Re: Slow index scan on B-Tree index over timestamp field

From: Igor Neyman <ineyman(at)perceptron(dot)com>
To: Caio Casimiro <casimiro(dot)listas(at)gmail(dot)com>
Cc: Jeff Janes <jeff(dot)janes(at)gmail(dot)com>, "pgsql-performance(at)postgresql(dot)org" <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Slow index scan on B-Tree index over timestamp field
Date: 2013-11-05 13:17:59
Message-ID: A76B25F2823E954C9E45E32FA49D70EC4283C108@mail.corp.perceptron.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

From: Caio Casimiro [mailto:casimiro(dot)listas(at)gmail(dot)com]
Sent: Monday, November 04, 2013 4:33 PM
To: Igor Neyman
Cc: Jeff Janes; pgsql-performance(at)postgresql(dot)org
Subject: Re: [PERFORM] Slow index scan on B-Tree index over timestamp field

These are the parameters I have set in postgresql.conf:

work_mem = 128MB
shared_buffers = 1GB
maintenance_work_mem = 1536MB
fsync = off
synchronous_commit = off
effective_cache_size = 2GB

The hardware is a modest one:
CPU: Intel(R) Atom(TM) CPU 230 @ 1.60GHz
RAM: 2GB
HD: 1TV 7200 RPM (WDC WD10EZEX-00RKKA0)

This machine runs a slackware 14.0 dedicated to the Postgresql.

Thank you,
Caio
With just 2GB RAM, this:

shared_buffers = 1GB

and this:

effective_cache_size = 2GB

is too high.

You should lower those:

shared_buffers = 256MB
effective_cache_size = 1GB

and see how your execution plan changes.

Oh, and this:
maintenance_work_mem = 1536MB

is also too high.
Turning off fsync and synchronous_commit is not very good idea.

Regards,
Igor Neyman

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Jeff Janes 2013-11-05 16:10:19 Re: Slow index scan on B-Tree index over timestamp field
Previous Message Jan Walter 2013-11-05 12:25:06 Trees: integer[] outperformed by ltree