Alert for hung database -- ie enq HW:

From: Patterson, Joel <Joel.Patterson_at_crowley.com>
Date: Wed, 21 Nov 2012 10:52:18 -0500
Message-ID: <C95D75DD2E01DD4D81124D104D317ACA1C7647C832_at_JAXMSG01.crowley.com>



11.2.0.3.3 on Solaris 10.

A database got stagnant due to a connection pooling issue, some parameters where changed in the app, and the app started working again for a short time, but only until the connections again began piling up behind the processes that were left dormant in the first place -- so the app team stopped and started the application, which of course left all the original oracle processes that were messed up to begin with holding locks and whatever other resources, and were not able to communicate back to the client. Restarting the database fixed all that permanently.

DBOptimizer did identify an enq: HW on a particular table that is used all the time, but otherwise was unable to go farther, (no info in tabs etc on any other window).

So now the question has become, can we monitor for this and get an alerted? If so, does anyone have any good ideas on it? We do have EM.

Thank you,

Joel

--
http://www.freelists.org/webpage/oracle-l
Received on Wed Nov 21 2012 - 16:52:18 CET

Original text of this message