From: | Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Free-space-map management thoughts |
Date: | 2003-02-27 17:04:07 |
Message-ID: | 1046365447.4814.607.camel@camel |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thu, 2003-02-27 at 11:00, Tom Lane wrote:
> Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> writes:
> > Now that indexes are getting some reporting, my understanding is an
> > index would report fewer pages overall than it's associated table, but
> > those pages would be completely empty. However, given that they don't
> > reported non-empty pages, the percentage of freeable space to total
> > space would be unfairly lower (if I'm right in thinking that the back
> > end will assume that non-reported pages don't have empty space in them).
> > This would tend to hurt index management even though it's pages are the
> > best candidates for removal (100% empty). Is this a valid concern, or am
> > I misreading something?
>
> I'm not following your point... across relations, the proposed scheme
> only considers numbers of pages, not how much space is believed free in
> each such page. If anything I suspect it would over-favor the indexes.
>
I think I was thinking that a given table will always report more pages
than an index on that table, since tables can report 50% empty pages
while indexes only report 100% empty pages. This would cause tables to
generally be favored over indexes, even though the index pages have the
most to gain.
Robert Treat
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2003-02-27 17:24:50 | Re: Free-space-map management thoughts |
Previous Message | Merlin Moncure | 2003-02-27 16:43:09 | Re: Can pessimistic locking be emulated? |