RE: buffer advisor

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Fri, 12 Sep 2014 07:14:58 +0000
Message-ID: <CE70217733273F49A8A162EE074F64D901E0E9F2_at_exmbx05.thus.corp>

I think the biggest problem with the buffer advisor is that it's reporting stats since the instance started up, NOT for the interval. A piece of advice that "on average" looks okay for the instance might be bad for the critical period you're examining. I am also a little suspicious about how smart the advisor can be regarding tablescans and "small tables" - for example, if you increase the buffer then a large table can become a medium table (without changing size) and end up being buffered more aggressively than you'd like, kicking more useful data out of the cache.

Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
_at_jloracle

---

From: oracle-l-bounce_at_freelists.org [oracle-l-bounce_at_freelists.org] on behalf of Niall Litchfield [niall.litchfield_at_gmail.com] Sent: 12 September 2014 08:01
To: andrew.kerber_at_gmail.com
Cc: ORACLE-L; Mark Powell; Mark W. Farnham Subject: Re: buffer advisor

The feature also has a long, long history of not playing nicely with other Oracle features. A search of the MOS bug database will reveal a number of reasons that might preclude it in your environment.

--
http://www.freelists.org/webpage/oracle-l

Received on Fri Sep 12 2014 - 09:14:58 CEST