RE: Puzzle - Abandoned Terminal.

From: Walker, Jed S <Jed_Walker_at_cable.comcast.com>
Date: Tue, 15 Feb 2011 16:45:37 +0000
Message-ID: <BAA6E28B6241F046AED1E62D8697516C2FCE99C1_at_COPDCEXMB09.cable.comcast.com>



Thanks Jared, yes, I realized after sending it that my comment was a bit "round about". A lock would only exist in a transaction, and if the transaction wasn't committed or rolled-back then a "shutdown normal" would do no good (whether it was holding a lock or not). Sorry, a bit off the trail and not precise on that comment. Newbies, definitely read up on the various types of shutdown... it'll save you grief one day.

Jared - I like your title, is there a test for that? ;)

From: Jared Still [mailto:jkstill_at_gmail.com] Sent: Tuesday, February 15, 2011 9:02 AM To: Walker, Jed S
Cc: ORACLE-L
Subject: Re: Puzzle - Abandoned Terminal.

On Tue, Feb 15, 2011 at 6:22 AM, Walker, Jed S <Jed_Walker_at_cable.comcast.com<mailto:Jed_Walker_at_cable.comcast.com>> wrote: It would sit, unless you have a profile (or semething else -FW timeout, etc) that would kill it.

Oracle takes care of deadlocks by killing one of the sessions. If the user holds a lock (and there is no deadlock) then it will just sit there until the session is killed (this is why they gave us "shutdown immediate")

Just to clarify for newbies out there, what I think you are referring to is that prior to the availability of 'shutdown immediate', shutting down the database when there were open transactions required a shutdown abort.

Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist Oracle Blog: http://jkstill.blogspot.com Home Page: http://jaredstill.com

--
http://www.freelists.org/webpage/oracle-l
Received on Tue Feb 15 2011 - 10:45:37 CST

Original text of this message