On 29 Oct 2003, Robert Treat wrote:
> On Wed, 2003-10-29 at 17:24, Josh Berkus wrote:
> > If anybody has evidence that the FSM index management doens't work, then we'll
> > cut the paragraph. However, I'm inclined to trust Tom & Co., and my only
> > simple tests seemed to uphold the Lazy-Vacuum-ability of indexes.
>
> Tom has laid out at least one case where the potential for index growth
> exits, though I don't see it in a quick search of the archives...
>
> Tom, can you weigh in here?
I thought that was more the case where indexes may be up to 33% larger
than they would be if they were created staticly, but no more. Or
something like that. If the possible maximum size of a vacuumed index is
1/3 or so greater than the most compact size, I wouldn't consider that
bloated. not like the old way, where you'd have tons of dead nodes in the
btree index.