Re: PostgreSQL Disk Usage and Page Size

From: Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com>
To: "Saleh, Amgad H (Amgad)" <ahsaleh(at)lucent(dot)com>
Cc: Seum-Lim Gan <slgan(at)lucent(dot)com>, "'srn(at)commsecure(dot)com(dot)au'" <srn(at)commsecure(dot)com(dot)au>, pgsql-performance(at)postgresql(dot)org
Subject: Re: PostgreSQL Disk Usage and Page Size
Date: 2004-03-18 16:56:03
Message-ID: 20040318085310.S50452@megazone.bigpanda.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Thu, 18 Mar 2004, Saleh, Amgad H (Amgad) wrote:

> Stephan / Stephen
>
> We know about the overhead and do understand the math you've provided.
> This is not the question we're asking. We've just provided the table definitions as
> examples.
>
> The real question was, even with the 52 & 56 (assuming right),' I wouldn't get
> the same number of records per page for all 4k, 8k, 16k, and 32k pages.

On my system, I don't using your tests, IIRC I got 134 with TEST_1 and
like 128 or so on TEST_2 when I used strings of maximum length for the
columns.

>
> To make it more clear to you here's an example:
>
> For an 8k-page: we've got 120 records/page for both tables and other tables such as
>
> CREATE TABLE TEST_3 (
> F1 VARCHAR(10),
> F2 VARCHAR(12) );

Are you storing the same data in all three tables or different data in all
three tables? That's important because there's no difference in length
between varchar(5) and varchar(12) when storing the same 5 character
string.

Browse pgsql-performance by date

  From Date Subject
Next Message Bruce Momjian 2004-03-18 17:34:36 Re: [HACKERS] fsync method checking
Previous Message Tom Lane 2004-03-18 05:03:47 Re: severe performance issue with planner (fwd)