> We have been trying to resolve our SQL Server deadlocking issue without any
> luck. CAS keeps throwing deadlocks. Everything I'm seeing online says that
> our problem may be SQL Server specific due to the way they do locking

I've tracked down our problem in PostgreSQL, and the root problem is
the same: platform-specific locking semantics,
http://archives.postgresql.org/pgsql-general/2010-06/msg00804.php.
That said, it's clear that _all_ platforms can deadlock on the
insert/update behavior described in the thread above; some platforms
like Oracle simply handle it _much_ better.  PostgreSQL is bad, and
SQL Server seems somewhere in the middle.  I'm going to post some
recommendations to cas-dev in the design of the JPA storage layer that
might help mitigate this behavior going forward.

If anyone would like to evaluate a tool that was invaluable in
reproducing deadlocks in our environment, please review
https://filebox.vt.edu/users/serac/pub/cas-deadlock-gen.py.

M

-- 
You are currently subscribed to [email protected] as: 
[email protected]
To unsubscribe, change settings or access archives, see 
http://www.ja-sig.org/wiki/display/JSG/cas-user

Reply via email to