[2.0.14] Crash after saving multiplayer game after the host ended it

[joshinils]

colonelwill ended the stream and closed his game.

I chose to save the map instead of exiting, which proceeded, but then the game crashed.

factorio-current_crash_after_saving_will.log

(240.93 KiB) Downloaded 15 times

I tried to open a copy of the saved file which resulted in another crash.

factorio-current_loading_wills_save.log

(25.91 KiB) Downloaded 16 times

I do have an unopened copy of the save file, but feel uncomfortable uploading it publicly to the forum.

[Genhis]

Hello, please include the save file and your blueprint library. If you don't want to share them publicly, send me a DM.

[joshinils]

After more than 6 hours connected, I crashed. I did nothing at the time.

factorio-current.log

(227.31 KiB) Downloaded 10 times

I know my pc may have some sort of issue.

[Loewchen]

See 119577

[joshinils]

factorio-current.log

(137.79 KiB) Downloaded 17 times

The exact same thing happened. Will quit the game, I saved it, and then it crashed.

It did generate a zip file, but:

11-04-2024, 01-28-41.png (24.97 KiB) Viewed 312 times

[joshinils]

factorio-current.log

(144.16 KiB) Downloaded 10 times

This time, I was using Will's savegame as "creative mode" to design rail blueprints.

I got a notice, that the saving in background did not succeed, and tried to save manually, a few times, since it didn't work.

I disabled the blocking saving, tried again, and it crashed the main process (as expected).

Each crash in the log is the same.

Edit:
It generated a *.tmp.zip file, each try, each 39.3MB, but the save I started from is 178.3MB
The preceding autosaves are also 178.3MB each.

[Rseding91]

You're having an awfully lot of crashes compared to the amount of crash reports we're getting automatically... it seems.. off. Like maybe your computer is failing/having issues? It's always possible that the issues are in Factorio's code somewhere but i'm starting to have doubts.

[joshinils]

Like maybe your computer is failing/having issues? It's always possible that the issues are in Factorio's code somewhere but i'm starting to have doubts.

I have had these hardware doubts for a while now. Nothing much I can do about it.

I am going to use the drm free version now, and try to benchmark the autosaves for one tick, to see where/if the error got introduced, or if maybe it's something else.

[joshinils]

So, I tried to figure out which autosave was the last good one.
Wrote a small bash script that runs a benchmark for one tick, and saves the output in different files.
Each Autosave is tested like this each 5 times. (and a bunch more during writing of the script)

Sadly, some have mixed results, where at least one log end in Goodbye, and at least one in some unexpected error. Sometimes the error are different for the same autosave, which as I gather indicates hardware issues rather than Factorio.

But, the overwhelming amount of errors I saw looked like this:

2024-11-04_18-18-55---_autosave4.zip---2024-11-05_01-03-12.log

(5.26 KiB) Downloaded 12 times

Code: Select all

Trying to make chunk at unreasonable position [26, 131107]

grepping for "unreasonable" results in 72 results out of 174 logs
24 autosaves end in error (124 logs)
3 autosaves are mixed (15 logs, 11 error, 4 ok)
7 are ok, 3 of which are autosaves, the other are regular saves by myself as sanity check (35 logs)

Other errors are all over the place, even on the same saves.

Weird to me is that the 3 ok autosaves are not at the beginning, or from a continuous range, it's number 30, 2, 3. but 1 is all errors. and mixed are 23, 24 and 26. The fatal Autosave was 19, but the oldest -also named 19- is also all errors.


Edit:
I did the same thing on two other smaller machines now, they both produced identical results, but some are different from my desktop pc. Each save has the same log each time and the errors seem to stem only from the savegames themselves (chunk at unreasonable position as above, or "Error RuntimeError.cpp:24: Decompression failed: not enough room in the output buffer.")

Conclusion: it's my desktop pc that has an issue, I just don't know what or where.

[joshinils]

Just to end this, I did a memtest on my ram, it failed, many many errors.

Tested each of my two sticks, one passes, one errors almost immediately.
I've ordered new RAM, and I'm just using the one good stick for now.