This may or may not be related to the connection pool. That is
certainly issue that has been improved in the trunk, but the behavior
(if I remember right) was not something you would typically see in
testing, only stale transactions and such on servers that have been
live for a while.
This particular problems sounds like a good old deadlock with a thread
A waiting on thread B and thread B is trying to get read something
thread A has a lock on. With custom code it's impossible to tell
without reviewing the code...
On a side note, there have also been some improvements to service
engine job concurrency support in the trunk that may address these. Is
there a reason you are using the old release4.0 branch?
-David
On Apr 24, 2009, at 5:11 PM, BJ Freeman wrote:
need to upgrade the pool to what the trunck uses.
not an easy task.
can't find the fAQ on it
hopefully Jacques will jump in since he supplied the info last time.
mayo sent the following on 4/24/2009 4:04 PM:
I noticed that we were missing steps on enabling XA transactions
for MS SQL.
I followed the official install steps
(http://msdn.microsoft.com/en-us/library/aa342335(SQL.90).aspx) and
still
having the same error. I am now using the MSSQL JDBC 1.1 driver.
Should the profiler show the use of the XA procedures? Do I need
to set the
pool-xa-wrapper-class in <inline-jdbc>?
--
BJ Freeman
http://www.businessesnetwork.com/automation
http://bjfreeman.elance.com
Systems Integrator.
