Re: DB File Sequential Read Waits

Geomancer,

You mention checking.

Do you subscribe to the notion that a checking account is the only purchasing mechanism that you should employ ?

Just to explain the point of the above:
a) Your question was completely outside the context

     of the previous posting, despite the arbitrary link
     that you attempted to build

b) If you had read and thought about the WHOLE post,

    you wouldn't have needed to ask the question.

On a more serious note - I don't think I know of anyone who subscribes to the notion that "one should only tune to reduce logical I/O?"

Personally I hold that tuning is an activity that takes place before production, and the target is a "Nash equilibrium" between minimising the work done by data access, and minimising the work done to support the structures that are needed to minimise the work done by data access - given that different strategies produce different levels of contention at different degrees of concurrency.

After production, the popular pastime of all DBAs is trouble-shooting - and then I subscribe to the notion of addressing processes that are consuming extreme amounts of resource or suffering extreme amounts of contention. But only if they are performing outside their SLAs, or may be causing other processes to perform outside their SLAs.

A sound strategy for trouble-shooting is not something that can be summarised usefully in a sound-bite. Unfortunately, too many people take sensible ideas and parrot them in sound-bite format - usually in a way that turns reasonable advice into crass stupidity.

-- 
Regards

Jonathan Lewis
http://www.jlcomp.demon.co.uk

  The educated person is not the person
  who can answer the questions, but the
  person who can question the answers -- T. Schick Jr


One-day tutorials:
http://www.jlcomp.demon.co.uk/tutorial.html


Three-day seminar:
see http://www.jlcomp.demon.co.uk/seminar.html
____UK___November


The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html


"Geomancer" <pharfromhome_at_hotmail.com> wrote in message
news:cf90fb89.0401040937.52d574e9_at_posting.google.com...


> There are three things I can think of that might help:

>

>     a)    Find queries where the number of disk reads is

>         significantly higher than the number of rows returned,

>         and check to see if the tkprof output shows the table

>         returning far fewer rows than the index, where the

>         disk figure in the tkprof output is roughly the same as

>         the INDEX line rows.  This indicates lots of table blocks

>         being visited unnecessarily, and MIGHT be fixed by adding

>         columns to indexes to avoid the excess table visits.

>

> Jonathan,

>

> You mention checking disk reads.

>

> Do you subscribe to the notion that one should only tune to reduce logical

I/O?