Re: fragmentation - a myth?

From: Thomas Pall <tpall_at_bga.com>
Date: 22 Nov 98 21:23:59 GMT
Message-ID: <365880ef.0@feed1.realtime.net>

Oracle says it very concisely in its Tuning Manual:

Prioritized Steps of the Tuning Method
The recommended method for tuning an Oracle database prioritizes steps in order of diminishing returns: steps which have the greatest impact on performance are listed first. For optimal results, therefore, tackle tuning issues in the order listed: from the design and development phases through instance tuning.

Step 1: Tune the Business Rules 
Step 2: Tune the Data Design 
Step 3: Tune the Application Design 
Step 4: Tune the Logical Structure of the Database 
Step 5: Tune the SQL 
Step 6: Tune the Access Paths 
Step 7: Tune Memory Allocation 
Step 8: Tune I/O and Physical Structure 
Step 9: Tune Resource Contention 

Step 10: Tune the Underlying Platform(s) After completing the steps of the tuning method, reassess your database performance and decide whether further tuning is necessary.

Fragmentation is all they way in Step 8.

Jorge Meirim (transparente_at_mail.telepac.pt) wrote:
: (1) You can analyze whatever you want if you : - Choose RULE in init.ora
:
: - If you build rule oriented Queries.
: (2) Unbalanced B* Tree's. Makes sense that you reorg (depending on queries)

: (3) Analyzing objects to see Row Migration is a good ideia. Sometimes
: chaining can't be avoid.

: (4) I don't call myself DBA

: Jorge Meirim

: >Has anyone done timings to see how much
: >performance improves when they
: >(1) re-analyze the table
: >(2) rebuild the index
: >(3) de-fragment the objects?
: >
: >I'm going to do some myself soon.
: >
: >I know theoretically you reduce head
: >movement, etc, but I wonder how much that
: >matters these days.
: >
: >Any results would be very appreciated.
: >
: >Thanks,
: >
: >Ben Krug, DBA
: >benkrug_at_yahoo.com
: >
: >
: >
: >-----------== Posted via Deja News, The Discussion Network ==----------
: >http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own

-- Received on Sun Nov 22 1998 - 15:23:59 CST