We are experiencing an increase when we reach 450 users/500 users online. This problem only happens when we exceed and maintain this amount of users. When we reach 500 exactly, the server exceeds almost 25% of packetloss and makes it impossible to play and stay on the server.
The dedicated hardware does not even reach 50% usage, having, we are not under ddos attack, nor any other reason why it may cause this problem.
Hardware:
12th Gen Intel(R) Core™ i9-12900, 24 cores.
128 GB RAM
1TB SSD
1Gbps dedicated internet network.
Actually we already tried to disable many random things that can cause this, such as: anticheat, voip, different scripts, etc, etc. And we are still without any solution.
We also tried temporarily disabling proxy + anti-ddos mitigation, but this had no effect either.
We tried several other things to diagnose the problem, taking out different profilers, pcaps, etc, etc, and we do not see any anomaly, neither in the network traffic, nor in the server scripts (profiler).
We tried to start the server without resources stremeable, and it kept happening.
Today we formatted the computer from Ubuntu 22.04 to Debian 10, and it keeps happening. At the same time we installed txAdmin to have a log of “svMain/svSync/svNetwork”, which I attach its graph.
Environment: OS: Linux 5.10.0-21-amd64 CPU Model: Intel Gen Intel® Core™ i9-12900 CPU Stats: 16c/24t - 0.8 GHz CPU Usage: 39% Memory: 19% (24.29/125.60) Version: FXServer-master v1.0.0.6231 linux
We don’t know if it’s a hardware problem, or a wrongly programmed resource inside the server. But it is completely unfeasible to track this problem. Attached are pictures of the current performance of the server.
I think it is not a hardware problem, but the packet loss would be with the dedicated one, not only with the FXServer…
I also attach a graph of WinMTR that we have not lost any packet at times when in the FXServer if we lose (+6 hours of WinMTR).
Host - %
Sent
Recv
Best
Avrg
Wrst
Last
192.168.0.1 - 0
38555
38555
0
0
3
0
No response from host - 100
7787
0
0
0
0
0
No response from host - 100
7787
0
0
0
0
0
No response from host - 100
7787
0
0
0
0
0
No response from host - 100
7787
0
0
0
0
0
host39.181-89-51.telecom.net.ar - 6
31413
29569
5
37
1819
116
200.0.17.173 - 0
38575
38575
4
11
26
10
No response from host - 100
7787
0
0
0
0
0
No response from host - 100
7787
0
0
0
0
0
146.99.235.45.in-addr.arpa - 1
38566
38564
4
11
30
11
________________________________________________
______
______
______
______
______
______
Image from the server-side where we lost only 1 packet with the server-dedicated.
I know that with all this you will not be able to give me any solution, I am just trying to look for help to see where to trace this problem. I don’t know what else to do to try to get a diagnosis of where the high packet loss is coming from when reaching 350 players, 400 players, 500 players, etc, etc, etc.
If there is a way to do it in Windows, which one is the best and how exactly is it done?
this is not the right resource monitor, you should be looking at the server side resource monitor via svgui which i am not sure if it even exists in linux