Re: Block Corruption in empty pages in Oracle 11g

From: ddf <oratune_at_msn.com>
Date: Sun, 15 Aug 2010 15:29:00 -0700 (PDT)
Message-ID: <8fd9c3c1-c3ee-42a6-ab5b-66f98aa4640a_at_z30g2000prg.googlegroups.com>

On Aug 15, 11:30 am, zigzagdna <zigzag..._at_yahoo.com> wrote:
> On Aug 14, 11:51 pm, zigzagdna <zigzag..._at_yahoo.com> wrote:> On Aug 14, 2:57 pm, Mladen Gogala <gogala.mla..._at_gmail.com> wrote:
>
> > > On Fri, 13 Aug 2010 21:32:32 -0700, zigzagdna wrote:
> > > > Can some one explain following ooutput from dbv;
>
> > > Have you tried the documentation?
> > > --http://mgogala.byethost5.com
>
> > Yes,  I have read the documentation. Based on the documenation; my
> > block corruption is in unused blocks; i.e.; thes eblocks are not being
> > used by current data and indexes. But, I get errors when doing delete
> > or truncate; so documentaion is not correct or dbv is provding
> > incorret information.
>
> Based on documentaion, I cannot figure out
> Questions remain why corrupt table I can do:
> 1. select without error.
> 2. insert without error
> 2. but truncare gives error
> 3. delete of all the rows gives error; one may be able to delete some
> rows; but do not know which ones.
>
> So either dbv output is misleading or I do not unserstand what does it
> mean?

Currently unused blocks are not necessarily above the highwater mark; they may have had data in them at one time and your delete from <table_name> may be scanning all allocated blocks for that table including the currently unused ones which would cause your error. Truncate processes every block allocated to the table, again providing reason for the error to surface. Simply because those blocks are empty now does not indicate in any way that they were never used to contain data. Thus, dbv is not misleading, it's your concept of empty (unused) blocks with respect to your situation.

David Fitzjarrell Received on Sun Aug 15 2010 - 17:29:00 CDT