kernel is too old for Cake. HND models have a 4.1.* kernel.
 
There is no need to change anything. There are 2 changes I made:
  1. Renamed the label from "Game Transferring" to "Learn-From-Home". No other functional difference.
  2. Modified the default "Game Downloads" iptables rule to send traffic to the "File Transferring" priority instead of "Game Transferring".
All the original uselessness of Learn-From-Home still applies now. The category can be used for whatever traffic you like, because nothing will ever reach it as long as it is lower priority than Streaming and Web Surfing. But to minimize confusion, I decided I want the name to match the QoS tab when using the Customize list.
Thanks for the reply!! Appreciated...
 
Another small update pushed to develop:

- Re-add the qdisc check that broke things this morning.
- Change the minimum quantum/burst to use the formula from OpenWRT sqm-scripts. This basically assumes a worst-case scenario minimum for an ATM-based connection.

Practically speaking, the minimum quantum and burst for a connection with a MTU between 1489 and 1536 would be 1749. Previously, the minimum would have been 1600. I don't expect this to have much effect on anything but I don't find any reason to second-guess the expertise over there. You always want to ensure you can burst a full packet, and the worst-case overhead is associated with ATM-based connections.

Time for bed now! o_O
 
It is not an HND model. Run uname -r and if it is 2.6.36, then the kernel is too old for Cake. HND models have a 4.1.* kernel.
thanks dave for your fast support, sadly its 2.6.36 ;-)
 
My kitchen chromecast, which has ip 10.0.0.67, is inside a rule that directs the subnet 10.0.0.65/29 to streaming.
However, it's traffic is being redirected to other categories.
In fact, when I filter it's ip, their traffic don't show on the list. Nothing shows.

Could it be related for it is being using IPV6?

Characters limit: https://pastebin.com/iAAET5a4
 
My kitchen chromecast, which has ip 10.0.0.67, is inside a rule that directs the subnet 10.0.0.65/29 to streaming.
However, it's traffic is being redirected to other categories.
In fact, when I filter it's ip, their traffic don't show on the list. Nothing shows.

Could it be related for it is being using IPV6?

Characters limit: https://pastebin.com/iAAET5a4
It’s certainly possible that the traffic is using IPv6, but you should still be able to see it in the connections list. Or because the rule is too high in the list and a lower rule is overriding it.

Also, in studying your iptables rule, I saw a rule with ports 83,853. Should that be 53,853?
 
It’s certainly possible that the traffic is using IPv6, but you should still be able to see it in the connections list. Or because the rule is too high in the list and a lower rule is overriding it.

Also, in studying your iptables rule, I saw a rule with ports 83,853. Should that be 53,853?
Indeed good catch!

It's ipv6 certainly, when filtering the device it only shows its ipv4 Traffic, because the nat ip is ipv4(10.0.0.67$).
Ipv6 Traffic can't be filtered. In this case it must using only ipv6 and the rule i created don't apply :(
 
Another small update pushed to develop:

- Re-add the qdisc check that broke things this morning.
- Change the minimum quantum/burst to use the formula from OpenWRT sqm-scripts. This basically assumes a worst-case scenario minimum for an ATM-based connection.

Practically speaking, the minimum quantum and burst for a connection with a MTU between 1489 and 1536 would be 1749. Previously, the minimum would have been 1600. I don't expect this to have much effect on anything but I don't find any reason to second-guess the expertise over there. You always want to ensure you can burst a full packet, and the worst-case overhead is associated with ATM-based connections.

Time for bed now! o_O

could be placebo but the dev changes seem to have overall performance improvements since latest stable update
 
could be placebo but the dev changes seem to have overall performance improvements since latest stable update
I definitely atest to that....getting A bufferbloat even with VPN and none of my devices (streaming wireless) are lagging. Awesome work @dave14305
 
I always get a kick out of graphs like this showing QoS priorities at work. I posted one not too long back.

The blue upload line is a OneDrive upload (5th priority). The green line is a speedtest with Comcrap (4th prio). When it’s time for the upload test, the lower priority upload concedes bandwidth to the higher priority speed test. Then it resumes its normal load.

1615594482948.png
 
Last edited:
Another situation - the only connected device on my network right now is using WhatsApp video call. It is being thrown to web traffic, what the heck is STUN?

Edit: this is my children device which have Google's family link installed, maybe this program creates some kind of tunnel through their servers?

And offtopic - why is it connecting to a LAN ip? I don't have that subnet on my network, my subnet is 10.0.0.0/24
 

Attachments

  • Screenshot_20210313-094920892.png
    Screenshot_20210313-094920892.png
    203.3 KB · Views: 93

Latest threads

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top