Server operators, what are you running?
gamester_5
Join Date: 2008-04-17 Member: 64094Members
<div class="IPBDescription">Just curious</div>What is your equipment specs and what are you running for server software, and what bandwidth pipe are you on (up/down)?
Just curious what is out there and see if we can build a second machine. Also would like input on what equipment would be needed to build a beast of a server for NS2.
Also please list your server game name. Thanks in advance
Just curious what is out there and see if we can build a second machine. Also would like input on what equipment would be needed to build a beast of a server for NS2.
Also please list your server game name. Thanks in advance
Comments
Our Specs: see my Sig. 3 assigned cores of one Intel Core i7 920 (2,6 GhZ), 2.5 GB DDR3 1600 MhZ RAM, Windows Server 2008 R2 with SP1. no special Server-Software just put HLDS Update-Tool on an Vanilla updated Windows Server 2008 R2 OS. using NS2-GMOvermnd v4 and a Server-Restart Routine. (NS2-Monitor)
Bandwith: 100mbit down / 10 mbit up
Server runs well so far, minimal lags...
Greetz
XuXu - exit1.at
CPU: Intel i5 2400
MOBO: MSI H67ms-E43 with the ever famous Sandy Bridge Chipset :-)
HD: Hitachi 2TB Sata III drive with 6GBps
Memory: 8GIGs Gskill RipJaws x DDR3 1333
PS: Cooler master 600W Power supply
OS: Windows Server 2008 Enterprise
HLDS Update-Tool
NS2-GMOvermnd v5
Bandwidth is 25 mbit Down / 25 mbit Up
Server name: Team #156 - West Coast
Server runs great unless there are 100 hydra, 20 turrets and burning DI everywhere :-)
What are the server specs? Software used? Bandwidth?
I am going to start monitoring it now since I saw an earlier post saying that when reaching 10+ players CPU usage is 100%. I am curious to see if that is true.
MOBO: MSI H67ms-E43 with the ever famous Sandy Bridge Chipset :-)<!--QuoteEnd--></div><!--QuoteEEnd-->
Someone is going to have fun rebuilding it when the replacement boards come out.
That would be me! :-) Newegg has already contacted me about a replacement board so I am going to do a cross ship so I will have the new board in hand and swap out. Maybe have the server offline for 30 minutes.
I am only using the SATA III ports so it really didnt impact me but I dont want a faulty board down the road.
It was recalled the day after I had placed my order :-) It actually has been a very good board so far.
Do post and let everyone know what you see, I'm interested in see what sort of CPU/RAM usage people are seeing to help decide if it is feasible/cost-effective to put up an NS2 server (I really want to but if it is still quite RAM/CPU heavy then it may have to wait a while).
God knows, I don't run it. All I know is that I played on it the other day and it was great. Joined later on and it was the same.
Well with looking at Resource Monitor it really only uses one of the 4 cores, once in a while it will access one of the other cores but very rare and thats on a full server with TON of stuff laid out (Hydra's, Turrets)
It will go from 2% (Idle) to around 27% at maximum with 12 people on the server. Once you get 8-10 people on it tends to max out core 1.
If usage could be spread among the other cores I think performance would increase. Or even go heavier on memory usage to increase performance
I have nothing else running on this server so it is all NS2. Memory usage is not to strong either.
The trick seems to be to restart the server daily. I've had the experience that the longer you let the server run, the more there is very annoying "jittering".
Hardware is virtualized Core i7 920 on Windows 2008 R2.
It will go from 2% (Idle) to around 27% at maximum with 12 people on the server. Once you get 8-10 people on it tends to max out core 1.
If usage could be spread among the other cores I think performance would increase. Or even go heavier on memory usage to increase performance
I have nothing else running on this server so it is all NS2. Memory usage is not to strong either.<!--QuoteEnd--></div><!--QuoteEEnd-->
The overall CPU usage is a meaningless number for NS2. It's all single threaded, so the only thing that matters is the per-core CPU usage.