Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    Inconsistent Network Performance / Connectivity using Iperf3

    Scheduled Pinned Locked Moved General pfSense Questions
    2 Posts 2 Posters 427 Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • B
      Boethius
      last edited by Boethius

      I'm not getting results that make any sense to me when trying to troubleshoot my slow network. On one hand I assume I don't know how to use the tools properly. On the other, I wonder how I can be using tools wrong such as fast.com.

      Strange results I am getting include...

      When testing on fast.com...
      750-1.1Gbps download on my WiFi connected Pixel 5
      1.2Gbps upstream on my Ubuntu VM (1Gbps fiber link, never see above ~920Mbps on the speedtest direct to the BGW320-500)

      And Iperf3
      Almost always 250Mbps TCP ceiling no matter what (unless VM to VM, where it's 17Gbps)
      Totally wacky UDP test results (Usually with insane jitter, and pfSense traffic graph showing asymmetric flows when crossing interfaces)
      ubuntuserver-to-pf-udp.png
      asymetry.png

      Especially slow are file transfers from pfSense itself (e.g. downloading a pcap).

      mynettopo.jpg

      Here is my network. I have a pair of gigabit interfaces on Juniper and igb2,3 of the APU4c4 configured as LAGG.0. The LAGG is a trunk carrying all VLAN ID's for my network.

      Connecting the Proxmox host is a pair of 10G DACs (Intel 82599) to the 2x10Gb SFP+ uplink module on the Juniper.

      What am I doing wrong or misinterpreting?

      1 Reply Last reply Reply Quote 0
      • stephenw10S
        stephenw10 Netgate Administrator
        last edited by

        Testing to or from pfSense directly will always be slow. Especially on an APU where it's pushed to route 1Gbps anyway. Running iperf itself uses significant CPU.

        What do you see if you run iperf between different internal subnets with hosts in each?

        Steve

        1 Reply Last reply Reply Quote 0
        • First post
          Last post
        Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.