Re: Lots o' I/O

From: Clarence Gardner <clarence(at)silcom(dot)com>
To: Stephan Szabo <sszabo(at)megazone23(dot)bigpanda(dot)com>
Cc: pgsql-general(at)postgresql(dot)org, Sean Leach <sleach(at)netlojix(dot)com>
Subject: Re: Lots o' I/O
Date: 2003-02-15 00:15:02
Message-ID: Pine.LNX.4.44.0302141610131.21703-100000@liberty.sba2.netlojix.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Fri, 14 Feb 2003, Stephan Szabo wrote:

> On Fri, 14 Feb 2003, Clarence Gardner wrote:
>
> >
> > I have a database that was populated about two months ago, and one
> > particular table has begun causing problems. It's got about 20,000
> > records, all fixed length of about 1 kbytes. If we do any operation
> > that involves a sequential scan of that table (e.g., select count(*)),
> > it now takes about 20 seconds, and according to linux vmstat, reads
> > 275000 disk blocks (275 mbytes). The database is vacuumed each night.
>
> What does vacuum full verbose <table> show? And how big is the actual
> data file?
>

The FULL made the difference -- the table now performs like the copy.
Despite a nightly vacuum analyze, we've never done a vacuum full. The
docs (http://www.postgresql.org/docs/view.php?version=7.3&idoc=0&file=sql-vacuum.html)
almost, but not quite, come out against it....

We're going to do a full vacuum weekly now.

Thanks, all.

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Bruce Momjian 2003-02-15 00:17:43 Re: Dropping column silently kills multi-coumn index (was
Previous Message Tom Lane 2003-02-14 23:51:07 Re: pg_stat_user_tables