Server memory overflow crash

FiveM Server Development Server Discussion
, ,
1

Hello

We have a big problem with our server. In random times the luv_tcp9 and the fivem process (linux) starting to eat our memory (129 GB) and on 100% memory the server chrases. We tried to kick all players when the memory overflow starts, but no changes. We started to search for a buggy script, but when this problem started we dont updated nothing on the server in 2 days. Today crashed other friendly server with the same way.

What/Who can cause this problem?

OS: Debian 11
Build: 5402

Ps: We tried windows and the server crashed with the same diagnose.

2

Please provide some actual crash dump, heap trace, or similar from Windows. There’s no equivalent data on Linux that can show such.

Likely cause could be a novel denial-of-service vulnerability, but this is impossible to tell without info.

3

We need to migrate to windows first, but we will do it now. After first crash I will send you the dump file.

4

Hi,
Same crash today morning.
FXServer filling up memory and crash down.

OS: WS2019
Build: 5371

5

Why no dump file? Why no heap trace? Why the weird old (EOL, even) server build?

What are you expecting us to do with that little info?

6

Currently waiting for new crash for gathering requested information.

1 Like
7

Hi,

We have successfully made an ETW trace during a crash. Unfortunately the fxserver process did not generated any dump files in the crashes folder.

OS: Windows Server 2019
Build: 5562

Thanks

2022-06-28_16-12-39_Administrator.etl (59.3 MB)

8

@nta we uploaded the dump file.

9

Have you found a solution?

10

Disable “apparmor” it is automatically put by debian 11.