Re: Extents in SYSTEM tablespace

There is the ol' Steve Adams argument about cluster chaining, but I've never done any benchmarking - one would imagine (given a dictionary cache) that its at the esoteric end of the performance spectrum.

Cheers
Connor

"Thomas Kyte" <tkyte_at_us.oracle.com> wrote in message news:9unma402t3v_at_drn.newsguy.com...
> In article <u0t9oscttkv3d0_at_corp.supernews.com>, "Van says...
> >
> > If you looked at the thrown extents right after a clean install using
> >Oracle's pctincrease settings you'd still see some objects with several
> >thousand extents. You can probably ignore this, or you can edit sql.bsq
> >before you create a database to remove the problems. Starting in 9i
Oracle
> >recommends not editing sql.bsq. I've done it in 8 and 8i databases and
plan
> >to try it in 9i. It's up to you.
>
> why? do you notice any increased performance or benefits?
>
>
> > The same objects that started with a lot of thrown extents in 8 and
8i
> >are the ones that show up again in 9i.
> >
> >Van
> >
> >
> >"Rick Denoire" <100.17706_at_germanynet.de> wrote in message
> >news:t2kq0u00j9jg44omnvkcmu8bu8df1avdjf_at_4ax.com...
> >> Hello!
> >>
> >> I took a look at the number of extents in the tablespaces of an Oracle
> >> DB (8.1.7) and found out that there are several thousand of them.
> >> Actually, only the TEMP tablespace has more extents.
> >>
> >> But this should not be so, since the objects in the SYSTEM tablespace
> >> are almost never deleted or added. But this is a development DB, so it
> >> happens that tables are created and deleted quite often. Could this be
> >> the reason for the high fragmentation in this tablespace?
> >>
> >> By the way, which is the best way to get rid of fragmentation?
> >>
> >> Thanks
> >>
> >> Rick
> >>
> >
> >
>
> --
> Thomas Kyte (tkyte@us.oracle.com) http://asktom.oracle.com/
> Expert one on one Oracle, programming techniques and solutions for Oracle.
> http://www.amazon.com/exec/obidos/ASIN/1861004826/
> Opinions are mine and do not necessarily reflect those of Oracle Corp
>
Received on Thu Dec 06 2001 - 17:01:46 CST