Direct answer

If many chunks constantly unload and reload near moving players, the server may do extra I/O and entity lifecycle work instead of saving time.

How to check it

  1. Observe memory growth, loaded chunk counts and storage activity when players move in/out of a small region.
  2. Check per-world chunk options such as delay-chunk-unloads-by, min-chunk-unload-count and min-chunk-unload-fraction.
  3. Identify any plugins, force-loaded chunks or farms deliberately holding areas active.

What to change

chunks:
  delay-chunk-unloads-by: 10s  # example documented current default
  1. Start with the world defaults and change only one supported option after a measured reason emerges.
  2. A longer unload delay may reduce load/unload churn but keep more chunks resident in memory.
  3. If memory pressure is high, investigate persistent chunk tickets and view distance before broadly forcing delayed unloads.

Software and version notes

Compatibility and safety

These settings live in config/paper-world-defaults.yml and may be overridden per-world. Chunk unloading, chunk disk saves and world autosaves have different performance costs.

Verify the result

Track chunk residency, memory use, I/O, and MSPT over the same gameplay route; verify farm/teleport behavior remains correct.

References