Re: Proving or Debunking the need for rebuilding

hpuxrac wrote:
> hasta_l3_at_hotmail.com wrote:
> > > >
> > > > Is it possible what the OP is seeing is a delayed block cleanout?
> > >
> > > I think it is entirely possible. I'm wondering if simply doing a
> > > count(*) that uses the index in question, before performing each test,
> > > might shed light one way or the other. The idea is to visit all blocks
> > > and clean them.
> > >
> >
> > We are deleting in a bunch of small transactions.
> > We should have (almost) no delayed block cleanouts, right ?

You previously said a later phase has massive deletes. The thing about delayed block cleanout is... it's delayed. You see the effect in a later access to the blocks, like a select, then don't see it after that. If the complaint is everything slows down always until an index is rebuilt, that would be something else.

Maybe you need to reiterate what you do when, and what slows down, I'm confused.

Now to get way more speculative: the fast cleanout happens when no one else is using the block - is it possible you are hitting the same block in different small transactions concurrrently, and consecutively hitting blocks, such that a bunch of small transactions overlap into a large bunch of blocks?

> >
> > --- Raoul

>

Dang, I'm always 55 miles away from the right book. Fortunately, he also breaches the subject in one-on-one.

jg

--
@home.com is bogus.
"We would like to ask for apologizes to people who supported this"
http://www.argeniss.com/woodb.html
How about an apology from the people who came up with the idea?