Re: Finding space-bound objects?

Why on Earth are you using PCTINCREASE anyway?

It's the worst invention since the Ford Edsell. Steer clear.

It's there to stop the acquisition of a bazillion extents. So is a good dba. A positive value for it is designed to clear up tablespace fragmentation -which a positive value causes. LMTs cure (by preventing) fragmentation far more effectively.

Therefore, whilst you may be right about the PCTINCREASE screwing up the space calculations in sundry scripts, the probability is that the script writers didn't think anyone was going to be mad enough to use PCTINCREASE in the first place, so didn't worry about it!

Regards
HJR "Tim Kearsley" <tim.kearsley_at_milton-keynes.gov.uk> wrote in message news:725736ef.0204300252.2ce88692_at_posting.google.com...
> Hi all,
>
> I've looked at several scripts which purport to produce a list of
> tables, indexes etc. which can't extend within the tablespace, because
> the largest contiguous piece of free space is less than the
> NEXT_EXTENT value for the segment.
>
> However, unless I'm not seeing something here, this is too simplistic
> for the real world isn't it?
>
> What happens if the segment has a PCT_INCREASE value non-zero? Surely
> in that case the next extent it wants to create will be the size of
> the last extent it created * (1 + PCT_INCREASE/100) won't it? Or have
> I misunderstood the meaning of PCT_INCREASE or the way it is
> implemented?
>
> For example if the table BIGTAB has a PCT_INCREASE value of 50 and has
> INITIAL_EXTENT and NEXT_EXTENT values of 16384 and 16384 respectively,
> won't the extents be allocated with these sizes:
>
> 16384
> 16384
> 24576
> 36864
> 55296
> etc...
>
> I appreciate that these might not be the exact values, as presumably
> the real values would have to be multiples of the block size, but you
> can see the principle.
>
> Any comments or input very welcome!
>
> Regards,
>
> Tim Kearsley
> Database Manager,
> Milton Keynes Council
Received on Tue Apr 30 2002 - 18:47:51 CDT