ThinkerQAQ

NOTE

Java OOM Examples and Diagnosis

How to distinguish heap retention, allocation pressure, Metaspace, direct/native memory, and thread-related out-of-memory failures.

JavaCreated Updated 1 min readhistorical

This is a historical learning note and may contain outdated or incomplete understanding.

1. OOM Is a Symptom Class

OutOfMemoryError does not always mean the same resource is exhausted. Possible categories include Java heap, class metadata, direct/native allocation, native thread creation, or VM-specific resource limits.

2. Heap Retention vs Allocation Pressure

A leak/retention problem means live reachable data grows unexpectedly. An allocation-pressure problem can allocate enormous temporary volume even when most objects die correctly.

A heap dump and dominator/retained-size analysis are useful for retention. Allocation profiling/JFR is often better for excessive churn.

3. Common Retention Sources

  • unbounded caches/maps/queues;
  • listeners/callbacks never removed;
  • ThreadLocal values on long-lived pool threads;
  • class-loader leaks;
  • request/session objects retained by static structures.

4. Native/Process Memory

If RSS grows while Java heap remains healthy, inspect direct buffers, thread stacks/count, native libraries, allocator behavior, Metaspace, and JVM native memory.

Do not increase -Xmx blindly: that can reduce native headroom and make a container OOM sooner.