Strange behavior (server lag related)



  • Two server different operating system same hardware (both running as virtual machines on esxi host).

    First server running on Windows 2008R2 assigned 4 cores @ 3.4GHz and 8GB of ram.
    Second server running on Ubuntu 12.04.1LTS assigned 2 cores @ 3.4GHz and 4GB of ram.
    Both servers are using the VMXNet3 network cards connected to a 100/100Mbit connection.
    Chivalry configuration is identical.

    The chivalry server running on the Windows server is experience more lag and higher ping then the one running on the linux server.
    Would be nice if anyone could shed some light on this matter… or if they noticed the same results.

    Server #1: [server:2tnkrhih]31.44.232.92:7777[/server:2tnkrhih]
    Server #2: [server:2tnkrhih]31.44.232.93:7777[/server:2tnkrhih]



  • try entering this command

    netsh interface tcp set global autotuninglevel=disabled

    reboot and test.



  • @dragunov:

    try entering this command

    netsh interface tcp set global autotuninglevel=disabled

    reboot and test.

    I don’t think that will do much for the performance since Chivalry is using pure UDP and setting autotuninglevel to disabled basically just sets the TCP receive window to a fixed size (64KB).

    As far as I know there isn’t much you can tweak when it comes to UDP traffic on windows, it seems more likely its related to how the chivalry server is utilizing the resources. And it seems like it prefer Windows 2003/WINE over Windows2008R2.



  • If you’re certain it’s pure UDP then yes there’s no point, if you’re not certain then give it a go.



  • Sgt.TAtari,

    I am having the same lag and higher ping on windows 2008 r2 server. At first i i thought was cpu but the amd quad core is around 35% when the server is full 32/32. I checked the bandwidth graphs around 3 mb up/down out of 50. Anything else I can check on config files maybe?

    Thanks

    Davee.


Log in to reply