Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle 9i datafile sizes compared with 8i

Re: Oracle 9i datafile sizes compared with 8i

From: Austin <austint_at_vsnl.com>
Date: Thu, 8 Jul 2004 22:55:00 GMT
Message-ID: <I0K2C3.446@news.boeing.com>


We are using LOCALly managed tablespaces for the 9i database as against DICTIONARY managed for 8i. could this change in extent management clause cause an inflation of datafiles?
After giving a smaller initial extent, but an 'autoextend on' clause, I end up with a final tablespace size about 4 times the size of the 8i tablespace. I have been using the output of /usr1/oracle/tablespace.sql to make the comparisons

"Joel Garry" <joel-garry_at_home.com> wrote in message news:91884734.0407021326.3068e27b_at_posting.google.com...
> "Austin" <austint_at_vsnl.com> wrote in message
news:<I06oqA.GsB_at_news.boeing.com>...
> > Hi DBAs,
> >
> > I am facing some problems importing an 8i export file into 9i instance.
> >
> > I have created tablespaces with same sizes in 9i before the import (used
the
> > "create tablespace.." command from the export dmp file). I find that the
> > tablespaces are getting full prematurely into the import. I would expect
the
> > 9i tablespace to hold the same amount of tables/records as the 8i
tablespace
> > of the same size. But after I set an initial extent to about twice the
size
> > of the 8i tablespace and retry, the import works fine.
> >
> > My question is - Is there a difference in the way 9i datafiles are
managed
> > compared to 8i datafiles, or have I missed setting some parameter on the
9i
> > instance?
>
> Sybrand's answer is the most likely, but it is also possible you
> simply have wrong pct[free|used] parameters in the tables or
> tablespaces. It is possible that a lot of update activity in the 8i
> has enlarged rows, filling up data blocks, then when you do the import
> there needs to be room for that expanded data plus free space for
> future updates (which have already been done in the past). Using the
> large initial extent might hide this because you aren't running into
> Oracle trying to extend segments in too small of a tablespace.
>
> btw, use LMT's if you are not. (They default in 9, but we don't know
> what you've done.)
>
> jg
> --
> @home.com is bogus.
> http://www.nasa.gov/externalflash/cassini/indexmain.html
Received on Thu Jul 08 2004 - 17:55:00 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US