Re: Excessive Logical and Physical I/O

> > If I need to be corrected, then I need to be corrected.
>
> Oh well, in that case... do something about that paper showing 50%
> improvements in speed because of an index rebuild!!! ;-)

I was waiting for someone to take me to task for that paper....I wrote that paper many moons ago, so the information is dated. And I don't go back and update my papers every time I learn something new. I had been meaning to add some more information to that paper, but just havn't gotten around to it yet. Your comments prompted me into action today. If you go to my website (http://www.peasland.net), you will see that I did add an addendum to that white paper.

> In my book, a
> transaction needs to be as big and as long as it needs to be, and if that's
> very big and very long because that's the logic of what a transaction is in
> your business, then so be it.

Definitely. I've always subscribed to the theory that a transaction is "one logical unit of work". So the start of the transaction to the COMMIT or ROLLBACK should reflect that unit of work, not 1/2 or 1/8 or 1/999 units of work.

Cheers!
Brian

-- 
===================================================================

Brian Peasland
dba_at_remove_spam.peasland.com

Remove the "remove_spam." from the email address to email me.


"I can give it to you cheap, quick, and good. Now pick two out of
 the three"