[2.0.43] Crash saving (SignalIDBase.cpp:320: Unknown enum value: 21)

Place for things which are bugs but we have no idea how to solve them. Things related to hardware, libraries, strange setups, etc.
User avatar
distortions864
Fast Inserter
Fast Inserter
Posts: 128
Joined: Thu Apr 20, 2017 12:56 am
Contact:

[2.0.43] Crash saving (SignalIDBase.cpp:320: Unknown enum value: 21)

Post by distortions864 »

Logfile and save attached.
Was crashing during player join.
It stopped crashing, so maybe it was related to a specific player joining.
Attachments
factorio-previous.log
(21.47 KiB) Downloaded 34 times
_autosave108.zip
(31.79 MiB) Downloaded 31 times
M45 Science: A Factorio Community since 2017! https://m45sci.xyz/
User avatar
boskid
Factorio Staff
Factorio Staff
Posts: 3941
Joined: Thu Dec 14, 2017 6:56 pm
Contact:

Re: [2.0.43] Crash saving (SignalIDBase.cpp:320: Unknown enum value: 21)

Post by boskid »

Without reproduction steps there is not much that can be done here, provided save file passes consistency and the stack trace shows that one of circuit networks got a signal of a corrupted type and it would crash when trying to attempt any saving of the game in that state. This looks pretty much 1/0 magic.
User avatar
distortions864
Fast Inserter
Fast Inserter
Posts: 128
Joined: Thu Apr 20, 2017 12:56 am
Contact:

Re: [2.0.43] Crash saving (SignalIDBase.cpp:320: Unknown enum value: 21)

Post by distortions864 »

Interesting, thanks for taking a look!
That is really odd, because it crashed several times but it went away.
The server is a dell xeon, with ECC enabled.

Maybe someone was intentionally doing it?
M45 Science: A Factorio Community since 2017! https://m45sci.xyz/
eugenekay
Filter Inserter
Filter Inserter
Posts: 504
Joined: Tue May 15, 2018 2:14 am
Contact:

Re: [2.0.43] Crash saving (SignalIDBase.cpp:320: Unknown enum value: 21)

Post by eugenekay »

distortions864 wrote: Sat Apr 19, 2025 6:12 pm Interesting, thanks for taking a look!
That is really odd, because it crashed several times but it went away.
The server is a dell xeon, with ECC enabled.

Maybe someone was intentionally doing it?
ECC Memory reduces the rates of memory-read errors, and adds the ability to report errors if detected.... but silicon Ages with time. The CPU model is a Xeon E5-2680 v4 - first launched 9 years again in 2016. I recommend running a Memtest cycle for good measure - the results may surprise you.

It is improbable that a remote-crash while processing SignalIDs is feasible before the Map has been downloaded & executed by the Client.
Post Reply

Return to “1 / 0 magic”