RE: Backup Controlfile and ArchiveLog (Shutdown immediate leading to requiring open resetlogs?)

Now I'm interested.  

Anticipating a blow by blow description if you want help understanding what is happening, I have a couple questions:  

1. Did you shut down just the database or the whole machine?
2. Did you let Oracle complete the shutdown immediate, or did you kneecap any processes, especially including LGWR (log writer, whatever name it goes by on your release)?
3. Is any sort of write cache enabled on the devices supplying the persistent media for your redo logs or controlfiles?
4. Did you replace the current controlfile with anything else?
5. Did you use the recover command with . using backup controlfile?

Even a shutdown abort *should* proceed with a normal instance recovery unless the situation is not normal in a very fundamental way. From time to time precise situations lead to mythology that I have to convince my customers out of; whether such mythology is spawned by Oracle's competitiors or merely the uninformed, I like to nip it in the bud. Oracle's recovery model has been very, very solid since at least 6.0.36.something, so if you're got a variant situation I hope we can pin it down to a tightly defined case.  

Regards,  

mwf  

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Jeremy Paul Schneider
Sent: Friday, May 04, 2007 1:21 PM
To: Hemant K Chitale
Cc: oracle-l_at_freelists.org; Joel.Patterson_at_crowley.com Subject: Re: Backup Controlfile and ArchiveLog  

In my last email I was not talking about the blog post... that post was something entirely different that I tested last night.  

Monday night we tested with shutdown immediate. It did require us to recover the online log file and open resetlogs. If I have time this afternoon then I'll do it again and post the output...  

-Jeremy
   

--

http://www.freelists.org/webpage/oracle-l Received on Fri May 04 2007 - 14:02:49 CDT