Graphics stops updating but game continues

[run4flat]

Hello everybody,

I've been playing Mineclonia with a few other members of my household and I've had issues on one of the three machines we use. Here are what I hope are the relevant system and game specs:

• Windows 11 Enterprise

• Intel UHD Graphics with 128MB memory

• Luanti 5.14.0

• OpenGL 4.6.0

• The Active Renderer describes itself as "4.6.0 / opengl / SDL 2.32.2"

• Not sure which version of Mineclonia, but I installed it early January. Now I see there is an update.

I'll download the update and report back if it seems to fix it, but in the meantime...

The problem is that while I am playing the graphics will suddenly stop updating. It has pretty much universally occurred while I am walking and not in combat (thank goodness). The graphics stop updating, but the audio continues to work and indicates when I am walking or stopped. This happened recently while I was playing with somebody else, a game hosted on their machine. My player was still around and I could see through their perspective that I could turn, walk, do actions, etc. I simply did not have the graphics updating on my machine. This issue occurs both when I play singleplayer and when I play in a world hosted on another machine. It seems to occur randomly, roughly once every two hours, but once or twice it has frozen twice in 15 minutes. I have read through the debug.txt file and it seemed too sparse to figure out what is going wrong.

For now, my only solution has been to Alt+F4 and close Luanti, then rejoin/restart the game.

Two questions: (1) Does anybody have any ideas about what might be causing this, how to trace it, etc? (2) Is there a more convenient way to restart my graphics besides exiting the game?

[Krock]

Is this issue limited to Luanti 5.14.0 or does it also occur on older versions? FYI, 5.15.0 is the most recent.

EDIT: I suppose this is another instance of buggy Intel drivers on Windows. You could place the software renderer opengl32.dll next to luanti.exe to circumvent this issue. That's however not a proper solution (like reinstalling a new driver).

1. Close Luanti
2. Use the setting opengl_debug = true
3. Wait for the issue to occur.

What would help us more

1. Use Luanti 5.15.0
2. Use the setting video_driver = opengl3 AND opengl_debug = true
3. Start Luanti in renderdoc and do reproduce the issue
4. Take a snapshot of the scene using PrintScreen (you can test this early to ensure it works)
5. Share the snapshot for inspection

[run4flat]

Krock, thanks for this. I just started playing in the beginning of January, so I can't say if this occurred with previous versions. I will start with the setting you suggest and see if I can reproduce the issue. Never heard of RenderDoc, so I'll give that a spin, too. Once I can get details about the bug, I'll upgrade to 5.15.0 and see if it is still an issue.

I'm coming to Windows from Linux, so I was hoping to manage most of my software through a package manager. Do you know how long it typically takes for the latest version of Luanti to be available via winget? I can download and run the Luanti installer myself to upgrade to 5.15, I was just curious.

[run4flat]

Hello Krock,

Still on 5.14.0.

I am pretty sure I could set those parameters in a configuration file, but I'm not quite sure where that configuration file lives, so I decided to go spelunking through the Luanti settings menu. After checking 'Show technical names' and 'Show advanced settings' I found the Advanced > Advanced settings. There, I have ticked the checkbox next to opengl_debug and set video_driver to opengl3.

I will keep you posted!

[run4flat]

Alright, I just triggered the issue. Still on 5.14. There are a ton more lines in the log file, but nothing here gives me a clear idea of where the real problem lies. The issue happened almost immediately after I decided to harvest some wheat. This was logged thus:

Code: Select all

2026-02-04 00:57:49: ACTION[Server]: singleplayer digs mcl_farming:wheat at (197,4,-311)

This was followed by 290 lines of this message (except that the seconds incremented a few times):

Code: Select all

2026-02-04 00:57:49: WARNING[Main]: Irrlicht: GL: 8246 824d API_ID_LINE_WIDTH deprecated behavior warning has been generated. Wide lines have been deprecated. glLineWidth set to 2.000000. glLineWidth with width greater than 1.0 will generate GL_INVALID_VALUE error in future versions

But there are gazillions of those lines throughout the debug.txt file, so that might be a red hearing.

At that point I closed the game and it logged

Code: Select all

2026-02-04 00:57:54: ACTION[Server]: [persistent_map] Player singleplayer left, data cleaned up
2026-02-04 00:57:54: ACTION[Server]: singleplayer leaves game. List of players: 
2026-02-04 00:57:54: ACTION[ServerStop]: Server: Shutting down

[Krock]

Thank you for testing.
glLineWidth with thickness 2 very likely originates from drawing the pointed node outline (setting selectionbox_width). Thus, it seems logical that it appears plenty of times in the log. Whereas that deprecation should be addressed in the future, I believe that it is either unrelated or just happens to be the final "trigger" of a deeper issue.
At the moment I can only suggest to reproduce it in renderdoc (needs the opengl3 driver, by setting) and/or to perform a system memory check e.g. by memtest86+ (assuming that the iGPU uses your system memory).

EDIT: Alternatively, use a higher resolution texture pack (e.g. 256px) and/or a higher quality shadow shader (see settings) to check whether it occurs faster (if it's a VRAM OOM issue).