I run a server with a 20k by 20k block custom world. It was created in 1.13.2, but some areas haven't been explored since. Those areas are bugging out, and around 20% of chunks in them are being reset back to vanilla chunks.
I started noticing this bug in 20w17a, however, it may have been going on for longer. I see that in that snapshot's changelog, something was changed about how blocks are stored.
I've attached a glitched region file - I would recommend opening it in a new world, preferably superflat so you can see which areas are reset (and which ones are terrain that was reset on my server)
[03:00:37] [Server thread/ERROR]: Couldn't load chunk [639, 36]
java.lang.NullPointerException: null
[03:00:37] [Server thread/ERROR]: Couldn't load chunk [639, 37]
java.lang.NullPointerException: null
[03:00:37] [Server thread/ERROR]: Couldn't load chunk [639, 38]
java.lang.NullPointerException: null
[03:00:37] [Server thread/ERROR]: Couldn't load chunk [638, 40]
java.lang.NullPointerException: null
Note: No stacktraces are logged, possibly because the wrong logging methods are called (last parameter is Object
and not Throwable
)? The class logging this is called yf
in 20w18a (Mojang name: net.minecraft.server.level.ChunkMap
).
Linked issues
relates to 3
Attachments
Comments 4
Might be the same issue as MC-181511
Added a link to MC-181507 since it appears some chunks in the provided region file also don't have a DataVersion
. However, I am not seeing the same stack trace. Errors were logged which I have added to the description of the report, they look similar to the one in MC-181511, however because both don't have stack traces, it is difficult to tell if both are caused by the same underlying bug.
Could you attach the world (or a world demonstrating this issue) for testing purposes?