On Fri, 2008-06-27 at 16:02 +0100, Simon Riggs wrote:
> The patch splits into two parts:
> * the machinery to *not* dirty a page when we set hints
> * behaviour modifications now that we can tell the difference between
> dirty and hinted pages
>
> Nobody has yet come up with any comments about the first half, which is
> good. The second part is clearly where much debate will occur. I'm going
> to literally split the patch into two, so we can get the machinery into
> CVS and then fiddle and argue over the second part over next few months.
The first "half" is actually quite large, but that makes it even more
sensible to commit this part now.
The enclosed patch introduces the machinery by which we might later
optimise hint bit setting. It differentiates between hint bit setting
and block dirtying, when the distinction can safely be made. It acts
safely during VACUUM and correctly during checkpoint. In all other
respects it emulates current behaviour.
The actual tuning patch can be discussed later, probably at length.
Later patches will be fairly small in comparison and so various people
can fairly easily come up with their own favoured modifications for
testing.
--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support