Re: db cache size in oracle 9ir2

"Mark D Powell" <Mark.Powell_at_eds.com> wrote in message

> I would consider the ability of the spfile to remember parameter
> changes applied dynamically via alter system commands a good thing.
>

Hi Mark:

See my post on this topic elsewhere in this thread.

There is nothing wrong with having changes you make remembered automatically. Unless you do lots of changes which you *wish* to only be temporary... In which case, the boot's on the other foot, and what you see as an advantage is my bugbear.

It is usual when they introduce a significant bit of new functionality (which auto-remembering is) to give you a parameter to control whether or not, and how, that new functionality should be implemented (think cursor_sharing, for example). They didn't do that for the spfile, however. Short of not having an spfile at all, you are stuck with alter system actually taking on a completely different job.

As another example... Alter system used to do exactly the same thing as alter session, with the only difference being that the scope of the dynamic alteration of something was different (all sessions versus one session). Now the implications are quite different... alter system can cause the spfile to change, whereas alter session can't. Why? If they can introduce dot notation to allow one spfile to govern multiple instances, why wasn't that dot notation extended to allow alter session commands to become permanent across instance shutdowns or session logouts, too?

I realise we're stuck with what we've got. And it isn't all bad. But it isn't all good, either. And I remain not one of the spfile's greatest fans.

Regards
HJR Received on Sat Nov 22 2003 - 23:22:18 CST