I´ve a question about how to avoid a application deadlock in the following situation:
Having two applications (x and y) using two entities (a and b) in many-to-many relation.
x accesses a by primary key and then b from a. y accesses b by primary key and then a from b.
If x and y access the same entities at the same time we have a deadlock.
Since both applications are jms driven it is not a very big problem because we get a rollback an a try again. But we get a lot of error messages ...
I´m looking for a smarter solution. It would be very easy if it will be prossible to lock the entity a and its dependent b entities in an atomic operation. Y will then wait until x finished its transaction.
But how can I write such a finder method? EJBQL does not help?