Re: Slow sequential scans on one DB but not another; fragmentation? - Mailing list pgsql-general

From Stephen Harris
Subject Re: Slow sequential scans on one DB but not another; fragmentation?
Date
Msg-id 20070330221600.GA7506@pugwash.spuddy.org
Whole thread Raw
In response to Re: Slow sequential scans on one DB but not another; fragmentation?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
On Wed, Mar 28, 2007 at 11:36:27AM -0400, Tom Lane wrote:
> Stephen Harris <lists@spuddy.org> writes:
> > INFO:  "sweep_users": found 835831 removable, 972662 nonremovable row versions in 2890304 pages

> Oy, that's one bloated table ... only one live row in every three or so pages.
>
> Probably a CLUSTER is the most effective way of cleaning it up.  Once

OK, we were doing a code release today and so had a change window open.
We ran a cluster command on this table.  It took 15 minutes to complete,
and the number of pages went down to 27,000 - ie by a factor of 100.
"select count(*)" took 4 seconds instead of 220, giving us a 55 times
speed increase.

We'll keep our eye on this but since it was relatively quick we might
schedule a weekly cluster "just in case".

Thanks!

--

rgds
Stephen

pgsql-general by date:

Previous
From: csaba
Date:
Subject: Re: PANIC: unexpected hash relation size
Next
From: Ron
Date:
Subject: Anyone know where I can get an 8.2.3 binary for ubuntu?