Re: tuning problem

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Sun, 11 Feb 2001 08:22:38 -0000
Message-ID: <981879549.29910.0.nnrp-07.9e984b29@news.demon.co.uk>

Have you dumped and compared the statistics on the two sets of tables and indexes.

Although your compatibility (from your next post) does change some 'hidden' features, the first check has to be the numbers.

I would guess that the clustering factor on some of the indexes varies significantly. There may also be large enough differences in free-space below hwm and freed-space in indexes.

Can you send us the plan(s) that you get on production, and the plan you would like to see (by hinting it) so we can see the paths AND COSTS that occur on production; and then the plan that appears on DEV.

--
Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases
Publishers:  Addison-Wesley

Reviews at: http://www.jlcomp.demon.co.uk/book_rev.html

Received on Sun Feb 11 2001 - 02:22:38 CST