Re: dba_extents

From: utkanbir <hopehope_123_at_yahoo.com>
Date: 12 Feb 2004 00:25:22 -0800
Message-ID: <f6c90ebe.0402120025.67bad648@posting.google.com>

Hi ,
I have found the reason.
This is one of the dump output:
*** 2004-02-12 09:08:49.119
Start dump data blocks tsn: 5 file#: 5 minblk 4840 maxblk 4840 buffer tsn: 5 rdba: 0xc9c9c9c9 (807/641481) scn: 0xc9c9.c9c9c9c9 seq: 0xc9 flg: 0xc9 tail: 0xc9c9c9c9 frmt: 0x01 chkval: 0xc9c9 type: 0xc9=unknown Hex dump of corrupt header 4 = CORRUPT
Dump of memory from 0x60000000007B4E00 to 0x60000000007B4E14 60000000007B4E00 C9C9C9C9 C9C9C9C9 C9C9C9C9 C9C9C9C9

[................]
60000000007B4E10 C9C9C9C9                             [....]

End dump data blocks tsn: 5 file#: 5 minblk 4840 maxblk 4840

All the block is filled with C9! I have more than 3000 blocks like this. All of them are inside an index tablespace .

Is it possible to find why oracle refers some of these blocks ? The block flag is unknown (c9) so how does oracle find these blocks? Is there any tablespace block list in a header or somewhere which oracle checks in order to get the free blocks in a data file? If so is it possible to see it by using a v$view or similiar?

Most importantly is it possible to overcome with this without recreating the whole tablespace?

Kind Regards,
hope

"Michel Cadot" <micadot{at}altern{dot}org> wrote in message news:<402a01f7$0$28152$626a14ce_at_news.free.fr>...
> "utkanbir" <hopehope_123_at_yahoo.com> a écrit dans le message de
> news:f6c90ebe.0402110103.6a618c_at_posting.google.com...
> > Hi ,
> >
> > I have some corruption messages related to blocks. I have checked the
> > dba_extents in order to find the extents of these blocks , but
> > unfortunately these blocks dont belong to any extent. How could this
> > be possible?
> >
> > This is the query that i run:
> > select * from dba_extents where file_id=5 and corrupt_block_number
> > between block_id and block_id+blocks+1
> >
> > Kind Regards,
> > hope
>
> What is your Oracle version?
> When do you get the corruption message?
>
> Imho, if a corruption occured if a free block then it doesn't matter as
> the block will be reformatted when it'll be used.
>
> Regards
> Michel Cadot
Received on Thu Feb 12 2004 - 02:25:22 CST