[Linux] Port won't reopen on server crash

DZaleks · 2023-10-23T20:00:40.333Z

Hello there,

I switched hosting a few days ago and I noticed some issues. On Server crash (they’re caused when my anticheat pulls an update) the fx process cant be started again.

example on how it should work:
= Restart at 10 am =
= Resources Loading =
= Resources Done =
= Anticheat checks update =
= Update found =
= Fx process being terminated =
= Resources Loading again =
= Resource Done
= New update installed =

But now, after the Fx process is terminated it wont start again.

[          svadhesive] Authenticating server license key...
[          svadhesive] Server license key authentication succeeded. Welcome!
[    c-scripting-core] Creating script environments for sessionmanager
[           resources] Started resource sessionmanager
[    c-scripting-core] Creating script environments for monitor
[      script:monitor] [txAdmin] Resource v6.0.2 threads and commands set up. All Ready.
[           resources] Started resource monitor
[      net-tcp-server] Listening on socket failed - libuv error EADDRINUSE.
[      net-tcp-server] error on connection: address already in use
Could not bind on 0.0.0.0:30120 - is this address valid and not already in use?
[ citizen-server-impl] Error: Could not bind on 0.0.0.0:30120 - is this address valid and not already in use?
[ citizen-server-impl] 
> txaEvent "serverShuttingDown" "{＂delay＂:5000,＂author＂:＂txAdmin＂,＂message＂:＂Server restart (Crash).＂}"

This is printed in a loop until i manually kill the fx servers process and run the start shell again.

This issue is not related to any anticheat update since I seem to be the only one with this problem.
I think I may forgot some previleges?

server is started as root btw

DaveGaffer · 2023-10-23T23:29:10.853Z

I know with Linux, ports can be a little finicky. A process opens the port and keeps it open even after the program crashes. You can check to see what’s got the port tied up with the netstat command. If it’s not installed, Ubuntu and Debian have it in their repos in the net-tools package.

Simply type sudo netstat -lp | grep 30120 and you’ll see what’s currently tying up that port. An example from my sever:

$> sudo netstat -lp | grep 30120
tcp   0   0 0.0.0.0:30120   0.0.0.0:*  LISTEN  1385043/ld-musl-x86
udp6  0   0 0.0.0.0:30120   [::]:*             1385043/ld-musl-x86

This shows the binary chugging along hanging on to port 30120,

Hope this helps, though it doesn’t solve your problem, you can at least get a look under the hood.

DZaleks · 2023-10-26T08:05:39.386Z

i’ll check this when error appears next time.

DZaleks · 2023-10-26T22:11:51.657Z

DaveGaffer:

sudo netstat -lp | grep 30120

I did this before restarting, after crash and a few times after the error

Error: Could not bind on 0.0.0.0:30120 - is this address valid and not already in use?

This are the outputs. Are you able to read out any untypical behavior here?

image813×927 89 KB