Hi Scott, Kevin, Alvaro, and everybody else:
Many thanks for your help and advices. After vacuuming
and reindexing all tables, my nightly run zipped right
through at amazing speed.
Regards,
Tena Sakai
tsakai@gallo.ucsf.edu
-----Original Message-----
From: Scott Marlowe [mailto:scott.marlowe@gmail.com]
Sent: Wed 11/14/2007 1:46 PM
To: Kevin Grittner
Cc: Tena Sakai; pgsql-admin@postgresql.org
Subject: Re: [ADMIN] postgres bogged down beyond tolerance
On Nov 14, 2007 3:33 PM, Kevin Grittner <Kevin.Grittner@wicourts.gov> wrote:
> >>> On Wed, Nov 14, 2007 at 3:16 PM, in message
> <dcc563d10711141316u53338170l5d31d7723c8dc30b@mail.gmail.com>, "Scott Marlowe"
> <scott.marlowe@gmail.com> wrote:
>
> > On Nov 14, 2007 2:26 PM, Tena Sakai <tsakai@gallo.ucsf.edu> wrote:
> >> INFO: vacuuming "public.allele"
> >> INFO: "allele": found 2518282 removable, 1257262 nonremovable row versions
> >> in 31511 pages
>
> >> Total free space (including removable row versions) is 161940948 bytes.
>
> > It looks to me like your tables were bloated. After running vacuum
> > full your cron should run faster now.
>
> Isn't it usually a good idea to REINDEX after using VACUUM FULL to
> recover from this level of bloat?
Depends on your indexes. These didn't look bloated to me, but that
wasn't the whole output of vacuum full either. Yes, reindex is a good
idea after a vacuum full though.