Re: coalescing free space
Date: 1996/01/22
Message-ID: <Pine.SUN.3.91.960122082616.26384A-100000_at_seatimes>#1/1
On Sat, 20 Jan 1996, Carlos Augusto Leite Netto wrote:
> Steve Butler <sbut-is_at_seatimes.com> wrote:
> >
> >On Thu, 18 Jan 1996, Chuck Hamilton wrote:
> >
> >> I remember reading somewhere that oracle is supposed to automatically
> >> coalesce free space if the blocks are contiguous. In other words if I
> >> have 3 10m spaces in the same datafile that are immediately next to
> >> each other, it's supposed to create one 30m space out of it. Does this
> >
> >There was a feature mentioned in this group about a month ago. Seems
> >that if the DEFAULT pctincrease on a tablespace is 0, then ORACLE does
> >not coalesce free space. I've set my default to 1 (% that is) and
> >override it back to 0 on each table. This seems to cure the problem
>
> I think that's better to have pctincrease set to 0 and loosing the
> automatic glueing of free extents. You can easyly glue the free extents
> using a small peace of .sql script, but you can't easyly fix the problems
> caused by non-zero pctincrease. If you ever drop/create tables and
> pctincrease is not-zero, you're get a total fragmented free tablespace
> and you'll need to drop most tables to fix it. Much more work.
If you read carefully, you will see that the pctincrease at the TABLE level is zero. It's the default at the tablespace level that causes the problem with non-coalescing of free space. Anyone who drops and then creates tables without specifying the pctincrease at the table level is simply asking for trouble.
+----------------------------------------------------+ | Steve Butler Voice: 206-464-2998 | | The Seattle Times Fax: 206-382-8898 | | PO Box 70 Internet: sbut-is_at_seatimes.com | | Seattle, WA 98111 Packet: KG7JE_at_N6EQZ.WA | +----------------------------------------------------+All standard and non-standard disclaimers apply. All other sources are annonymous. Received on Mon Jan 22 1996 - 00:00:00 CET