Re: Does oracle 8i round extents up for new tables?

I think I remember that article. And I think it says that 8i (and all Oracle versions) will still perform the "round up to 5 blocks" trick *unless* you specify MINIMUM EXTENT (for Oracle 8) or use locally managed tablespaces (for 8i).

If all you've got is 'create tablespace blah datafile '/bing/bong/blah.dbf' size 10m', and you're then simply doing 'create table X (...) storage (initial 256K) tablespace blah;' then you'll get rounding whatever version you're using

It's an extremely bad idea not to have MINIMUM EXTENT set if you insist on using dictionary-managed tablespaces, and its an even worse idea to specify a storage clause when creating the segments. Use locally managed tablespace, and the segment storage clause becomes utterly redundant. Even in 8.0, the MINIMUM EXTENT starts (effectively) modifying the segment storage clause.

Regards
HJR

--

Oracle Resources : http://www.geocities.com/howardjr2000
========================================


"Leigh" <lsatchell_at_geelongcity.vic.gov.au> wrote in message
news:2603dc0c.0110310408.77f38ad_at_posting.google.com...


> Created some new tables with initial extent size 256k and when the

> tables created the extent count is one and bytes are 286720 bytes for

> all but a few of them instead of the expected 262144 (256k). I read at

> http://www.ixora.com.au/tips/creation/extents.htm that 7.3 rounds up

> to the nearest 5 blocks however thought 8i would not do as is has not

> for a few of the tables. There is no min_extlen set for the tablespace

> either and min_extents are 1. Can anybody please shed some light on

> this one since I'm tying to get an extent size to match my multiblock

> read count.  DB is 8.1.7.2 with 8k block size.

>

> Many thanks, Leigh.