Oracle Total Recall performance vs Materialized Views

From: Guillermo Alan Bort <cicciuxdba_at_gmail.com>
Date: Mon, 29 Jul 2013 12:59:37 -0300
Message-ID: <CAJ2dSGQjP3Da62s9k98y+NP_z5sxy3uAhWeQwYohMZJZs49abQ_at_mail.gmail.com>



Hi,
  We are having a discussion about the following:

  we have a set of tables, from these tables we need to have a snapshot of the table at 00:00:00 exactly (all the tables at the same time). Right now we are using materialized views and we refresh them nightly. Of course this is a weak solution as if any refresh fails, we loose consistency.

  as I don't like square wheels and the architect can't seem to build a round one, I suggested using Oracle Total Recall, which based on the documentation does everything we need (including guaranteeing consistency and stuff like that).

  The architect claims that the performance on the queries with flashback (with FDA enabled) is unacceptable for the application. I am certain he didn't perform any testing, and I will do so as soon as I finish building the environment, but I'm sure there are some people out there who have already implemented total recall in production so my question to you is what is the performance overhead of FDA (I've read about the overhead during commit, and the restriction regarding DDLs; I'm wondering about the performance of queries).

Thank you very much,

Alan.-

--
http://www.freelists.org/webpage/oracle-l
Received on Mon Jul 29 2013 - 17:59:37 CEST

Original text of this message