Re: Automatic Segment Space Management

From: joel garry <joel-garry_at_home.com>
Date: Fri, 5 Mar 2010 10:45:22 -0800 (PST)
Message-ID: <61b0d502-a415-4d11-8818-e6dd073d9e62_at_t9g2000prh.googlegroups.com>



On Mar 5, 7:40 am, Mladen Gogala <n..._at_email.here.invalid> wrote:
> On Fri, 05 Mar 2010 04:19:16 -0800, vsevolod afanassiev wrote:
> > On PCTFREE parameter: I think it is better to avoid chained rows and not
> > only for performance reasons:
>
> Speaking of the chained rows, how many people here have actually used the
> ANALYZE utility to list chained rows and base the analysis on facts? At
> one time, there was a query, using th VSIZE function to get you a rough
> estimate for the number of chained rows. That was replaced by ANALYZE
> table list chained rows into <chained rows table, created by utlchain>,
> the utility to produce the list of chained rows. I use it only when the
> developers complain that "reads from the table are slow".
>
> --http://mgogala.byethost5.com

Since it is easy to do this in dbconsole, I do it about every month or two. There are a few tables that get mass updated irregularly, and it lets me know if something has gone unpredictable. There are a couple tables that have a data pattern that makes shrink worthwhile (though less than the tool predicts). It's not obsessive to fix growing problems before people complain. It's just my preciousssssss.

jg

--
_at_home.com is bogus.
http://www.wired.com/threatlevel/2010/03/schmidt-cyberwar/
Received on Fri Mar 05 2010 - 12:45:22 CST

Original text of this message