Re: Online backup: Backup online redologs?

From: Charles Fisher <Charles.Fisher_at_alcoa.com>
Date: Thu, 17 May 2001 15:15:29 GMT
Message-ID: <Pine.GSO.4.31.0105170921570.1445-100000@unknown>

On Thu, 17 May 2001, Howard J. Rogers wrote:

> For that reason, the use of the trace file is the *recommended* way of
> restoring all controlfiles, becaue it cheats... containing the paths
> to all the datafiles, Oracle knows where to look for them. It then
> finds the highest SCN of any datafile, and uses that as the SCN to
> write into the header of the new Controlfiles. Net result: the
> controlfiles are re-constructed, with an SCN that means that a
> resetlogs is not required. It's a much 'cheaper' way of performing
> the recovery, and all prior backups and archives are still useable.

Here is what the Couchman 8i OCP Cert book has to say about the value of a controlfile trace:

"In general, you won't need to create many new control files this way, but in case you do, you know how to do it. Plus, you should understand how to do this for the OCP exam."

Notice that it does not say anything like "controlfile traces are critical components of a backup." It almost dismisses them as trivial.

This Oracle documentation is exceptionally poor. I knew when I first read it that the advice was bad.

In a related question, can the dbverify command be used to check the consistency of a hot backup of a redolog?

The behavior of this utility seems a little odd; even though my db_block_size is 8192, dbv will not run on the redologs unless I use a block size of 1024.

Here is the output of dbv on one of my hotbackup datafiles:

:/pkg/prdcrm/backup:root> dbv file=platinum.dbf blocksize=8192

DBVERIFY: Release 7.3.4.4.0 - Production on Thu May 17 10:02:57 2001

DBVERIFY - Verification starting : FILE = platinum.dbf Page 1 is marked media corrupt
Page 2 is marked software corrupt
Page 3 is marked software corrupt
...
Page 12798 is marked media corrupt
Page 12799 is marked media corrupt
Page 12800 is marked media corrupt

DBVERIFY - Verification complete

Total Pages Examined         : 12800
Total Pages Processed (Data) : 0
Total Pages Failing   (Data) : 0
Total Pages Processed (Index): 0
Total Pages Failing   (Index): 0
Total Pages Empty            : 0
Total Pages Marked Corrupt   : 12800
Total Pages Influx           : 0
-------------------------------------------------------------------------------

Since this datafile can be recovered successfully, I guess that corrupt pages are not a problem.

Here it is for the hot copy of the redolog:

:/pkg/prdcrm/backup:root> dbv file=redoprdcrm01.log blocksize=1024

DBVERIFY: Release 7.3.4.4.0 - Production on Thu May 17 10:07:57 2001

DBVERIFY - Verification starting : FILE = redoprdcrm01.log Page 1 is marked software corrupt
Page 2 is marked software corrupt
Page 3 is marked software corrupt
...
Page 10238 is marked software corrupt
Page 10239 is marked software corrupt
Page 10240 is marked software corrupt

DBVERIFY - Verification complete

Total Pages Examined         : 10240
Total Pages Processed (Data) : 0
Total Pages Failing   (Data) : 0
Total Pages Processed (Index): 0
Total Pages Failing   (Index): 0
Total Pages Empty            : 0
Total Pages Marked Corrupt   : 10240
Total Pages Influx           : 0
-------------------------------------------------------------------------------

I get similar corruption errors on datafiles from a down instance.

Charles J. Fisher - Consultant
Alcoa Davenport Works
(319) 459-2512 Received on Thu May 17 2001 - 10:15:29 CDT