[Bug Report] Additional Server Crashes on Startup if Both a Running Server and the Additional Server Have the "logfile" Dvar Enabled

Hadi77KSA

So, apparently T4 has this problem with additional servers crashing on startup since update r3755, as per this topic: Problems Hosting Multiple Servers by Balrog85
Upon some testing, the problem seems to occur in the following cases:

2 MP
2 CO-OP
2 ZM
1 ZM & 1 CO-OP

but doesn’t occur for 1 MP & 1 CO-OP/ZM.

I managed to bypass this problem by simply editing the cfg file of the additional server and setting logfile to "0" (check edit2 for an alternative solution), and reported the fix to DJseanyB who was looking for a fix to the problem:

Upon investigating what update r3755 changed for T4, I’ve found this which I believe is relevant for this problem:

Rename old console log files with .001, .002, etc.

My guess is that the additional server crashes upon trying to rename console.log (CO-OP, ZM) or console_mp.log (MP) found in PLUTONIUM_PATH_OF_SERVER\storage\t4\main that’s in use by the already-running server, but I don’t have the skill to verify if that’s what’s actually happening.

I came across the fix when I noticed that starting servers without executing a cfg file made the servers start with no problem (aside from not having the dvars set that the cfg file sets, even manually started a map on both servers without either of them crashing), noticed that the additional server crashes when manually executing the cfg file while there’s a server running properly, then skimmed through the dvars in the cfg file which I believed might’ve been problematic, and finally, through a process of elimination by commenting out dvars and uncommenting them, I came to the conclusion that only one server from the same app may run with the logfile dvar enabled after testing with all its possible values.

To answer some questions which might get asked:

I had no custom scripts/textures installed for T4, as I don’t play that game.
I’ve used cfg files with different names.
I’ve used different ports.
I’ve tried making sure the g_log dvar in the cfg file is set to different names between the servers.

Edit (2024/05/03)

I reckon that using separate Plutonium installs for each server could also work as an alternative workaround from the user’s side for this problem. I will not elaborate further on this, as I consider this to be something only an advanced user would be capable of accomplishing successfully.

Edit2 (2024/08/07)

Today, Discord user DrSTaTTiK mentioned a dvar called g_console_log which determines the name of the console.log/console_mp.log file. I tested if setting the value of the dvar to a unique name per server circumvents the problem of additional servers closing when logfile is enabled on multiple servers, and it worked! I suggest using that as the solution to this problem. Write it in the server cfg file and use a unique name per server.

set g_console_log "console.log"                 // This should be unique per server.

Balrog85

Thanks. I never would've thought to try that either.

Edit: Yep, can confirm logfile 0 has fixed all my servers. Thanks again!