When I was running WB 1.9.9 beta I couldn't get /wb fill to complete because it would throw errors about not being able to unload chunks.

With WB 1.9.10 beta the errors are gone, however I can see the available memory decrease as the fill service is running, until there is none left and the fill service will stop until I restart the server.

I'm on Spigot 1.14.1.

Here is a log of my startup where you can see the memory issue: https://pastebin.com/raw/FY5mJTTs

I am also having a similar issue. Paper Spigot 1.14.1, it runs fine and i've legit tried 2GB -> 8 GB by 2GB increments. Crashes every so often with no reason - its going perfectly then just crash. Is there a solution?

Having similar kind of trouble with 1.9.10. as described here: https://www.reddit.com/r/admincraft/comments/butudl/optimizing_pregeneration_for_114_and_big_worlds/

Tested on spigot 1.14.2 on the 1.14 compatible release and am having this exact same issue. Even changing the freq value to super low like 5 only slightly extends the life of the fill but then it still eventually completely crashes.

I'm going to suggest try the new PaperMC build for those running 1.14.2, update your server jar to #83 or above, several memory leaks where fixed which addresses many chunk/chunk cache issues, there have been MASSIVE updates in the past week alone.

PaperMC/Paper@5c7081f

This was a chunk related Paper issue not a WorldBorder issue. A developer on the Discord acknowledged these issues and they where worked on since release #56

The issue no longer exists here. Chunks should be fully optimized by release #85 from what I was told.

Running with JVM params: "%JRE_HOME%\java.exe" -d64 -server -XX:+UseConcMarkSweepGC -XX:+UseParNewGC -XX:+CMSIncrementalPacing -XX:ParallelGCThreads=7 -XX:+AggressiveOpts -Xms1G -Xmx8G -jar "!APP_HOME!\server.jar" nogui

/wb reload
/wb world fill 1000 208 false
/wb confirm

Results:

Average: 130-230 chunks/run | Running for 2 days

My issues were also solved by paper build 58, so I'm contented :)