SRCDS Steam group


Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Choke assistance
#1
Sad 
Hello. Since i didnt continue my old post a while ago ill start a new one with updated data.

Ive recently upgraded the kernel on the specified server, hoping that it would solve my trouble, but wrong again.

Spec:
Intel Dualcore 3Ghz
2 GB Ram
Ubuntu 8.01 / 2.6.26.8-rt16

sys_ticrate 2500 / Pingboost 3 / 25 slot

Iam running the server with the realtime scheduler as suggested in an an other thread in here, and iam also using the idler.

Server is running smooth and stable at 1000fps, not terrible load on the cpu. Pings are looking fine.

To the trouble again. Iam using a fibre connection provider and i have a steady connection at 90Mbit up and down. On this connection i have 2 100tic css servers running without any trouble. Each server has its own connection, no router, only a switch.

The hlds server runs smooth with an updaterate of 30. Everything above this creates chokes around 100 for the clients.
When i try to use what i think its the optimal value (101,101,25000) chokes are unbareble. No loss, no ping raising. Just choke.

Ive tried increasing sys_ticrate up to 10000, 8000, 4000 and still no change. Ive tried both pingboost 2 and 3 and no change. Since my 2 srcds servers works all fine, this feels really strange. Same hardware in all 3 servers. Same connection.

I could really use some good advice here. I must have missed something, or something in my config is terrible wrong. Or else, my servers arent compatible with hlds.

Thanks
/Mike

sv_voicequality "5"
sv_alltalk "0"
sv_enableoldqueries 1
sys_ticrate 2500
sv_maxcmdrate 101
sv_mincmdrate 50
sv_minrate 0
sv_maxrate 10000 //want 25000
sv_minupdaterate 10
sv_maxupdaterate 30 //want 101
fps_max 0
sv_unlag 1

This is all my net settings in the config. Is something missing?
Reply
#2
sv_maxrate 25000 should fix it.
Reply
#3
With sv_maxrate at 25000 i still cant raise the sv_maxupdaterate. Even if i try raising it to 50 the choke starts to come. At 15 players the choke can go as hight as 100 when spawning and in smoke and flash.
Reply
#4
Then it is your internet.
Reply
#5
I thought about that to, but wouldn't my css servers choke as well? None of the has any signs of choke, neither has my dods that are running at the same connection.

Really strange.
Reply
#6
wowwen Wrote:I thought about that to, but wouldn't my css servers choke as well? None of the has any signs of choke, neither has my dods that are running at the same connection.

Really strange.

What's your CPU usage? I think since you run that many servers, it could cause this.
Reply
#7
I run each gameserver on separate physical servers. So the one running old hlds, thats the only server running on that mashine.

But all servers are identical. Same hardware and same kernel and dist.
Reply
#8
Ahh okay. I get it.

Does it get chokes whenever there is bounch or just like a few?

And 1,6/cz is not like Source/Orangebox when it comes to rates. Very various stuff. - Therefore it might act differently than the other servers, probably a configuration issue somewhere. I don't see anything wrong with you server.cfg. Maybe a FPS booster?
Reply
#9
FPS booster on linux other that the commandlineparameter?

For the post: lots of slots for the cpu. Connection shouldn't be the problem.
Reply
#10
The chokes comes mostly in spawn, and where there is lot of instant action, a rush for example or smoke and flash. Dust2 chokes terrible when you run from CT spawn up the hill. Iam not the real player here, just the administrator so i dont really now whats normal but according to the players, this is a non choke map.

The server has been running full today.
Ive monitored the cpu usage using top and it newer passes 50% usage.

This with ticrate 2500 and pingboost 2. Do you mean its a weak processor for 25 slots?
Reply
#11
I forgot he was on Linux, my fault.
Reply


Forum Jump:


Users browsing this thread: 2 Guest(s)