New Pfsense install results in greatly reduced upload/download speeds
-
Indeed you should only use one for any specific traffic flow if you're looking for best throughput.
-
Also check your MTU settings they can really bonk up things
-
@stephenw10 Thanks for that. I'll dump the traffic shaper and keep the limiters to see what happens
-
@JonathanLee I didn't enter any values for those but it seems they are currently set to 1500, which is recommended for Xfinity Internet. Does this need to be changed/removed?
-
@spotlizard set it to 1500 on WAN see if that helps, I thought it would be 1492
-
@JonathanLee Will do. Thanks
-
@spotlizard I'm not convinced that a bufferbloat test, performed using a web based tool, is telling the whole truth.
I didn't know about bufferbloat until I saw some youtube video about it a few years back. And I remember that I constantly got A+ rating and pretty much zero added latency under load whenever I tested it. Starting from a very low latency of around 1-2 ms unloaded... Back then my fiber connection was 100 then 250 and until a year ago I had 1 Gig fiber.
Now I have a 10G connection with my fiber terminating directly in pfsense. And when I just now ran the test I only got a B rating with 32 ms added latency in the uplink (zero down). But the speed isn't even close to the 10G I have. And it never is with any tests that I have done using a web browser (typically topping out around 3500-4000 Mbit/s). Running ookla's application however, I can get 8+ Gbit in both directions.
So, just for fun I added a limiter set at 3400 Mbit/s in the uplink and then I get an A+ rating with no added latency during load.
BUT, all I have done now is to limit my internet connection to the capability of my PC running a web based test tool. I don't think this will improve the gaming experience for myself or anyone in the household one bit...
Meaning, the test method needs to be considered as well. And as already mentioned, running tests from pfsense is not the best idea, as it's not meant for that.
-
@Gblenn I would be inclined to agree with that. I connected 3 different laptops to the LAN port of the appliance with no other network devices connected. I got staggeringly different results based on the spec of the machine connected
-
So this has been a voyage of discovery and I think I found the problem (NOT Pfsense).
About 2 weeks ago the 'bomb cyclone' hit western Washington. Our power glitched on and off for about 10 seconds before we lost it completely for 24 hours. Once the power was restored I found that two of my cable TV boxes no longer functioned (stuck in a permanent boot loop with no way to break out). This resulted in my finally giving in and calling Xfinity (Comcast). As they always do, the first thing they did was sent a 'reset' signal to my modem. When this didn't work they monkeyed around with some other stuff in the background. Long story short I ended up going to an Xfinity store and getting replacement RV boxes to fix the problem. All of this happened before I stood up Pfsense.
Yesterday afternoon I removed into my cable modem, looking for and WoS type settings I could tweak and that's when I found it. They had taken the modem OUT of Bridge mode. A normal reset doesn't change this, so their tech must have done it in the background as part of their troubleshooting. Once I changed it back and restarted Pfsense I saw a significant increase in performance. Not quite back to where it was in the old setup but I'm a lot happier now.
My sincere thanks for all your help on this.
-
@spotlizard said in New Pfsense install results in greatly reduced upload/download speeds:
They had taken the modem OUT of Bridge mode. A normal reset doesn't change this, so their tech must have done it in the background as part of their troubleshooting. Once I changed it back and restarted Pfsense I saw a significant increase in performance.
Wouldn't this have been visible in the pfsense dashboard? Your WAN IP would change to something very different than the public IP normally showing up there... probably a 192.168.1.N subnet?
