From: | Pavan Deolasee <pavan(dot)deolasee(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Vacuum, visibility maps and SKIP_PAGES_THRESHOLD |
Date: | 2011-05-27 15:07:09 |
Message-ID: | BANLkTikvVom5qU1-NuEaK23jDQzKHaATBw@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Fri, May 27, 2011 at 7:36 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Pavan Deolasee <pavan(dot)deolasee(at)gmail(dot)com> writes:
>> My statistical skills are limited, but wouldn't that mean that for a
>> fairly well distributed write activity across a large table, if there
>> are even 3-4% update/deletes, we would most likely hit a
>> not-all-visible page for every 32 pages scanned ?
>
> Huh? With a typical table density of several dozen tuples per page, an
> update ratio in that range would mean that just about every page would
> have something for VACUUM to do, if the modified tuples are evenly
> distributed. The case where the skip optimization has some use is where
> there are large "cold" sections that have no changes at all.
>
I was pretty sure that I would have done my maths wrong :-) So that
means, even for far lesser update ratio, we would pretty much scan
every block and vacuum many of them for a typical well distributed
updates. Hmm. That means the idea of a single pass vacuum is
interesting even after visibility maps.
Thanks,
Pavan
--
Pavan Deolasee
EnterpriseDB http://www.enterprisedb.com
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2011-05-27 15:08:42 | Re: dblink crash on PPC |
Previous Message | Robert Haas | 2011-05-27 15:07:05 | Re: [ADMIN] pg_class reltuples/relpages not updated by autovacuum/vacuum |