I think it was only a few weeks ago that
I posted a list of errors in an article about
statspack that Oracle Magazine published
in January.
You may be interested to see that another reader
sent a letter to the editor about the same article,
which has been published, with a response, in
the March issue.
<<quote>>
I found an error in "Advanced Tuning with Statspack" in your
January/February 2003 issue. Rich Niemiec writes that in
the event of a wait on a segment header to increase the
Freelist groups. Freelist groups only apply to Oracle9i RAC
systems, and have nothing to do buffer busy waits on
non-RAC systems.
<<end quote>>
<< my comment>>
It is, of course, unfortunate that this 'correction' is actually
wrong. Although the manuals have persistently stated, until
very recently, that freelist groups were relevant only to OPS,
they have been effective in single instance Oracle for a long
time - possibly as far back as 7.2.3
<<end of my comment>>
<<Rich Niemiec's reply>>
Freelist groups do, in fact, have some benefits apart from
Oracle9i RAC. MetaLink says that Freelist groups can
have a positive impact in an exclusive environment
(non-RAC) by helping reduce contention on the segment
header. However, I should point out that segment header
block contention can be addressed without multiple Freelist
groups, for example, by increasing the pctfree/pctused gap
or by partitioning the segment.
<<end of reply>>
<<My comments>>
Surely the point of a 'readers comments' section is
to get feedback from the readers, add value to the
article by including readers' experiences or extra
observations, allow expression of readers' attitudes
and feelings (without an aggressive editorial response),
and, if an actual error does get noted, allow a correction
to be published.
So why publish this erroneous correction ? It did not
add value to the article, it didn't even expose the need
for clarification of a point. It merely elicited a repetition
of some material that had already been stated in the
article (which originally said: "... If this is the case, increase
the freelist groups or increase the pctused to pctfree gap"...)
plus the claim that the author had to be right because the
comment was taken from Metalink.
It would, quite arguably, have been valid to reply with
something like: "Your comments do, indeed reflect
a well-known, and often documented, misconception
about freelist groups. However, the behaviour changed in
Oracle 7.X.X, and single-instance Oracle will take
advantage of multiple free list groups. Be careful,
however, that you set FREELISTS and FREELIST
GROUPS to relatively co-prime numbers, or you
will lose some of the benefits etc....."
So what has the dialogue achieved:
a) It hasn't improved the knowledge of the
readers.
b) It has told readers that they had better not
write in if they think that there is an error in
one of the articles unless they want to run
the risk of looking stupid in public.
c) It has given Rich Niemiec the chance to
say "I'm right, you're wrong - yah, boo, sucks".
I'm not impressed..
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
Coming soon one-day tutorials:
Cost Based Optimisation
Trouble-shooting and Tuning
Indexing Strategies
(see http://www.jlcomp.demon.co.uk/tutorial.html )
____UK_______March 19th
____UK_______April 8th
____UK_______April 22nd
____USA_(FL)_May 2nd
Next Seminar dates:
(see http://www.jlcomp.demon.co.uk/seminar.html )
____USA_(CA, TX)_August
The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Jonathan Lewis
INET: [EMAIL PROTECTED]
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: [EMAIL PROTECTED] (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).
