Kevin-
On Sep 21, 2006, at 8:39 AM, Kevin Sutter wrote:
Hi,
Attempting to verify some of the locking mechanisms within OpenJPA
and I
don't see the SingleJVMExclusiveLockManager class available in our
source
tree... Oversight or not going to happen? Regardless, can you
give me a
scenario where this would have been useful? I guess if you're
assuming that
the only access would be via this single JVM, then you could safely
lock
in-memory without accessing the database. But, is that a very likely
scenario?
It is probably going to remain held-back. Regardless, you are correct
about the usage scenario: it is basically an in-memory lock manager
that assumes single-JVM access to the database (e.g., for a pure-java
file-based database like Derby).
The doc indicates that this is still a valid option for the
openjpa.LockManager setting:
sjvm: An alias for the
org.apache.openjpa.kernel.SingleJVMExclusiveLockManager
<../apidocs/org/apache/openjpa/kernel/
SingleJVMExclusiveLockManager.html>.
This lock manager uses in-memory mutexes to obtain exclusive locks
on object
ids. It does not perform any database-level locking. Also, it does not
distinguish between read and write locks; all locks are write locks.
Thanks,
Kevin
