Re: I/O tuning... Allocating spindles to databases

  • From: Jared Still <jkstill@xxxxxxxxx>
  • To: cmarquez@xxxxxxxxxxxxxxxx
  • Date: Thu, 15 Sep 2005 20:09:10 +0100

On 9/14/05, Marquez, Chris <cmarquez@xxxxxxxxxxxxxxxx> wrote:
>  I have seen to big SAN implementations.
>  The prevailing wisdom (sales pitch) is "the SAN is huge and everything will
> be cached"
>  and "you will have more I/O than you can imagine".

One of the key statements in the 'SANE SAN' paper is that you
should just pretend the cache doesn't exist.

It isn't too difficult to write a simulation that will render the cache useless.

Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist
11+ years of trying to appear to know what I'm doing.

Other related posts: