[nark]

Hello,

We have a self-hosted 32 slot, 128 tick CS:GO Server powered by a i7-7700K Processor running @ 4.20 Ghz with 8 GB RAM, Ubuntu 20 managed by Pufferpanel with 800 mbps bandwith behing a NAT (you should see this from the startup command in the logs)

Our game database (score, timeplayed, headshots, c4 defused, etc) and website (with fastdl) are on another server via LAN with a gigabit connection between them.

We redirect the incoming traffic via NAT (Port forward)

We run the server for the past year, with no major problems (except some yellow, red spikes at sv and var when full server and a lot of nades are thrown)

Our major problems started a few weeks ago when the server keeps randomly crashing (connection timed out) and restarting.

We installed the debugging solution and run https://crash.limetech.org/ to debug the errors but we got stuck and thus the reason for my post

So here are the errors:
https://ibb.co/XCzhT7B

And the detailed links (from top to bottom):

https://crash.limetech.org/ggtobvjrazf5
https://crash.limetech.org/xlkscm7gnhe7
https://crash.limetech.org/vddczpiol63k
https://crash.limetech.org/ehc42v3vhg6r
https://crash.limetech.org/ubqo5mwkjbqc
https://crash.limetech.org/2iynb7ri6wo3
https://crash.limetech.org/qsde2nkxzfnu
https://crash.limetech.org/dfbd2sw4xekj
https://crash.limetech.org/7kfkmmt3cl3f
https://crash.limetech.org/iatwic6ywmzv
https://crash.limetech.org/2ovj4e3ugdyp
https://crash.limetech.org/2g5a7p5hl636
https://crash.limetech.org/rch344d36ae6
https://crash.limetech.org/pjax5bbenwnl
https://crash.limetech.org/q4psflmzvx44
https://crash.limetech.org/hi25vsqpncll
https://crash.limetech.org/c7kq24vthyd3
https://crash.limetech.org/675voj457lw3

In the links you can see our startup line and errors

Many of them are related to:

engine.so + 0x1c5b1a
vphysics.so + 0x5703f
libc-2.31.so + 0x80873
libtier0.so!Plat_ExitProcess + 0x1b
linux-gate.so!__kernel_vsyscall + 0x9
libtier0.so!Plat_ExitProcess + 0x1b

But are very random.

The server does not crash when it's full (32/32) but more when it's 1/2 full so I don't think it's a resource problem (although in one error dump we saw something regarding low memory)

We just did a plugin install of the bellow plugin but it's to soon to know if things are going better
https://forums.alliedmods.net/showthread.php?p=2656975

Any ideas are much appreciated

Thank you!

[nark]

Quick update after this fix:
https://forums.alliedmods.net/showthread.php?p=2656975

We just hit 48 hours without a crash.

Maybe this post will help someone

[Bacardi]

HL2 SRCDS is old program and it require daily restart because of memory leaks.
It has been that since day 1.

So do not attempt run SRCDS server 24/7 without restart.

[dyxL]

Originally Posted by Bacardi HL2 SRCDS is old program and it require daily restart because of memory leaks. It has been that since day 1. So do not attempt run SRCDS server 24/7 without restart.

is quit command for enough for this?

[Bacardi]

I assume people have somekind watchdog what start program automatically when it close.

Yes, you quit program. And make sure it goes back up.