Re: Performance issues when the number of records are around 10 Million

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: venu madhav <venutaurus539(at)gmail(dot)com>
Cc: Jorge Montero <jorge_montero(at)homedecorators(dot)com>, pgsql-performance(at)postgresql(dot)org
Subject: Re: Performance issues when the number of records are around 10 Million
Date: 2010-05-26 04:25:56
Message-ID: AANLkTilPBkZmPk-ILSYjM5y47DNIcrKcWoL38l3yye1B@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Wed, May 12, 2010 at 1:45 AM, venu madhav <venutaurus539(at)gmail(dot)com> wrote:
> [Venu] Yes, autovacuum is running every hour. I could see in the log
> messages. All the configurations for autovacuum are disabled except that it
> should run for every hour. This application runs on an embedded box, so
> can't change the parameters as they effect the other applications running on
> it. Can you please explain what do you mean by default parameters.
> autovacuum = on                         # enable autovacuum
> subprocess?
> autovacuum_naptime = 3600               # time between autovacuum runs, in
> secs

The default value for autovacuum_naptime is a minute. Why would you
want to increase it by a factor of 60? That seems likely to result in
I/O spikes, table bloat, and generally poor performance.

There are dramatic performance improvements in PostgreSQL 8.3 and 8.4.
Upgrading would probably help, a lot.

The points already made about LIMIT <some huge value> are also right on target.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message David Jarvis 2010-05-26 06:13:45 Re: Random Page Cost and Planner
Previous Message Robert Haas 2010-05-26 04:13:42 Re: Function scan/Index scan to nested loop