SPRING BOOT · HARD CODEX

Spring Boot Interview Questions — Hard

Q: How does Hibernate batch writes and what settings unlock it?

Requires: hibernate.jdbc.batchsize=50, ordered inserts (orderinserts=true), IDENTITY generation DISABLES batching for inserts (use SEQUENCE w/ pooled optimizer hi-lo style). Versioned data enables bat

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

50 Theory Questions5 Free Model Answers

THEORY QUESTIONS & SOLUTIONSShowing 5 of 50 questions

Q1How does the Spring context startup actually sequence, and what breaks it?

Scan/parse bean definitions → BeanFactoryPostProcessors (property placeholders, config class parsing - CGLIB enhancement of @Configuration enforcing proxyBeanMethods semantics) → instantiate non-lazy singletons in dependency order → BeanPostProcessors wrap proxies → SmartLifecycle.start().
Breakers: circular deps (now fatal by default; allow-circular-references escape hatch), eager @PostConstruct doing IO, static initializers touching beans.
Lazy init defers but shifts failures to request time - trade-off articulation expected.

Q2Explain @Configuration(proxyBeanMethods) semantics and the cglib interlude.

Full mode (default): config class is CGLIB-subclassed so internal @Bean calls intercept, returning the SAME singleton instead of invoking method - preserves container semantics.
lite mode (proxyBeanMethods=false / @Bean on non-@Configuration classes): methods execute literally → new instance per call; allowed when no intra-config references; faster startup, used heavily in native/AOT guidance. Bug signature: two "singletons" unequal - explain mechanism to ace this.

Q3What is the transaction synchronization + JPA flush ordering dance in services?

Within tx: Hibernate flushes BEFORE query execution touching affected tables (FlushMode.AUTO queries flush pending changes to keep results consistent), and always at commit.
@TransactionalEventListener(AFTER_COMMIT) sees committed state but detached entities - must pass ids/data explicitly, not managed graphs.
Pitfall: enqueueing message inside tx reading uncommitted rows elsewhere → phantom reads for consumer; design consumers idempotent or move reads post-commit.

Q4How does Hibernate batch writes and what settings unlock it?

Requires: hibernate.jdbc.batch_size=50, ordered inserts (order_inserts=true), IDENTITY generation DISABLES batching for inserts (use SEQUENCE w/ pooled optimizer hi-lo style).
Versioned data enables batched updates; statement rewrites via reWriteBatchedInserts=true (MySQL/pgjdbc) collapsing multi-values.
Verify: datasource-proxy/sql statement count metrics before/after - prove the win numerically.

Q5What are the concurrency semantics of @Transactional across threads/reactive boundaries?

Transaction bound to thread via ThreadLocal (DataSourceTransactionManager) - spawning threads inside a tx method runs code OUTSIDE it silently.
Reactive: ReactiveTransactionManager propagates via Reactor context; mixing JDBC (blocking) inside reactive chains needs schedulers bounding blocking work.
Async bridging patterns: pre-load ids inside tx, process outside, final update tx with optimistic checks.

Unlock the remaining 45 Spring Boot (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

One-time settlement () · Lifetime access · Zero subscription

CROSS-DIFFICULTY NAVIGATION

Continue Preparing Spring Boot

EASY LEVEL

Foundations & Core Concepts

MEDIUM LEVEL

Mechanisms & Practical Patterns