12-17-2004, 05:38 AM
Is anyone else out there having problems with severe choke on their servers when running HL2DM?
When one of my two HL2DM servers gets about ten players (or more, or course), I start to see very high levels of choke. When I connect as a player and turn on net_graph 3, I observe choke levels often in the range of 20-50 when more than ten players are on the server.
I've got a machine that should be able to handle HL2/Source dedicated servers with no problem. We have a rack-mounted dual-processor AMD64 Opteron 244 server with 2GB of DDR, routed through a Cisco switch to a fiber optic backbone (hosted at The Planet, in Dallas). I have checked the server remotely when it the choke is occurring, and the CPU and other vital resources are running at only 10% or less of capacity -- so I don't think I'm maxing out the hardware by any means.
I also run several other HLDS-based server instances on this machine, and none of them show any significant (usually 0 or 1) choke, even when the HL2DM servers are full.
I'm a reasonably experienced server operator , and have tried playing with all the obvious cvars (sv_maxupdaterate, sv_minupdaterate, sv_minrate, sv_maxrate), but to no avail.
Here are my questions:
1) Is everyone else having this problem, or is it just me?
2) Is there a solution?
3) Is it a bug, and is the development team working on it?
You can see the server IPs to test them out and see what I mean here:
http://www.sunlitgames.net/serverstatus_public.php
Any help would be greatly appreciated.
[I&S] Pvt. John
http://www.SunLitGames.net
When one of my two HL2DM servers gets about ten players (or more, or course), I start to see very high levels of choke. When I connect as a player and turn on net_graph 3, I observe choke levels often in the range of 20-50 when more than ten players are on the server.
I've got a machine that should be able to handle HL2/Source dedicated servers with no problem. We have a rack-mounted dual-processor AMD64 Opteron 244 server with 2GB of DDR, routed through a Cisco switch to a fiber optic backbone (hosted at The Planet, in Dallas). I have checked the server remotely when it the choke is occurring, and the CPU and other vital resources are running at only 10% or less of capacity -- so I don't think I'm maxing out the hardware by any means.
I also run several other HLDS-based server instances on this machine, and none of them show any significant (usually 0 or 1) choke, even when the HL2DM servers are full.
I'm a reasonably experienced server operator , and have tried playing with all the obvious cvars (sv_maxupdaterate, sv_minupdaterate, sv_minrate, sv_maxrate), but to no avail.
Here are my questions:
1) Is everyone else having this problem, or is it just me?
2) Is there a solution?
3) Is it a bug, and is the development team working on it?
You can see the server IPs to test them out and see what I mean here:
http://www.sunlitgames.net/serverstatus_public.php
Any help would be greatly appreciated.
[I&S] Pvt. John
http://www.SunLitGames.net