Does anybody has a more detailed answer at this question ?

All that exists in the documentation 2.1.7 is:
increment - generates identifiers of type long, short or int that are unique only when no other process is inserting data into the same table. Do not use in a cluster.

Thanks, Cristian.

hibernate assigns the identifier, as opposed to you or the db. the identifier is incremented each time and you cannot do this in a cluster because app1 will try to insert a row w/ an identifier property that has already been used by app2.

hibernate team, are there non-obvious technical challenges to a "distributed increment" identifier ? is there just not enough demand for this?

Hi Dennis,

Do you know the algorithm that Hibernate uses to increment the id ?
I mean, does it use a select max(id) to determine the start point or what ?

Thanks, Cristian.

Hi everybody,

I have found into Hibernate in Action book (page 91) the following explanation:

At Hibernate startup, this generator reads the maximum primary key column value of the table and increments the value by one each time a new row is inserted.
The generated identifier is of type long, short, or int.
This generator is especially efficient if the single-server Hibernate application has exclusive access to the database but shouldn’t be used in any other scenario.

Now everything is clear regarding this subject.

See you soon, Cristian.

Right, the "highest value" is held somewhere within the JVM, with a lock around it. If some other JVM (or anything else) had write access to that table it wouldn't work.