Re: Fractured Block 8.0.5

Hi Jonathan,

I suspect that the disk is faulty. The buffer cannot have been reconstructed from redo, because this happened during a buffer read, not instance recovery. The fact that you did not get an ORA-1578 error indicates that the situation was resolved with a 'buffer read retry'. However, the tail value could never have been valid. Therefore this is not an in-flux buffer being written by another node of an OPS database, or a fractured block due to a system crash.

Check the hardware diagnostic logs for the disk, and if you find any errors or warnings, I suggest you have it replaced.

Regards,
Steve Adams

http://www.ixora.com.au/

http://www.oreilly.com/catalog/orinternals/

http://www.christianity.com.au/

On Fri, 10 Dec 1999 08:41:10 -0000, "Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> wrote:

>
>I came on this error message in an alert log
>recently, a couple of days after the event:
>
><begin quote>
>***
>Corrupt block relative dba: 0x26c085d6 file=155. blocknum=34262.
>Fractured block found during buffer read
>Data in bad block - type:6. format:2. rdba:0x26c085d6
>last change scn:0x0000.00970ace seq:0x2 flg:0x00
>consistancy value in tail 0x00000001
>check value in block header: 0x0, check value not calculated
>spare1:0x0, spare2:0x0, spare2:0x0
><end quote>
>
>
>When I did a block dump of the block (155,34262) it was
>perfectly sound (and I am fairly sure it would not have been
>subject to a drop/create cycle in between).
>
>My guess at the moment is that there was a sot
>corruption in the SGA that Oracle detected before
>writing the block out, so it re-read the block and
>applied redo to bring it up to date.
>
>Has anyone else seen this happen, and can
>anyone confirm my hypothesis.
>
>--
>
>Jonathan Lewis
>Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk
>
>
>
Received on Fri Dec 10 1999 - 20:59:53 CST