Re: Locally managed tablespaces and migrated rows

"Nuno Souto" <nsouto_at_nsw.bigpond.net.au.nospam> wrote in message news:3af6c86c.24242540_at_news-server...
> On 7 May 2001 13:23:24 GMT, xmark.powell_at_eds.com.x wrote:
>
> >object taking too many extents. I think the factual information for how
 many
> >extents is too many and how much wasted space is too much is currently
> >non-existent. The numbers you use should probably be 'confort' level
> >numbers.
>
> Yeah, good point. Fact is: with LMTS and uniform allocation, all
> extents are same size. That doesn't automatically match every
> situation. There are many instances in PS where one knows certain
> tables are gonna be empty or nearly, so they should have a much
> smaller extent.
>
> But you should hear the local ORACLE education guys: one would think
> LMTS is the best thing since sliced bread! They are recommending them
> across the board as some sort of universal panacea... I'm having
> tremendous trouble convincing a newbie DBA that it is not the case.

I trust that's not a dig at certain, er, known individuals, Nuno!! LMTS *is* nearly as good as sliced bread (with Vegemite, natch), and I'd recommend them heavily -but not stupidly. If you've different extent needs, use different tablespaces.

Remembering that one of the reasons we like LMTS is it's total lack of fragmentation, I recall one instructor doing an install of some Oracle app or other -and I'd persuaded him of the virtues of LMTS, so naturally, he set up all his tablespaces with 1m extents... net result: one app that used to take 55Mb or so suddenyl decided to chew up 650Mb. And 90% of those 1Mb extents were fresh air! I'd prefer fragmentation myself!! (Those Apps boys wouldn't know one end of a data dictionary from the other, bless them!)

But that's just daft *implementation* of LMTS, not an indictment of LMTS themselves.

Regards
HJR
>
> Cheers
> Nuno Souto
> nsouto_at_bigpond.net.au.nospam
> http://www.users.bigpond.net.au/the_Den/index.html
Received on Mon May 07 2001 - 20:03:45 CDT