1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  2. Greetings Guest!!

    In order to combat SPAM on the forums, all users are required to have a minimum of 2 posts before they can submit links in any post or thread.

    Dismiss Notice

Some things I've noticed....

Discussion in 'UO Resources' started by gjohnson5, Dec 15, 2009.

  1. gjohnson5

    gjohnson5 Guest

    I'm going to keep some notes here and yall can read if you will. I have noticed after upgrading my OS from Windows XP x64 to Vista X64 that my ping times to UO servers have jumped from 50-60 to over 200!

    Yes the game is not very playable in this state. I have changed nothing about the hardware. The OS upgrade, and post sp2 hotfixes installed from http://kbupdate.info Also some issues on tcp autotune which was a feature introduced in Vista

    So I am checking the state of my TCP settings now
    .....
    Microsoft Windows [Version 6.0.6002]
    Copyright (c) 2006 Microsoft Corporation. All rights reserved.

    C:\Windows\system32>netsh interface tcp show global
    Querying active state...

    TCP Global Parameters
    ----------------------------------------------
    Receive-Side Scaling State : enabled
    Chimney Offload State : disabled
    Receive Window Auto-Tuning Level : normal
    Add-On Congestion Control Provider : ctcp
    ECN Capability : disabled
    RFC 1323 Timestamps : disabled



    C:\Windows\system32>netsh int tcp set global chimney=enabled
    Ok.


    C:\Windows\system32>netsh interface tcp show global
    Querying active state...

    TCP Global Parameters
    ----------------------------------------------
    Receive-Side Scaling State : enabled
    Chimney Offload State : enabled
    Receive Window Auto-Tuning Level : normal
    Add-On Congestion Control Provider : ctcp
    ECN Capability : disabled
    RFC 1323 Timestamps : disabled


    C:\Windows\system32>netsh interface tcp set global ecncapability=enabled
    Ok.


    C:\Windows\system32>netsh interface tcp show global
    Querying active state...

    TCP Global Parameters
    ----------------------------------------------
    Receive-Side Scaling State : enabled
    Chimney Offload State : enabled
    Receive Window Auto-Tuning Level : normal
    Add-On Congestion Control Provider : ctcp
    ECN Capability : enabled
    RFC 1323 Timestamps : disabled


    C:\Windows\system32>netsh interface tcp set global timestamps=enabled
    Ok.


    C:\Windows\system32>netsh interface tcp show global
    Querying active state...

    TCP Global Parameters
    ----------------------------------------------
    Receive-Side Scaling State : enabled
    Chimney Offload State : enabled
    Receive Window Auto-Tuning Level : normal
    Add-On Congestion Control Provider : ctcp
    ECN Capability : enabled
    RFC 1323 Timestamps : enabled


    C:\Windows\system32>netsh interface ip show global
    Querying active state...

    General Global Parameters
    ---------------------------------------------
    Default Hop Limit : 128 hops
    Neighbor Cache Limit : 256 entries per interface
    Route Cache Limit : 128 entries per compartment
    Reassembly Limit : 16761024 bytes
    ICMP Redirects : enabled
    Source Routing Behavior : dontforward
    Task Offload : enabled
    Dhcp Media Sense : enabled
    Media Sense Logging : enabled
    MLD Level : all
    MLD Version : version3
    Multicast Forwarding : disabled
    Group Forwarded Fragments : disabled
    Randomize Identifiers : enabled
    Address Mask Reply : disabled

    Current Global Statistics
    ---------------------------------------------
    Number of Compartments : 1
    Number of NL clients : 7
    Number of FL providers : 4


    C:\Windows\system32>netstat -e -p tcp
    Interface Statistics

    Received Sent

    Bytes 363659048 27727840
    Unicast packets 284948 213960
    Non-unicast packets 81176 5864
    Discards 0 0
    Errors 0 0
    Unknown protocols 0

    Active Connections

    Proto Local Address Foreign Address State
    TCP 127.0.0.1:27015 AMD64:49160 ESTABLISHED
    TCP 127.0.0.1:49160 AMD64:27015 ESTABLISHED
    TCP 127.0.0.1:49245 AMD64:49246 ESTABLISHED
    TCP 127.0.0.1:49246 AMD64:49245 ESTABLISHED
    TCP 127.0.0.1:49248 AMD64:49249 ESTABLISHED
    TCP 127.0.0.1:49249 AMD64:49248 ESTABLISHED
    TCP 192.168.1.5:50181 host36:http CLOSE_WAIT
    TCP 192.168.1.5:50182 host36:1380 CLOSE_WAIT
    TCP 192.168.1.5:50183 host36:1380 CLOSE_WAIT
    TCP 192.168.1.5:50185 159.153.233.79:5001 ESTABLISHED
    TCP 192.168.1.5:50232 gw-in-f138:http TIME_WAIT

    C:\Windows\system32>netstat -t

    Active Connections

    Proto Local Address Foreign Address State Offload
    tate

    TCP 127.0.0.1:27015 AMD64:49160 ESTABLISHED InHost

    TCP 127.0.0.1:49160 AMD64:27015 ESTABLISHED InHost

    TCP 127.0.0.1:49245 AMD64:49246 ESTABLISHED InHost

    TCP 127.0.0.1:49246 AMD64:49245 ESTABLISHED InHost

    TCP 127.0.0.1:49248 AMD64:49249 ESTABLISHED InHost

    TCP 127.0.0.1:49249 AMD64:49248 ESTABLISHED InHost

    TCP 192.168.1.5:50237 host36:http CLOSE_WAIT InHost

    TCP 192.168.1.5:50238 host36:1380 CLOSE_WAIT InHost

    TCP 192.168.1.5:50239 host36:1380 CLOSE_WAIT InHost

    TCP 192.168.1.5:50241 159.153.233.77:5001 ESTABLISHED InHost


    C:\Windows\system32>

    So I am satisfyied that chimney offload has been enabled at this point for UO and my interface is showing no errors.
    I don't see the MTU setting so I will set it

    C:\Windows\system32>netsh interface ipv4 set subinterface "1" mtu=1500 store=
    sistent
    Ok.


    Setting registry values ss per http://www.speedguide.net/read_articles.php?id=2574

    HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters

    EnableTCPA = 1
    SynAttackProtect = 1
    TCPNoDelay = 1

    The Receive Side Scaling should be enabled for dual core CPU's for parallel processing. Since this is a single core AMD 64 , I will disable...

    C:\Windows\system32>netsh interface tcp set global rss=disabled
    Ok.


    Reboot after registry changes and check connection...
     
  2. gjohnson5

    gjohnson5 Guest

    Just for giggles , I turned off my Windows 7 laptop and tried UOSA again and the latency was reduced to 60ms.....

    Also since typing this there was an update to my WNRD3300 wireless router and updated to the latest firmware

    http://kb.netgear.com/app/answers/detail/a_id/12689/kw/wndr3300 latest firmware/r_id/100109


    I'm not sure what Bill Gates has concocted for tcp compatibility in Windows 7, but I would suspect there are issues with the tcp stack in Win 7 that needs hotfixes for. Also the firmware update seems to certified for Windows 7 , yet the ping time increases. I will turn the laptop on and try again

    Bascailly I just wanted to post this to show that ping times may not be network congestion. They could be incompatibilities with your network products
     
  3. Beer_Cayse

    Beer_Cayse Guest

    and a lot of this is why I won't be upgrading to W7 any time soon. Along with the fact I (according to sources) need 3 separate licensed copies - 1 for each system I have. <sigh> Still fighting that one, tho.

    As for UOSA - either EC or CC - on XP/SP3 I have no real performance issues, only ones related to the EC client itself. <shrug>
     
  4. gjohnson5

    gjohnson5 Guest

    Totally false alarm. The next time I loaded UOSA , the ping times were back to 200ms.... Changed DefaultTTL to 64 in
    [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters]

    Reboot again
     
  5. gjohnson5

    gjohnson5 Guest

    Come to find out Microsoft has a hotfix for using the netsh command and causing some values to go out of range
    http://support.microsoft.com/kb/967224

    However I have already downloaded and installed that hotfix and my values were still incorrect set. SO it seems that hotfix is broken, so setting manually


    Changing values as per speedguide tcp/ip patch
    http://www.speedguide.net/files/vista/SG_Vista_TcpIp_patch.cmd
    [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\Tcpip\Parameters]
    TcpMaxDataRetransmissions changed to 7
    TcpTimedWaitDelay changed to 30
    TcpUseRFC1122UrgentPointer changed to 0
    TcpFinWait2Delay changed to default 120

    So need to remember if using the netsh command , I need to check the registry for these values... sigh rebooting
     
  6. gjohnson5

    gjohnson5 Guest

    Need to update all the changes I've made. I didn't notices other values being out of range from using netsh and have changed them.

    as per the speedguide patch mentioned above
    [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\ServiceProvider]
    CHO >> SG_Vista_TcpIp_Patch.reg [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\ServiceProvider]
    ECHO >> SG_Vista_TcpIp_Patch.reg "LocalPriority"=dword:00000004
    ECHO >> SG_Vista_TcpIp_Patch.reg "HostsPriority"=dword:00000005
    ECHO >> SG_Vista_TcpIp_Patch.reg "DnsPriority"=dword:00000006
    ECHO >> SG_Vista_TcpIp_Patch.reg "NetbtPriority"=dword:00000007

    Some more common settings
    [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters]
    EnablePMTUBHDetect was 255 now set to 0
    EnablePMTUBHDetect was set to 255 is now 1
    QualifyingDestinationThreshold was like 4 billion and is now set to 3.
    This seems to be an undocumented setting that deals with determining if a destination host is unreachable and hence killing the connection. Sometimes computers without default gateways may run into this problem.
    This registry setting fails to be found in technet documents. Also, this is new logic introduced by Microsoft in Vista that does not appear in XP :-(
    Tcp1323Opts was 2 now 3. The timestamps add 12 bytes to the frame of each packet , however from enterprise backups I still do, this option in long transmissions which saturate the network , this option increases speed.
    I will keep it at 3
    KeepAliveInterval and KeepAliveTime are both 4294967295 for some strange reason. Probably the above bug from using netsh. I will change to default technet values
    http://technet.microsoft.com/en-us/library/cc957548.aspx
    http://technet.microsoft.com/en-us/library/cc957549.aspx

    another reboot...
     
  7. gjohnson5

    gjohnson5 Guest

    To be honest the only reason I upgraded to Vista is because you CANNOT UPGRADE from XP to Win 7. But I failed to remember that in x64 land , there is no upgrade from XP to Vista either. The installer FORCES YOU to install a new copy of the OS. In x32 land , you can do a straight upgrade.

    This upgrade probably wasn't necessary, but my disk controller aha-29160 doesn't have signed drivers yet for Win 7. So I will have to ebay a 29320 or wait for drivers.
     
  8. Lefty

    Lefty Lore Keeper
    Stratics Veteran

    Joined:
    May 17, 2008
    Messages:
    894
    Likes Received:
    67
    This is true as far as buying box, but when talking new systems this is possible. For instance my distributors often offer systems that were downgraded from Vista to XP pro and have a free upgrade to Windows 7.

    So if the system had downgrade rights, the upgrade rights would apply to the original allotment.
     
  9. gjohnson5

    gjohnson5 Guest

    It's an old box that I will have to buy a copy of Windows 7 and install. I will need a disk controller as well before I can do that, but it's a trivial swap. Add the new disk controller and installer drivers , change the scsi cable to the new scsi controll and turn on box. Should be that simple

    ANyway Is there some issue with UOSA on vista x64? For some reason my chars will run and PHYSICALLY stop. Its not like I'm telling the character to stop running. It's like the server is telling my char to stop running. Maybe record a video of the game play and post it? So I cannot run away from paragons or anything fast really SO Ishnear , Ter Mer and Abyss are not really playable anymore. 2d seems fine.