Re: Summary: Controlling "ad hoc" queries

From: <pihlab_at_hhcs.gov.au>
Date: 15 Sep 92 09:35:08 +1000
Message-ID: <1992Sep15.093509.370_at_hhcs.gov.au>


In article <lapscgINN8jg_at_news.bbn.com> Nat Brooks, NBROOKS_at_BBN.COM writes:
>Has anyone out there successfully installed and supported an end user
>query tool that was actually used by end users?

Something I didn't include in the original reply is that we try to run end-user adhoc querries on multiple CPU machines (the more CPUs the better).

The reason for this is that an "evil" query will normally only thrash one CPU at a time so other "good" querries will still get a bite of the cherry from the other CPUs.

Of course, you still have disk I/O contention and such but our experience has been that "evil" querries are doing lots of full table scans and Oracle6 doesn't flush these through the SGA buffers as much as Oracle5 so "good" querries are still getting reasonable hits in the buffers.

The more CPUs you have, the more "evil" querries that can be tolerated.

-- 

Bruce...        pihlab_at_hhcs.gov.au
                 ^^
*******************************************************************
* Bruce Pihlamae  --  Database Administration                     *
* Commonwealth Department of Health, Housing & Community Services *
* Canberra, Australia                             (W) 06-289-7056 *
*******************************************************************
* These are my own thoughts and opinions, few that I have.        *
*******************************************************************
Received on Tue Sep 15 1992 - 01:35:08 CEST

Original text of this message