JAVA · HARD CODEX

Java Interview Questions — Hard

Q: Detail the lifecycle of HotSpot JVM thread synchronization locks (Lock Inflation).

Mark Word: An object's memory header contains a "Mark Word" storing locking state bits. Stage 1: Biased Locking (Deprecated/Removed): Thread ID is written into the Mark Word. Subsequent locking attemp

Q: Compare JVM Garbage Collection algorithms: Serial, Parallel, G1, ZGC, and Shenandoah.

Serial GC: Single-threaded; pauses all application threads ("Stop-the-World"/STW) for both young and old collections. Ideal for tiny memory footprints. Parallel GC: Multi-threaded young and old collec

Q: Explain memory barriers and compilation instruction reordering under the JMM.

Instruction Reordering: Compilers, JVM, and CPU architectures reorder execution instructions to maximize pipelining and cache efficiency, provided single-thread execution outcomes remain unchanged. Me

Q: Detail Classloading resolution, initializing sequences, and Class.forName() vs ClassLoader.loadClass().

Execution order: 1. Loading: Reads bytecode binary streams. 2. Linking: Verifies bytecode structure, prepares static fields with default values, and optionally resolves symbolic references. 3. Initial

Q: How do you trace and diagnose memory leaks, thread starvation, and CPU spikes?

Memory Leaks: Generate heap dumps using jmap -dump:live,format=b,file=heap.hprof [pid]. Analyze references using Eclipse Memory Analyzer (MAT) to identify leak suspects retaining massive heaps (retain

Deep runtime internals, memory models, distributed design, concurrency failure modes, and architectural decisions.

50 Theory Questions7 Implementation Folios5 Free Model Answers

THEORY QUESTIONS & SOLUTIONSShowing 5 of 50 questions

Q1Detail the lifecycle of HotSpot JVM thread synchronization locks (Lock Inflation).

Mark Word: An object's memory header contains a "Mark Word" storing locking state bits.
Stage 1: Biased Locking (Deprecated/Removed): Thread ID is written into the Mark Word. Subsequent locking attempts by the same thread skip synchronization operations completely.
Stage 2: Lightweight Locking: If another thread attempts acquisition, the lock inflates to lightweight. The JVM uses CAS to copy the Mark Word to the lock-seeking thread's stack. If it succeeds, the lock is acquired. If contention occurs, the thread spins briefly.
Stage 3: Heavyweight Locking (Inflated): If contention persists, the lock inflates to heavyweight. The Mark Word is rewritten to point to an OS-level monitor (ObjectMonitor). Threads are parked and queued in native OS wait states, incurring heavy kernel context-switching penalties.

Q2Compare JVM Garbage Collection algorithms: Serial, Parallel, G1, ZGC, and Shenandoah.

Serial GC: Single-threaded; pauses all application threads ("Stop-the-World"/STW) for both young and old collections. Ideal for tiny memory footprints.
Parallel GC: Multi-threaded young and old collection. High throughput, but causes significant STW pauses.
G1 (Garbage-First) GC: Divides the heap into equal region blocks. It targets regions containing the most garbage first, performing incremental, concurrent collections to meet configurable maximum pause times (-XX:MaxGCPauseMillis).
ZGC: A scalable, low-latency concurrent collector using colored pointers (storing metadata in reference bits) and load barriers (intercepting references to update object coordinates on-the-fly). STW pauses are sub-millisecond, independent of heap scale (supports terabytes).
Shenandoah: Similar to ZGC, uses concurrent evacuation barriers to compact the heap concurrently, minimizing pause times.

Q3Explain memory barriers and compilation instruction reordering under the JMM.

Instruction Reordering: Compilers, JVM, and CPU architectures reorder execution instructions to maximize pipelining and cache efficiency, provided single-thread execution outcomes remain unchanged.
Memory Barriers (Fences): Hard CPU instructions inserted by the JVM to enforce memory ordering:
- LoadLoad: Prevents subsequent reads from reordering before previous reads.
- StoreStore: Prevents subsequent writes from reordering before previous writes.
- LoadStore: Prevents subsequent writes from reordering before previous reads.
- StoreLoad: The strongest fence; forces all previous writes to main memory before subsequent reads are evaluated.

Q4Detail Classloading resolution, initializing sequences, and Class.forName() vs ClassLoader.loadClass().

Execution order:
1. Loading: Reads bytecode binary streams.
2. Linking: Verifies bytecode structure, prepares static fields with default values, and optionally resolves symbolic references.
3. Initialization: Executes static initializer blocks (static {}) and assigns real values to static fields.
Class.forName("MyClass"): Loads, links, and initializes the class, running static blocks immediately.
ClassLoader.loadClass("MyClass"): Only loads the class; linking and initialization are delayed until the class is instantiated or referenced.

Q5How do you trace and diagnose memory leaks, thread starvation, and CPU spikes?

Memory Leaks: Generate heap dumps using jmap -dump:live,format=b,file=heap.hprof [pid]. Analyze references using Eclipse Memory Analyzer (MAT) to identify leak suspects retaining massive heaps (retained size).
Thread Starvation/Deadlocks: Generate thread dumps using jstack [pid] or jcmd [pid] Thread.print. Examine thread states to locate deadlocks or threads blocked indefinitely on monitor lock acquisitions.
CPU Spikes: Profile CPU hot paths using async-profiler or JProfiler to capture on-cpu stack traces, identifying performance-intensive methods and loop bottlenecks.

Unlock the remaining 52 Java (Hard) questions

You've completed the 5 free sample questions. Get unrestricted lifetime access to every question, model answer, implementation challenge, and all 27+ technologies for a single payment.

Unlock Everything — Already Enrolled? Sign In

₹399 India / $9 International · One-time settlement · Zero subscription

CROSS-DIFFICULTY NAVIGATION

Continue Preparing Java

EASY LEVEL

Foundations & Core Concepts

MEDIUM LEVEL

Mechanisms & Practical Patterns