RB2011UiAS-2HnD-IN: Please help me understand the difference in throughput with these two master/slave switch configs.

Hi Forum,

I have a RB2011UiAS-2HnD-IN behind a pfSense router.
The RB2011 was my main router before I got pfSense. I want to use it as a simple AP + switch.

The Routerboard is set to WISP AP: bridge mode, the version is 6.28 reinstalled via netinstall.
ether1 is connected to the pfSense router which does the Firewalling, DHCP, NAT etc.

My WAN connection is 120Mbps/12Mbps

Configuration 1: speedtest from my computer gives around 50 mbps on Speedtest.
ether1 is master switch port of ether 2-5
ether6 is master switch port of ether 7-10
bridge1 ports are: ether1 and ether6 and wlan1

My understanding is that in this case the connection from my computer (ether2) to pfSense (ether1) goes through the switch1 CPU. Am I correct?
Ether1 as master switch port slow result.JPG
Configuration 2: speedtest from my computer gives 90+ mbps on Speedtest.
ether2 is master switch port of ether 3-5
ether6 is master switch port of ether 7-10
bridge1 ports are: ether1, ETHER2, ether6 and wlan1

In this case the connection from my computer to pfSense goes through the Routerboard CPU.
Ether2 as master switch port fast result.JPG
I have tested each config multiple times and I get the same results.
Can someone please explain to me why I see a difference in throughput?
In my opinion configuration 1 should give me the same throughput as configuration 2 because it uses the switch CPU.
What am I missing?

Thanks!
BenCo

post an export of your config (without sensitive info) to check it

Looking like you are saying that bridging ports give you better result than switching. I would suggest to remove the configuration without keeping defaults, don’t use quickset and set the ports manually. Then make the tests. Also check the speed negotiation on ports and make always btest from computer to 2011 also to see the differences. Watch profiler when testing.

Look at interface queues also. Check what happens if you swap hardware only queue for ethernet default queue.

Hi Guys.

I have reset my routerboard and this is an export.
Did not use QuickSet this time and just added a bridge and DHCP client on the bridge and then set up the switch ports.
With this base config I get the same results. Bridging is faster than switching when testing with speedtest.net and the dslreports speedtest.
Also tried changing the hardware queues to ethernet-default. Same result.
ether1: pfSense
ether2: computer

Configuration 1 (ether2 as slave port to ether1)
/interface bridge
add name=bridge1
/interface ethernet
set [ find default-name=ether2 ] master-port=ether1
set [ find default-name=ether3 ] master-port=ether1
set [ find default-name=ether4 ] master-port=ether1
set [ find default-name=ether5 ] master-port=ether1
set [ find default-name=ether7 ] master-port=ether6
set [ find default-name=ether8 ] master-port=ether6
set [ find default-name=ether9 ] master-port=ether6
set [ find default-name=ether10 ] master-port=ether6
/interface wireless
set [ find default-name=wlan1 ] l2mtu=2290
/port
set 0 name=serial0
/interface bridge port
add bridge=bridge1 interface=ether1
add bridge=bridge1 interface=ether6
add bridge=bridge1 interface=wlan1
/ip dhcp-client
add default-route-distance=0 dhcp-options=hostname,clientid disabled=no
interface=bridge1
/romon port
add disabled=no

This is the graph of the BTest application. Default settings. Receiving from bridge1 to my computer.
Both configs give me the same graph which displays full throughput.
BTest.JPG
I will have to take my routerboard to a friends house this weekend to rule out pfSense.

have you tested with another pc´s?? another gigabit NIC ??? updated drivers on NIC?? its very rare

i have tested the gigabit switch on rb 951g and rb 2011 and fast ethernet switch on 951ui and rb2011

they do wirespeed switching without problem