| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: New oracle myths?
Wow, there's so much there to talk about... :-)
As a member of this "small movement," I'll say one thing for starters. =
From
the article, "However, this shift toward CPU-based tuning DOES NOT mean =
that
the ratio-based tuning of the early 1990s has become totally useless."
Two points of clarification:
(1)=A0I'm pretty sure that ratio-based tuning of the early 1990s was =
never
"totally useless." It works sometimes, I'll grant you that. I'm on =
record as
saying it's "worse than useless." Here's why: it's because ratio-based
tuning works sometimes (some people celebrate that), and it fails =
abysmally
other times (I see these failures as a mark of an unacceptable method). =
When
a method inspires confidence but has this propensity for failure, then I
think it's worse than if the method never worked at all. It's the false
confidence that causes so much wasteful IT spending trying to make a =
method
work in circumstances where it can't.
(2)=A0I'm pretty sure that the ratio-based tuning of the early 1990s =
hasn't
"become" any more or less useless than it ever was. I think that the
usefulness of the method has not changed one single bit since it was
invented. I think that what HAS changed is the quality of the other
information that we CAN get out of the Oracle kernel. I do think that =
not
recognizing the value of this other information is a particularly
debilitating decision.
Okay, I'll say two things: You guys realize that 10046 "dumps" aren't in hex, right? :-)
Okay, okay, three things: I hope that if you've read Chapter 1 of
"Optimizing Oracle Performance" (the free chapter that's online at
oreilly.com), then you realize that what many of us are trying to =
accomplish
is NOT a focus on 10046 data. It's a focus on this:
(1) That, to optimize the performance of a BUSINESS, you have to focus =
on a
properly prioritized list of individually targeted tasks.
(2) That, to optimize an individually targeted task, you CANNOT rely on
aggregated data. Note that, fundamentally, (1) and (2) have nothing
whatsoever to do with Oracle.
(3) That, presently, the only performance diagnostic data that Oracle =
offers
(10g or not!) is aggregated data, EXCEPT FOR a very short (10-event) =
history
in the ASH stuff, and a COMPLETE history in the extended SQL trace data
(now, I guess, called the DBMS_MONITOR data). THIS is why I advocate =
using
the extended SQL trace data.
If you've read this article, or if you plan to read this article, =
please,
Please, PLEASE also read Chapter 1 from Jeff's and my Optimizing Oracle
Performance book. It's available free online at oreilly.com.
Have fun!
Cary Millsap
Hotsos Enterprises, Ltd.
http://www.hotsos.com
* Nullius in verba *
Upcoming events:
- Performance Diagnosis 101: 9/14 San Francisco, 10/5 Charlotte, 10/26
Toronto
- SQL Optimization 101: 8/16 Minneapolis, 9/20 Hartford, 10/18 New =
Orleans
- Hotsos Symposium 2005: March 6-10 Dallas
- Visit www.hotsos.com for schedule details...
-----Original Message-----
From: oracle-l-bounce_at_freelists.org =
[mailto:oracle-l-bounce_at_freelists.org]
On Behalf Of Darrell Landrum
Sent: Tuesday, August 31, 2004 3:54 PM
To: oracle-l_at_freelists.org
Subject: New oracle myths?
Take a look at:
http://www.dbazine.com/burleson25.shtml
This is a must see, because I believe (me, my own belief, not any =
company,
list, or other entity) there are folks on this list, umm, mentioned is a
word that comes to mind.
;)
-- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------Received on Tue Aug 31 2004 - 16:38:08 CDT
 |
 |