Different cache concurrent strategies for root entity and its collection (Hibernate with EHCache)?

Given example from Hibernate docs and modifying it so that root level entity (Customer) is read-only 开发者_如何学Pythonwhile one of its collections (tickets) is read-write:

@Entity
@Cache(usage = CacheConcurrencyStrategy.READ_ONLY)
public class Customer { 
    ... 
    @OneToMany(...)
    @Cache(usage = CacheConcurrencyStrategy.READ_WRITE)
    public SortedSet<Ticket> getTickets() {
        return tickets;
    }
    ...
}

Would collection of tickets get refreshed when accessing customer from cache?

If you modify one of the tickets of a given Customer somewhere else, yes. Why don't you test this? Let's assume Cutomer#1 has Ticket#1 and Ticket#2 in its tickets collection. I ran this code:

// loading in first session
Session session1 = HibernateUtil.getSession();
Transaction tx1 = session1.beginTransaction();
Customer c1 = (Customer) session1.load(Customer.class, 1L); // loads from db and puts in cache
for (Ticket ticket : c1.getTickets()) { // caches q2742145.Customer.tickets#1
    System.out.println(ticket);
}
Ticket ticket = (Ticket) session1.load(Ticket.class, 1L); // doesn't hit the db
ticket.setName("foo"); // do some change on Ticket#1
session1.save(ticket);
tx1.commit(); // Ticket#1 gets updated in the db and the cached association invalidated

// loading in second session
Session session2 = HibernateUtil.getSession();
Transaction tx2 = session2.beginTransaction();
Customer c2 = (Customer) session2.load(Customer.class, 1L); // hits the cache
Set<Ticket> tickets = c2.getTickets();
for (Ticket ticket2 : tickets) { // reloads tickets from db
    System.out.println(ticket2); 
}
tx2.commit();

Which shows the collection of tickets gets "refreshed".