Re: Problem: Extra blocks allocated during Initial Index Creation

From: Todd A. Wood <tawood_at_cimteam.cob.cummins.com>
Date: 1996/07/01
Message-ID: <31D7D258.41C6_at_cimteam.cob.cummins.com>#1/1


Brian P. Mac Lean wrote:
>
> Todd A. Wood wrote:
> >
> > I am having trouble figuring out why Oracle is allocating extra blocks
> > on some of my indexes.
> >
> > <-snip->
>
> I canā€™t remember all the details but if I remember right, Oracle does
> some rounding up, adds a few extra blocks, if the number of blocks you
> specified are not what it likes to use for boundaries. A fellow DBAer
> did the leg work and figured it out once. Where only talking about a
> couple of blocks here though, itā€™s not like you'll need to buy another
> disk drive or anything.
>
> brian.maclean_at_teldta.com

The problem is not with disk space at all. Typically we size an application for three years worth of data. An application has it's own tablespace for tables and another tablespace for indexes. The goal is to get an installation script that does not have to be modified.

I have an Excel spreadsheet which automatically generates the tablespace, tables, and indexes DDL scripts. When Oracle adds a "few" blocks the previously sized initial extents will not fit into the tablespace which was created to hold them. Of course a 5-10 MB database application is not going to make an impact on a 2GB disk.

If this is an internal round-up by Oracle then I will just add an adjustment factor of 3 blocks for each index I create.

Thank you, minus the sacasim provided. Just wanted to confirm my findings.

-- 
=======================================================================
   CCCCC     Todd A. Wood                tawood_at_cob.cummins.com     
 CC          Senior Systems Analyst      Cummins Engine Company, Inc.  
 CC ummins   Corporate CIM Team          500 Jackson Street, MC 60202  
 CC          Manufacturing Engineering   Columbus, IN  47201           
   CCCCC     http://www.cummins.com/     812-377-4637                  
=======================================================================
Received on Mon Jul 01 1996 - 00:00:00 CEST

Original text of this message