Crash while opening world / while ticking entity: "IllegalStateException: POI data mismatch: already registered at ev{x=##, y=##, z=##}"

Destroying all workplaces and killing all villagers in the world loaded in a previous snapshot seemed to resolve this issue, although some chunks appeared to have been deleted.

Same "point of interest" crash, this happens after optimizing world for the pre-release and clearing all caches.

ETA: Also happens if the world is NOT optimized.

Just happened to me too.  Stupid Snapshot!

Just had a ton of chunks regenerate on me in the 1.14 Pre-Release. World is totally fine on 19w14b but as soon as I update it regenerates the chunks.

Also had the POI error when loading certain chunks, would crash the server.

Crash will may occur for chunks that contain POIs (villager workplaces, beds, etc) that were created:

• in version before introduction of new villages (very probable)

• in this pre-release (large view range on freshly generated world increases probability)

We will try to re-create broken data, but villager assignments may be lost.

By "in this snapshot" do you mean in this pre-release, or in the snapshot before the pre-release? Or something else entirely? 🙂

It's spamming in the server console for me and regenerating chunks with chests/furnances.

--> attached log.txt.

This issue occurs only if there is "already" registered POI data upon restart. It means this issue never occurs if I remove the POI data (deleting the files from the poi world folder). But even with new computed data from this version pre-1, this issue will be happening again after restarting the server if there is any existing POI data, old one or not.

Crash happened to me when upgrading to 1.14-pre1:

Please DON'T remove POI data, it will only make this issue worse.

Confirmed for 1.14 pre-release 1

I have Got the Errors on my Server when Updating my Server to 1.14 Pre-release 1

Had this with a new world generated in the pre-release 1, so its not necessarily related to migrating old save data or optimizing.

This happened to me as well, and only affects to the entity, Villager.
Loading up other worlds from previous snapshots seems to work fine.

hopefully we will get this bug fixed in 1.14 pre-release 2

Confirmed. Some worlds from the previous snapshot can be loaded properly, but the world that already in pre-release 1 version can no longer be loaded again.

As soon as I loaded 1.14-pre1:

The game crashed whilst ticking entity
Error: java.lang.IllegalStateException: POI data mismatch: already registered at ev{x=-49, y=79, z=-17}

This may or may not be causing chunk corruption also, loading up a world with this issue appears to cause some chunks to completely break. Reloading the world reveals the chunks to have been completely reset.

It's worth noting that I have not yet experienced this happening in worlds without this issue, even when transitioning from 19w14b to 1.14-pr1. If you need a test world, I will be happy to provide one for you.

I can confirm that it will break your chuck(s).

Had 50% of a house deleted because of it, it complained about a POI being allready registered on 1727,79,-913 , which is an airblock just above my house ...

------------

[22:49:00] [Server thread/INFO]: Preparing start region for dimension minecraft:overworld
[22:49:04] [Server thread/ERROR]: Couldn't load chunk
java.lang.IllegalStateException: POI data mismatch: already registered at ev{x=1727, y=79, z=-913}
at aqc.a(SourceFile:72) ~[server.jar:?]
at aqc.a(SourceFile:43) ~[server.jar:?]
at java.util.stream.ForEachOps$ForEachOp$OfRef.accept(ForEachOps.java:184) ~[?:1.8.0_201]
at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193) ~[?:1.8.0_201]
at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193) ~[?:1.8.0_201]
at java.util.Iterator.forEachRemaining(Iterator.java:116) ~[?:1.8.0_201]
at java.util.Spliterators$IteratorSpliterator.forEachRemaining(Spliterators.java:1801) ~[?:1.8.0_201]
at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:481) ~[?:1.8.0_201]
at java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:471) ~[?:1.8.0_201]
at java.util.stream.ForEachOps$ForEachOp.evaluateSequential(ForEachOps.java:151) ~[?:1.8.0_201]
at java.util.stream.ForEachOps$ForEachOp$OfRef.evaluateSequential(ForEachOps.java:174) ~[?:1.8.0_201]
at java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234) ~[?:1.8.0_201]
at java.util.stream.ReferencePipeline.forEach(ReferencePipeline.java:418) ~[?:1.8.0_201]
at aqc.<init>(SourceFile:43) ~[server.jar:?]
at bxt.a(SourceFile:123) ~[server.jar:?]
at java.util.Optional.map(Optional.java:215) ~[?:1.8.0_201]
at bxt.a(SourceFile:122) ~[server.jar:?]
at bxt.a(SourceFile:102) ~[server.jar:?]
at bxt.d(SourceFile:76) ~[server.jar:?]
at aqa.a(SourceFile:165) ~[server.jar:?]
at bxo.a(SourceFile:113) ~[server.jar:?]
at us.f(SourceFile:416) ~[server.jar:?]
at java.util.concurrent.CompletableFuture$AsyncSupply.run(CompletableFuture.java:1590) [?:1.8.0_201]
at agf.h(SourceFile:135) [server.jar:?]
at agf.p(SourceFile:114) [server.jar:?]
at vd$a.p(SourceFile:438) [server.jar:?]
at vd.d(SourceFile:194) [server.jar:?]
at net.minecraft.server.MinecraftServer.p(SourceFile:698) [server.jar:?]
at agf.bf(SourceFile:99) [server.jar:?]
at net.minecraft.server.MinecraftServer.o(SourceFile:676) [server.jar:?]
at net.minecraft.server.MinecraftServer.a(SourceFile:467) [server.jar:?]
at net.minecraft.server.MinecraftServer.a(SourceFile:370) [server.jar:?]
at ug.d(SourceFile:218) [server.jar:?]
at net.minecraft.server.MinecraftServer.run(SourceFile:602) [server.jar:?]
at java.lang.Thread.run(Thread.java:748) [?:1.8.0_201]

-------------

Ticking on "trapped" villagers. Had to eliminate a villager in a player's automatic carrot farm in order to get the server to boot.

Confirmed. Crash when villagers are in sight.

Villagers(NPC) are not the problem, new 1.14 villages(location) are. Problem still occurs when you kill all NPCs
Did not have this problem with 1.13.2 villages, but a newly generated village is a problem.
When you delete the POI files you can play again, until you find a 1.14 village, then close your world and try to rejoin.
POI likely stands for Points of Interest, villages are a point of interest. A Pillager Outpost is likely aso a Point of Interest, but no crash at those locations.

Hope this helps.

No MuMe, points of interest are the blocks villagers can interact with: beds, bells, and all the working stations: cauldron, brewing stands, lectern, stonecutter, loom, grindstone, composter, cartography table, smithing table, barrel, smoker, blast furnace and fletching table.
The bug can occur with poi created either with this pre1, or with versions prior to 19w11a (the village overhaul).
But even if it seems to not trigger the problem at first, like boq said, do not delete the poi data, and be sure to have a world backup.

If I don't delete POI data I cannot play at all, as it crashes right after starting the world.
But my only reason to do this is to check for bugs with my Realms server before the full release. My Server uses multiple command systems and datapacks, so I need to test those before the server breaks after the automatic update of realms. As long as i delete POI data I can at least test the other things.

I may have found a better work round than removing the POI files.

I tried that and chunks kept regening every time I opened the game, even if I had removed blocks etc. Then I looked through posts on the mcr files and found one from way back that suggested renaming the level.dat_mcr to level.dat_mcr_old so I gave it a try and touch wood the game has been more stable since.

I have had one odd crash but not related to this but a fatal in the Java runtime environment which I will look into seperatly.

Hope this helps.

Just to join the mass.

Confirmed! Yesterday, when I first upgraded server to 1.14-pre1, I played normally. Today I cannot join the server (with server being crashed) because of this bug.

Can you mark my MC-147877 as duplicate if MC-147768 / MC-147783 / MC-147830 are marked?

My case was this:

In the morning I started MC to play for a few minutes. New version. Backup and load. I noticed there are now guardians around the monument and other fish in the ocean around. Then I went home and switched the game off. When I loaded my world when I came back from work, I found my house gone, only one side still there, surroundings seemed OK too. I checked with a bit of digging that the basement is gone too. best described by MC-147768.

@boq  Will the fix of this bring in the end any data loss to existing worlds prior to the villages overhaul?

What does being fixed in "Future Version - 1.14+" mean? Does it mean I have to wait until the full release of 1.14 rather than just the next pre-release?

maybe on 1.14 pre-release 2

"Future Version - xx" is a common name for the strict next upcoming version not labeled on Mojira yet. Will be 1.14-pre2 in this case.

Confirmed for Pre Release 1

I also had the same problem opening a world created in previous snapshot versions. It usually opens the 1st time in Pre 1, but the second time the crash occurs. It seems that the crash is always related to the villager entity in certain chunks created in previous versions of snapshots. If you drop a villager with an egg in the creative mode in these chunks, the crash also occurs instantly. As in the logs posted here, the villager entity is always referred in these errors.

• Entity being ticked -
  Details:
  Entity Type: minecraft: villager (bird)
  Entity ID: 222
  Entity Name: Dorfbewohner
  Entity's Exact location: -756.53, 66.00, -852.33
  Entity's Block location: World: (-757,66, -853), Chunk: (at 11.4,11 in -48, -54; contains blocks -768,0, -864 to -753,255, -849), Region : (-2.2, contains chunks -64, -64 to -33, -33, blocks -1024.0, -1024 to -513.255, -513)

By deleting the villager entity mentioned in the logs with nbtexplorer it is possible to enter the world without the crash, but some chunks are lost.

Unfortunately the issue is not fixed in pre2.

[14:37:27] [Server thread/INFO]: Starting minecraft server version 1.14 Pre-Release 2
[…]
[14:37:36] [Server thread/ERROR]: Failed to load POI chunk
java.lang.IllegalStateException: POI data mismatch: already registered at ev{x=15, y=79, z=-1}
        at aqc.a(SourceFile:81) ~[server.jar:?]
        at aqc.a(SourceFile:44) ~[server.jar:?]
        …

Edit: although, the aforesaid chunks are not being overwritten this time.

Edit: error message only appeared on first startup. Second try, no error.

Working as intended, corrupted data needs to be regenerated, but that does not destroy whole chunk.

OK, Mr. Bok, I'll try it, but I'm holding you responsible if Skelly, Bones, Ice Runner, Zenith, Rusty, Bedrock, Leonard, Otis, Lars, Jellie, Jeb_, Toast, Rueben, Drumstick's cousin Drumstick, Chocolate, Gabe, Goldie, or any of my other pet are harmed by this!!

!

If you test anything other than release without backups, it's your own fault.