java - Do I need to call JPA flush to check database constraints when my application is clustered? -

i read jpa caches sql instructions improve performance:

jpa providers hibernate can cache sql instructions supposed send database, until commit transaction. example, call em.persist(), hibernate remembers has make database insert, not execute instruction until commit transaction.

i have java ee 6 application deployed glassfish cluster 2 instances. in application there race condition 2 singletons expensive queries , cache results in database table. they're doing same work , trying write same record, exception:

java.sql.sqlintegrityconstraintviolationexception: ora-00001: unique constraint (someschema.sometable_pk) violated 

i decided easiest way deal catch , ignore exception:

// in ejb container-managed transactions. public entity getexpensiveentity(int entityid) {     entity entity = entitymanager.find(entity.class, entityid);     if (entity == null) {         try {             result = expensivequeries();             entitymanager.persist(result);             entitymanager.flush();         } catch (sqlintegrityconstraintviolationexception ex) {             // other instance created result, it.             result = jpa.find(result.getid());         }     }     return result; } 

i think call flush necessary because otherwise sqlintegrityconstraintviolationexception won't occur until transaction ends somewhere ejb call stack, past catching , ignoring. correct, valid use case flush? there better way handle this?

reference

• correct use of flush() in jpa/hibernate
• a beginner’s guide flush strategies in jpa , hibernate
• jpa , cmt -- why catching persistence exception not enough?
• jpa - create-if-not-exists entity? (see top-voted answer)