Fredrik Liden <flide..ranslate.com> wrote:
> Woah, I think I'll try not to go outside of Cayenne. That's a cool
> solution though.
> Did you try the optimistic locking and found it to be to slow? is there
> any risk of dead-locks?
No, I also use optimistic locking. The strategy below is only good for
regularly-scheduled updates, not for real-time updates.
Optimistic locking has almost no performance issues (it's just attaching
additional qualifiers to each update and delete request).
Optimistic locking doesn't actually lock anything in the database, it just
guarantees that the operation fails if the data has already changed.
That's why it's optimistic locking -- it optimistically assumes that any
update will succeed, and you simply have to perform recovery if the
operation fails.
If it fails due to an OptimisticLockException, it means that something
external to your program updated the record since you last selected it.
-Mike
This archive was generated by hypermail 2.0.0 : Fri Feb 18 2005 - 14:54:37 EST