From: Scott Marlowe <scott.marlowe@gmail.com>
To: Rafael Domiciano <rafael.domiciano@gmail.com>
Cc: pgsql-admin@postgresql.org <pgsql-admin@postgresql.org>
Subject: Re: Shared_buffers hint
Date: Mon, 6 Oct 2008 08:06:01 -0600
Message-ID: <dcc563d10810060706p302a2ec6rb6e84c5a1d8ef5c3@mail.gmail.com> (raw)
In-Reply-To: <3a0028490810060656o5d1482a0r1efdb594f3aeb16c@mail.gmail.com>
References: <3a0028490810060452m268d0777uc18e2e605ff0bd28@mail.gmail.com>
<dcc563d10810060642m175af9ddt7240e065b1b6d53@mail.gmail.com>
<3a0028490810060656o5d1482a0r1efdb594f3aeb16c@mail.gmail.com>
On Mon, Oct 6, 2008 at 7:56 AM, Rafael Domiciano
<rafael.domiciano@gmail.com> wrote:
> I'm using 8.3.3 version in this machine..!
> So, the shared_buffers set to 100Mb is ok? Or 500Mb is better?
> This server we call "Reporter Server", so it's used to do heavy reports for
> a few users only (I could say 5 users).
> What could be the best config for my server with:
> 2 Gb RAM
> 300 GB HD
> Postgres 8.3
> Dual Core
100M is probably adequate. 2G is a pretty small database server
memory wise. I assume by 300G HD you mean a single hard drive. Since
a single hard drive is going to limit the speed at which you can
access data from it, I'd leave shared_buffers at 100M and let the OS
cache data for you. Also, look at work_mem. You might want to set
those few heavy users to have more work_mem than the other users.
alter user heavyuser set work_mem=128000;
note that work_mem is per user sort, so it's quite possible to exhaust
main memory if you set it high for everybody and they all do sorts on
large sets suddenly.
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-admin@postgresql.org
Cc: scott.marlowe@gmail.com, rafael.domiciano@gmail.com
Subject: Re: Shared_buffers hint
In-Reply-To: <dcc563d10810060706p302a2ec6rb6e84c5a1d8ef5c3@mail.gmail.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox