Mekanism

Mekanism

111M Downloads

server crash when loading chunk after #5505 fix, Fixed TE flooded console

ROMVoid95 opened this issue ยท 12 comments

commented

Please use the search functionality before reporting an issue. Also take a look at the closed issues!

Issue description:

after fixing the issue in #5505 the server was able to boot into world. After tp to base with known Mekanism builds the server crashes on chunk load and console is flooded with
[Chunk I/O Executor Thread-1/INFO] [Mekanism]: Fixed TE from ....

This server is running sponge and can confirm non-sponge resulted in identical error. my apologies if this is not mekanism related just wanted to be sure

Steps to reproduce:

  1. Same to reproduce form #5505
    2

Version (make sure you are on the latest version before reporting):

Forge: forge-1.12.2-14.23.5.2838
Mekanism: 1.12.2-9.7.7.378
Other relevant version:
Sponge: spongeforge-1.12.2-2825-7.1.6

If a (crash)log is relevant for this issue, link it here: (It's almost always relevant)

[gist/pastebin/etc link here]
Crashlog -> https://pastebin.com/z6eZMYS8
Lastest.log -> https://gist.github.com/ROMVoid95/9d91b152ef0ec160d6db8058424bbb04

commented

Hey I'm getting the same crash report as above and am wondering if there is a fix or if you could point me in the right direction on how to fix it. When ever a player goes to the coords in the crashreport it instant crashes the server. This happens with or without sponge. I included the log when the player joins, thats when it crashes.

EDIT
I believe this is an issue with mekanism. I loaded the world up on our test server (same as live server) but I removed mekanism from the test client and test server and logged in and I was able to load the world and see the base. Removing mekanism fixed the crash issue so it is relating to mekanism.

Forge Version: forge-1.12.2-14.23.5.2847-universal.jar
Mekanism Versions:
Mekanism-1.12.2-9.8.3.390.jar
MekanismGenerators-1.12.2-9.8.3.390.jar
MekanismTools-1.12.2-9.8.3.390.jar
Sponge version: spongeforge-1.12.2-2838-7.1.8-RC3963.jar

Log with sponge: https://gist.github.com/Trnrr/a016bb029a66679511ce7c9d762e8958
Crash Report with sponge: https://gist.github.com/Trnrr/1c8761146b972ceacba757eb8e2570f0

Log without sponge: https://gist.github.com/Trnrr/d2678d29f578367e68855dca666fc496
Crash report without sponge: https://gist.github.com/Trnrr/ab742526edd50d554bf59c514a529434

commented

@thiakil Is there anything I can do to help you try and solve this? I've been racking my head around this for the last few days and nothing Im doing is working. The only solution I've found is wiping the players chunks, which is pretty shitty.

  • I've tested downgrading to mek version 389, 387 & 381 and all of them instant crash when you connect still
  • Some weird stuff happened while testing. Sometimes after certain restarts/version changes the chunk magically became unglitched. Not sure what causes this. Wasn't able to pin point it. Just happened most of the time between versions 390 & 389
  • All of the crashes on the various versions show the same crash report

Is there anything I can do or any version of mek I could downgrade to until this is fixed? I would really appreciate a reply.

commented

Either wait until I look at it code wise, keep downgrading, or try remove all Mekanism pipes/tubes/cables from their chunks

commented

Ok thank you for the reply. I really appreciate it. I will keep downgrading versions and see if anything changes. If that doesn't change anything I'll try to mcedit the pipes/tubes/cables out of the base. Thank you

commented

sorry about that, didnt mean to close this

commented

I wasn't able to get it fixed. Even after removing the items from the chunk and downgrading the mod I am still getting crashes. I could have missed some pipes but I don't think I did since I was wiping the chunk and that players base was in 1 chunk.

Are the pipes/tubes/cables whats affected? I'm planning on banning them until this gets fixed (if ever). Also do by chance know in what version of mek did you start changing the:

From the stacktrace this looks related to recent(ish) changes that were made around getting transmitters (cables/pipes/etc) to correctly load/refresh their connections when they're added to the world.

I want to downgrade, but since none of my downgrading fixed I'm unsure of what version to do. I went back probably 8-10+ files. @thiakil

commented

It could be anything really, the crash reports offer no useful information unfortunately

commented

Show me the crashlog from without sponge. There is nothing in that stacktrace pointing (directly) to Mekanism

commented

The fixed TE from is a one time log when your saved data is being fixed to the new IDs

commented

Heyo!

We had a few reports with users of our modpack having crashes when porting their worlds from a version to a newer version which seems to have added the TE data fixer. But... something somewhere is causing the worlds to be non-usable.

Report: https://pastebin.com/pEUUPx0m

Converted Classes

image

We are using 1.12.2-9.7.9.380 in our beta release of the pack (yes it's old) but I have tested myself with 1.12.2-9.8.1.383 and the same issue is about. Removing Mekanism allows the world to load w/o any issues, I also went back and downgraded all our mods to versions in our 3.0.8 of the pack and only updated Mekanism to 1.12.2-9.7.9.380 and the issue still occurs so I narrowed it down to here.

I don't have any actual logs of the mod causing the issue, but as it's always constantly firing when then data-fixer is running. It would suggest either a party mod or this mod is not liking a change happening in the world. Could this be looked into somewhat, please?

I can supply a world download if you wish to help with the debugging.

Thanks,
SevTech Developer

commented

Hiya Kieron,

From the stacktrace this looks related to recent(ish) changes that were made around getting transmitters (cables/pipes/etc) to correctly load/refresh their connections when they're added to the world.

I have some ideas, and will try to get to it this weekend

commented

No idea what is going on, but given we are not actively developing 1.12 anymore, I am going to go ahead and close this, as my best guess is that it had to do with some weird race condition/halt in data fixers, and data fixers are not even available yet to mods in 1.15.