Master Transaction Locks in Spring Data JPA: Pessimistic vs Optimistic

1. Overview

In this quick tutorial we discuss how to enable transaction locks for custom query methods and predefined repository methods in Spring Data JPA, covering lock types and lock timeout settings.

2. Lock Types

JPA defines two main lock types: pessimistic lock and optimistic lock.

2.1 Pessimistic Lock

When a pessimistic lock is used inside a transaction, the entity is locked immediately and released on commit or rollback.

2.2 Optimistic Lock

Optimistic lock does not lock the entity immediately; the transaction stores the entity state and assigns a version number. When updating, the stored version is compared with the current version, and if they differ an OptimisticLockException is thrown.

Other strategies such as timestamps, hash values, or checksums can also be used.

3. Enabling Transaction Lock on Query Methods

Use the @Lock annotation on the target query method and specify the desired LockModeType .

Example:

@Lock(LockModeType.OPTIMISTIC_FORCE_INCREMENT)
@Query("SELECT c FROM Customer c WHERE c.orgId = ?1")
List<Customer> fetchCustomersByOrgId(Long orgId);

To lock predefined repository methods (e.g., findAll or findById), declare the method and annotate it with @Lock:

@Lock(LockModeType.PESSIMISTIC_READ)
Optional<Customer> findById(Long customerId);

If a lock is enabled without an active transaction, a TransactionRequiredException is thrown. If a lock cannot be granted and the conflict does not cause a rollback, JPA throws LockTimeoutException without marking the transaction for rollback.

4. Setting Transaction Lock Timeout

When a pessimistic lock cannot be obtained immediately, JPA throws LockTimeoutException. To avoid this, specify a lock timeout using @QueryHints:

@Lock(LockModeType.PESSIMISTIC_READ)
@QueryHints(@QueryHint(name = "javax.persistence.lock.timeout", value = "3000"))
Optional<Customer> findById(Long customerId);

Refer to ObjectDB articles for more details on lock timeout scopes.

5. Conclusion

This tutorial covered the different transaction lock modes, how to enable them in Spring Data JPA, and how to configure lock timeouts. Applying the appropriate lock in the right place helps maintain data integrity under high concurrency. Use pessimistic locks for strict ACID compliance and optimistic locks when concurrent reads are acceptable.