Re: SMON Logic?

Chakravarthy KM Nalamotu wrote:
>
> Isn't Oracle's assumption that if PCTINCREASE is 0, all the extents
> that the users create will be of standard size WRONG?
> I have seen a few cases where one follows this, but most of the
> times this standard is not followed.
>
> Brett Neumeier (neumebm_at_hpd.abbott.com) wrote:
> : The assumption that Oracle makes is that, if you have
> : PCTINCREASE 0 as the default storage for a tablespace, you will
> : be creating all of the extents in the tablespace with a standard
> : size. [....]

        Evidently, yes, the assumption is wrong -- given that you aren't doing things that way. :-)

        However -- at my site, we have separate tablespaces for each distinct extent size we create. It works quite well for us; we have no problems with internal fragmentation within our tablespaces.   Anyone else could quite easily follow the same approach. Saying that "most of the time this standard is not followed" seems to me in some ways similar to saying "most of the time applications are not thoroughly designed before development begins." That is, sure, it's common to abandon standards; in many cases, though, it's a bad idea to do so.

        In any event -- if you do not like the behavior of Oracle with default PCTINCREASE 0, why not set PCTINCREASE for the tablespace to be 1 and simply declare explicit PCTINCREASE 0 for all of the objects created in the tablespace? (I always advise to declare explicit storage parameters for all database objects in any event. I suppose that this, too, might be an uncommon practice.)

-bn Received on Fri May 09 1997 - 00:00:00 CDT