Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: truncate command 8.1.7. locally managed tablespaces

Re: truncate command 8.1.7. locally managed tablespaces

From: Howard J. Rogers <howardjr2000_at_yahoo.com.au>
Date: Thu, 22 May 2003 06:09:32 +1000
Message-ID: <MSQya.39492$1s1.557672@newsfeeds.bigpond.com>

"Frank" <fvanbortel_at_netscape.net> wrote in message news:3ECBDB34.9090801_at_netscape.net...
> Howard J. Rogers wrote:
> <snip!>
> >
> > 50% PCTINCREASE has been the default value for Oracle since year dot.
> > It's a disgraceful default, and should always be changed to zero,
> > without exception.
> >
>
> The only exception being SYSTEM in pre-9i (and I'm just stating that
> because I do not know exactly how 9i would behave) instances. 9i with
> LMT SYSTEM would be OK, but DMT?

My turn to say I should have elaborated! Never muck around with SYSTEM's settings. They are weird, break all the 'good' rules, and yet it's been designed that way and works perfectly well. So I agree: SYSTEM is allowed its PCTINCREASE.

>
> If your system TS (on pre-9i) reaches max_extents, you're likely to
> get stuck between a rock and a hard place.

Exactly. Though I'd question why anything in SYSTEM needs to acquire lots of extents, hitting MAXEXTENTS not something you want to encounter.

> Saw that happening to a 7.x db once; it came to a grinding halt, and
> could not be brought back to life, restoring backups was the only
> option. Cannot think of a reason why that would have been improved
> in 8.0 or 8i.
>
> Of course, it depends... on the table being hit, for one.
> But, better safe than sorry.

Absolutely.

Regards
HJR
> --
> Regards, Frank van Bortel
>
Received on Wed May 21 2003 - 15:09:32 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US