RE: Oracle shutdown on Wintel

From: Mark W. Farnham <mwf_at_rsiz.com>
Date: Tue, 16 Feb 2016 14:25:48 -0500
Message-ID: <012901d168ef$d7d83bc0$8788b340$_at_rsiz.com>



I have always considered it a useful practice to stop Oracle before pulling the rug out from under it.  

Not that you really need to, because as was noted earlier in the thread crash recovery *should* work just fine.  

However, where a standard prelude to OS shutdown includes an Oracle shutdown (possibly immediate or abort); startup restrict; shutdown (normal), then the restart after reboot is as fast as it can been, and if an all storage snapshot or backup is taken before the OS restart you've got something to which no recovery need be applied.  

So plus one to Martin's practice. (You do probably need a bailout in case there is a long running transaction or rollback in progress that eat up too much of the planned patching window.)  

mwf  

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Herbener, Martin - Division of Engineering and Management Sent: Tuesday, February 16, 2016 12:14 PM To: Oracle-L_at_freelists.org
Subject: Re: Oracle shutdown on Wintel  

Daniel Hubler,  

The behavior you describe does indeed seem to be standard for Oracle on Windows (it's described further in 309230.1). As part of migrating to 12c we did some mitigation by implementing a Windows shutdown script that shuts down the Oracle Windows service before the OS actually quits. This was not my idea - there are other discussions of this out there in the www - but I didn't make a note of original writer.  

I can share the specific script/configuration we used if you want it.  

Martin    

Martin Herbener

Kentucky Department of Education

502 564 2020  

KDE on Facebook: <http://www.facebook.com/kydeptofed> www.facebook.com/kydeptofed

KDE on Twitter: twitter.com/KyDeptofEd  

--
http://www.freelists.org/webpage/oracle-l
Received on Tue Feb 16 2016 - 20:25:48 CET

Original text of this message