Re: LMT Autoallocate initial, tot # of extents

  • From: Barbara Baker <barb.baker@xxxxxxxxx>
  • To: Jared Still <jkstill@xxxxxxxxx>
  • Date: Mon, 6 Jun 2005 11:41:25 -0600

nope.  <sigh>  not yet
I'll go check metalink now.
thx.
Barb


On 6/6/05, Jared Still <jkstill@xxxxxxxxx> wrote:
> Ah, that's why I included the IIRC on the INITIAL size; too lazy
> to go look it up.  ;)
>=20
> Have you checked on MetaLink for possible bugs in a version=20
> you are using?
>=20
> 9204 is no longer the 'latest and greatest'.
>=20
> Jared
>=20
>=20
> On 6/6/05, Barbara Baker <barb.baker@xxxxxxxxx> wrote:
> > Thanks, Jared!  I missed that ref in the archives.  It looks like
> > exactly what's happening to me.
> >=20
> > However, in my research, I saw references that lead me to believe
> > there might  be some way to change the behavior.  From a Tom Kyte=20
> > column "LMT's take the initial,next, minextents, pctincrease to figure
> > out how much to initially allocate and allocate that much.  . . . .
> > if you start empty and grow -- it'll adjust.  If you say "give me=20
> > big", it'll start big"
> >=20
> > and this one from Jonathan Lewis:  "normal behaviour pattern is 16 x
> > 64K extents before switching to 1mb extents, but if you're importing
> > large objects with a large 'initial' size, or if your DDL requests a=20
> > large initial size, then Oracle can skip the smaller size extents and
> > leap straight in to 1mb, or even 1m extents."
> >
--
//www.freelists.org/webpage/oracle-l

Other related posts:

  1. Home
  2. oracle-l
  3. Archive
  4. 06-2005