Using jmap and Eclipse Memory Analyzer (MAT) to Diagnose Java OutOfMemoryError

When a Java application suffers from an OutOfMemoryError, the root cause must be identified; after learning to use the jmap command in a previous article, this guide shows how to analyze the resulting heap dump with Eclipse Memory Analyzer (MAT).

MAT is not a universal tool but can parse common heap dump formats like Sun/HP/SAP HPROF and IBM PHD; see the official documentation for details.

OutOfMemoryError typically arises from two scenarios: (1) memory leaks where dead objects are not reclaimed, and (2) memory overflow where live objects exceed the allocated heap, often due to improper -Xmx / -Xms settings or long-lived object references.

Generating Heap Information with jmap

jmap -dump:live,format=b,file=E:/jmap/heap.dump pid

The command creates a binary heap dump file (e.g., heap.dump) in the specified directory.

Importing the Dump into MAT

Open MAT and load the generated .dump file to start analysis.

Generating Analysis Reports

MAT can produce several reports, including Histogram, Dominator Tree, Top Consumers, and Leak Suspects, each visualized with charts and detailed object information.

Histogram

Lists classes, object counts, shallow heap size, and retained heap size, helping to identify which objects occupy the most memory.

Dominator Tree

Shows which threads and their associated objects dominate memory usage.

Top Consumers

Displays the largest memory‑consuming objects and their distribution.

Leak Suspects

Automatically highlights potential leak sources; in the example, a dark region accounts for 62% of a 79.7 MB heap, indicating suspicious code.

By navigating to List Objects → with outgoing references , you can inspect the reference chain of the suspected class.

Further analysis can involve viewing Path To GC Roots (excluding weak references) to trace why objects remain reachable.

The classloader/component "org.apache.catalina.loader.WebappClassLoader @ 0xa34cde8" occupies 19,052,864 (22.80%) bytes. The memory is accumulated in one instance of "java.util.HashMap$Entry[]" loaded by "<system class loader>".

This indicates that the Tomcat WebappClassLoader and a large HashMap are major memory consumers, which is normal unless custom classes dominate the heap.

MAT also provides views such as Shortest Paths To the Accumulation Point and Accumulated Objects by Class , revealing that java.util.HashMap instances are heavily retained.

Overall, the tool helps you understand heap composition, locate leaks, and guide code or architecture optimizations.