is it possible to traffic shape between devices

Started by clutchmaster, October 29, 2024, 01:04:27 PM

Previous topic - Next topic
Those two rules 3 and 4 attached on OPT3, what direction you have them set?

Regards,
S.
Networking is love. You may hate it, but in the end, you always come back to it.

OPNSense HW
APU2D2 - deceased
N5105 - i226-V | Patriot 2x8G 3200 DDR4 | L 790 512G - VM HA(SOON)
N100   - i226-V | Crucial 16G  4800 DDR5 | S 980 500G - PROD

Quote from: Seimus on October 30, 2024, 12:59:07 PM
Those two rules 3 and 4 attached on OPT3, what direction you have them set?

Regards,
S.

would i have to invert que i cloned them for my global limit to save time

Quote from: Seimus on October 30, 2024, 12:59:07 PM
Those two rules 3 and 4 attached on OPT3, what direction you have them set?

Regards,
S.
heres an image of the current config

( https://imgur.com/ndvIAxd )

Are you trying to limit speed between devices that are connected to the same switch?

Quote from: wiggleroom on October 30, 2024, 02:48:18 PM
Are you trying to limit speed between devices that are connected to the same switch?
im not trying to limit speed between devices on the switch... essentailly trying to load balance and prioritize gaming and video traffic for example  if opt4 is playing a game while also has a youtube video going then opt3 starts downloading a steam game i dont want his download to hog all bandwidth resulting in ping issues & packet loss or kicked from server and video buffering but if the network is rather quiet sure let him hit the limit

1. I assume OPT3 is your WAN, right? I have not checked all your directions and source/destinations, either.

2. How are you testing? If you use speedtest.net, you MUST set the little "connections" knob below the "Go" button to "single" instead of "multi".

That being said, here are my settings that work by defining rules between WAN and LAN, not via subnets. I just checked that when I limit my downstream pipe to a smaller value, this works (with single connection).

Intel N100, 4 x I226-V, 16 GByte, 256 GByte NVME, ZTE F6005

1100 down / 800 up, Bufferbloat A+

thanks guys for time realy appreciate it
heres how i have the ports configured

That correlates only slightly with your topology image. If all the LAN ports share the same subnet and are bridged, you should use the bridge ports for all firewall rules and set up the mandatory bridging tuneables.

There should be no rules set for the bridge member interfaces.
Intel N100, 4 x I226-V, 16 GByte, 256 GByte NVME, ZTE F6005

1100 down / 800 up, Bufferbloat A+

Quote from: meyergru on October 30, 2024, 04:35:00 PM
That correlates only slightly with your topology image. If all the LAN ports share the same subnet and are bridged, you should use the bridge ports for all firewall rules and set up the mandatory bridging tuneables.

There should be no rules set for the bridge member interfaces.
thanks for the help! when i started out on this everything i found didnt directly touched on this.. i was quite confused on how that was used

now its working, i take it now all i gotta do to isolate which port its happening on would be to change "any" to the specific ip address... than duplicate for each port or is that unnecessary for what im tryin to do

Again: Your LAN is LAN - you should not apply any rules, configurations a.s.o. on bridge member ports.

Think of them as ports on a switch. If you want to differentiate between things that are on your LAN, use their IPs or MACs on rules, not the bridge ports.

To not get confused, it would probably be best to not list those low-level member interfaces at all by removing their uppercase names (like OPT3) from the assignments. The bridge itself is defined on the physical devices names.

This is point 2 here, for a reason.
Intel N100, 4 x I226-V, 16 GByte, 256 GByte NVME, ZTE F6005

1100 down / 800 up, Bufferbloat A+

Quote from: meyergru on October 30, 2024, 05:30:49 PM
Again: Your LAN is LAN - you should not apply any rules, configurations a.s.o. on bridge member ports.

Think of them as ports on a switch. If you want to differentiate between things that are on your LAN, use their IPs or MACs on rules, not the bridge ports.

To not get confused, it would probably be best to not list those low-level member interfaces at all by removing their uppercase names (like OPT3) from the assignments. The bridge itself is defined on the physical devices names.

This is point 2 here, for a reason.
i dont see anything in relation to adding the second interface in the rules section

but bit a weird behavior now i run a buffer bloat test i get 20mbits/s while my steam download is getting 20mbytes/s
i have the pipe set for 20mbits and yes i have steam set to display bytes instead of bits


The bridge is one single interface. Like a switch in a consumer router.
Deciso DEC750
People who think they know everything are a great annoyance to those of us who do. (Isaac Asimov)

Quote from: clutchmaster on October 30, 2024, 07:03:23 PM
i dont see anything in relation to adding the second interface in the rules section

You did configure OPT3 in your shaper.

Quote from: clutchmaster on October 30, 2024, 07:03:23 PM
but bit a weird behavior now i run a buffer bloat test i get 20mbits/s while my steam download is getting 20mbytes/s
i have the pipe set for 20mbits and yes i have steam set to display bytes instead of bits

As I said: the shaper handles one stream only. If applications choose to use several of them, you are out of luck. When I limit my downstream to, say, 100 Mbit/s and try testing with Speedtest and multiple connections, I still get >300 MBit/s downstream. That is at least my experience.

What my shaper settings are good for, is handling bufferbloat. IDK about "fairness". The docs give multiple different configurations for specific purposes, but since you did not succeed with that (now we know that was probably for a different reason), I showed you my config aimed at reducing bufferbloat.
Intel N100, 4 x I226-V, 16 GByte, 256 GByte NVME, ZTE F6005

1100 down / 800 up, Bufferbloat A+

Quote from: meyergru on October 30, 2024, 08:01:40 PM
Quote from: clutchmaster on October 30, 2024, 07:03:23 PM
i dont see anything in relation to adding the second interface in the rules section

You did configure OPT3 in your shaper.

Quote from: clutchmaster on October 30, 2024, 07:03:23 PM
but bit a weird behavior now i run a buffer bloat test i get 20mbits/s while my steam download is getting 20mbytes/s
i have the pipe set for 20mbits and yes i have steam set to display bytes instead of bits

As I said: the shaper handles one stream only. If applications choose to use several of them, you are out of luck. When I limit my downstream to, say, 100 Mbit/s and try testing with Speedtest and multiple connections, I still get >300 MBit/s downstream. That is at least my experience.

What my shaper settings are good for, is handling bufferbloat. IDK about "fairness". The docs give multiple different configurations for specific purposes, but since you did not succeed with that (now we know that was probably for a different reason), I showed you my config aimed at reducing bufferbloat.

in other words im at square one, what im trying to do is fairness and deprioritize high speed downloads then i discover after reading this
( https://www.reddit.com/r/PFSENSE/comments/y833jl/i_want_to_limit_steam_to_50_mbps/ )
apparently steam just blows past whatever you set since it treats the term "megabit" as "megabyte"