Guava Cache: Deletion Strategy, Expiration and Common Pit...

TL;DR

• Scenario: Using Guava Cache for local caching in Java projects, need to understand deletion strategy, expiration mechanism and common pitfalls
• Conclusion: Guava Cache uses ‘lazy cleanup + LRU+FIFO’ strategy. Passive deletion and active deletion need to be used together
• Output: Complete deletion strategy explanation + expiration mechanism details + error quick reference

Guava Cache Data Deletion Mechanism

1. Passive Deletion

Size-based Deletion: When cache element count reaches the preset maximum capacity, automatically remove the least recently used entry according to LRU algorithm.

Time-based Deletion:

• expireAfterAccess: Entry automatically removed if not accessed within specified time
• expireAfterWrite: Entry automatically removed after creation or update exceeds specified time

Reference-based Deletion: When values are wrapped with WeakReference or SoftReference, they are automatically recycled by garbage collector when memory is insufficient.

2. Active Deletion

• Explicitly call invalidate method to delete single key
• Call invalidateAll method for batch deletion
• Listen for deletion events through RemovalListener
• Call cleanUp method for active cleanup of expired entries

Important: Lazy Deletion Mechanism

Guava Cache uses lazy deletion mechanism. “After expiration time, cache.size() still hasn’t decreased” because access/maintenance logic hasn’t been triggered, or cleanUp hasn’t been called.

Solutions:

• Supplement cache.cleanUp() in scheduled tasks or critical paths
• Or trigger recycling through access

Error Quick Reference

Symptom	Root Cause	Fix
size doesn’t decrease after expiration time	Lazy deletion not triggered	Call cleanUp() or trigger access
maximumSize reached but no eviction	LRU+FIFO strategy inconsistent with expectation	Print cache.asMap() for debugging
Memory usage continuously increases	No limit on maximumSize or expiration policy	Set reasonable parameters

Application Scenario Examples

• E-commerce platform product cache: Use expireAfterWrite to ensure price information updates regularly
• User session management: Use expireAfterAccess to automatically clean up long-inactive user sessions
• Hot data cache: Use size-based deletion to prevent cache from occupying too much memory