New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
What makes a dedicated server host smooth game servers?
Hey guys,
I was just wondering what technology or specification usually makes game servers stutter and lag sometimes? I always thought it was the processor and the connection speed, but I have 2 servers from 2 different providers with the same processor & specs, and one of them hosts smoothly 24/7, while the other stutters for 2-4 seconds once or twice a game (a 30 minute span).
Could it be the network provider? And how can I find the smoothest providers?
Comments
What's stuttering? The game loop or your connection? What's your location in relation to the two servers? if you have one at ovh ca and one at hetzner, you're gonna see a difference.
The connectivity to your server is what's causing the issue here, it's not the processor since you're using both of the same.
One is Kansas City, one is Atlanta. And it's the game loop, everybody would mention that their game stuttered
What game?
Are the specs the "exact" same on each server? Does each server have the same port speed?
Choose the one you've got the best ping to and a network which isn't known for packet loss more or less frequently. Further CPU shouldn't be running @ 100% during a game and RAM not full. Also check if the disks I/O is maxed out da hole time.
I'm hedging my bets on it being shitty HDD
Please don't offer advice if you don't know what you're saying.
there's more to just "cpu" you need to worry about. even if the specs are the same, the quality of the disks for example can make a huge difference.
Initially I would have said that the location / routing difference was causing the lag, but since you mentioned others experiencing the same problem for that specific location, it is either a network wide problem (major latency increases / packetloss sometimes) or one of the hardware components experiencing a bottleneck.
In the case of disks, SSD is highly recommended for gameservers that require a lot of reads/writes.
Thanks for all the responses guys.
The servers are both running:
Dual Intel Xeon L5420
16 GB RAM
500GB Hard Drive
100Mbit connections
How can I find out which providers use networks with common packet loss?
Hard drives might be different. Run ioping on both and post the results. If there are more players playing on one server then the other, then it's not a fair comparison.
The game was Team Fortress 2 btw, and they hold a different amount of people but the Kansas City has no stutters ever (from few people to a lot), and Atlanta stutters even when there are only a few.
And I will run that, thanks
root@d1:~# ioping .
4.0 KiB from . (ext4 /dev/disk/by-uuid/06fd4460-878b-475a-9476-5b560246bdbd): request=1 time=131 us
4.0 KiB from . (ext4 /dev/disk/by-uuid/06fd4460-878b-475a-9476-5b560246bdbd): request=2 time=131 us
4.0 KiB from . (ext4 /dev/disk/by-uuid/06fd4460-878b-475a-9476-5b560246bdbd): request=3 time=125 us
4.0 KiB from . (ext4 /dev/disk/by-uuid/06fd4460-878b-475a-9476-5b560246bdbd): request=4 time=121 us
4.0 KiB from . (ext4 /dev/disk/by-uuid/06fd4460-878b-475a-9476-5b560246bdbd): request=5 time=126 us
^C
--- . (ext4 /dev/disk/by-uuid/06fd4460-878b-475a-9476-5b560246bdbd) ioping statistics ---
5 requests completed in 4.4 s, 7.9 k iops, 30.8 MiB/s
min/avg/max/mdev = 121 us / 126 us / 131 us / 3 us
This is the IOPing for the stuttering server. I can't access the other server right now (sorry), but are these numbers slow in general?
Can you show me smartctl -a /dev/sda as well?
Let me know if you need the rest of the log, it was pretty long
Where's all the counters stuff like power on hours and errors?
https://kb.acronis.com/content/9136
I know what I am talking about. Disk I/O Speed also came to mind while commenting on this, but all the person mentioned above was that they both had the same processors, so I debunked the processors in my comment automatically, since he did say they were both the same.
@Jonchun also note how I put "exact" in quotation marks.
.
Then I'll ask you read the OP again.
Let me requote your response. You neither used the word "exact", nor did you use quotation marks.
@Jonchun
Whatever though, I'm not going to argue with you over something so silly. You win.
Best way to test is have everyone who wants to join your server download a test file from nginx on your server at peak times and see how fast it goes, or how slow it goes. Hardware may be fine.
5400rpm disks are bad for gaming, and in my opinion, bad for everything else too (Except backups maybe)
From that smart data, it looks like hard drive failure, no? The multi zone error rate is really high.
There is no reason for a decent game server to require hard-real-time HDD access. It should be hoarding everything into RAM. Besides, even if a less-than-perfectly-healthy HDD misses a seek, 2-4 seconds is a heck of a long time for it to recover.
My money is on the network quality as the most likely distinguishing factor between the two servers. Sometimes these kinds of things can be traced to a piece of intermittently-failing networking hardware that glitches by randomly failing, realizing that it failed through a watchdog mechanism, and resetting itself.
Usually setting up the kernel right helps a lot