List
I spent last week at an official Oracle Education Oracle9i Performance
Tuning Class, and here is some of the non-technical stuff I learned.
- Oracle is teaching the wait interface more and more. In fact, they are
updating the curriculum next month to emphasize the wait interface even more
(lucky me).
- Just how the wait interface is emphasized may depend quite a bit on the
instructor, despite what the materials say. My observation is that our
opinions are based on what we have experienced and our interpretations of
those experiences. So we will probably still have some instructors that will
still feel that the wait interface is a passing fad and if you really want
to straighten out a database, you need to get in there and improve the BHR
(Buffer Hit Ratio).
- My instructor was John Hibbard. He is excellent, and I would highly
recommend him. He went well beyond the class materials to providing papers
he has researched and presented himself, as well as other sources, including
papers from Cary Milsap and Jonathan Gennick who participate on this list.
When you get through his class, you really feel you have been taken to a
whole new level of Oracle knowledge. He is also heavily involved in
selecting and preparing the official Oracle training materials for the
courses he teaches. Besides Performance Tuning, he teaches several other
Oracle classes. Most of the people in my class happened to be more
experienced with Oracle, and John did a good job of answering advanced
questions with some depth, but not leaving the newbies in the dust.
- A funny observation on buffer hit ratio vs. wait interface. The last
day of class is an opportunity to take a really screwed-up database and
apply a little of what you have learned. The first scenario is titled
"Buffer Cache". So you run the workload assignment and STATSPACK and look at
the BHR and say "wow, that is bad", increase the buffer pool, and rerun the
workload and STATSPACK. The BHR hasn't changed much, so the tendency is to
dumbly bump the buffer pool even more and go again. Then you look down at
the top 5 waits section just below on the first page of the STATSPACK report
and see that the big wait item is "Scattered Read". Then you go "dope slap"
and realize this schema is missing some critical indexes and table scanning
it's little heart out. I just found it ironic that some people have reported
that some of the Oracle instructors emphasize the BHR too much when the
first Workshop Scenario has a great example of why focusing on BHR can't
solve many problems. But again, we have experience vs. interpretation of
experience. A real died-in-the wool BHR fanatic would probably claim that
BHR had solved the problem because the first indication that something was
wrong was spotting the bad BHR, which led to other investigations.
Dennis Williams
DBA
Lifetouch, Inc.
dwilliams_at_lifetouch.com
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: DENNIS WILLIAMS
INET: DWILLIAMS_at_LIFETOUCH.COM
Fat City Network Services -- 858-538-5051 http://www.fatcity.com
San Diego, California -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing).
Received on Sat Oct 05 2002 - 17:13:22 CDT