pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Simon Riggs <simon@2ndquadrant.com>
To: Christopher Petrilli <petrilli@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>
Cc: jd@commandprompt.com, "Marc G. Fournier" <scrappy@postgresql.org>
Cc: pgsql-performance@postgresql.org
Subject: Re: Sustained inserts per sec ... ?
Date: Wed, 06 Apr 2005 09:09:58 +0100
Message-ID: <1112774998.16721.1000.camel@localhost.localdomain> (raw)
In-Reply-To: <59d991c405040515555c77d1e@mail.gmail.com>
References: <20050401140421.P18194@ganymede.hub.org>
	<1112388818.3487.34.camel@jd2.commandprompt.com>
	<59d991c4050401125956d6069d@mail.gmail.com>
	<59d991c40504040648457808e9@mail.gmail.com>
	<26454.1112629979@sss.pgh.pa.us>
	<59d991c405040409094e7bbd17@mail.gmail.com>
	<26669.1112631807@sss.pgh.pa.us>
	<59d991c40504041846bc39a6@mail.gmail.com>
	<5217.1112668589@sss.pgh.pa.us>
	<1112730509.16721.918.camel@localhost.localdomain>
	<59d991c405040515555c77d1e@mail.gmail.com>

On Tue, 2005-04-05 at 18:55 -0400, Christopher Petrilli wrote:
> On Apr 5, 2005 3:48 PM, Simon Riggs <simon@2ndquadrant.com> wrote:
> > B-trees aren't unique to PostgreSQL; the explanation developed here
> > would work equally well for any database system that used tree-based
> > indexes. Do we still think that MySQL can do this when PostgreSQL
> > cannot? How?
> > 
> > Do we have performance test results showing the same application load
> > without the degradation? We don't need to look at the source code to
> > measure MySQL performance...
> 
> http://www.amber.org/~petrilli/diagrams/comparison_mysql_pgsql.png
> 
> That chart shows MySQL (using INSERT against MyISAM tables) and
> PostgreSQL (using COPY) running with the exact same code otherwise.
> Note that MySQL does hit a bit of a wall, but nothing as drastic as
> PostgreSQL and actually maintains something "more flat".  The red and
> blue dashed lines are the 95th percentile point.

Interesting comparison. Any chance of separating the graphs as well, I'm
interested in the detail on both graphs.

Could you estimate the apparent periodicity on the PostgreSQL graphs?

> My suspicion is that what we're seeing is WAL issues, not particularly
> index issues.  The indices just fill up the WAL faster because there's
> more data.  This is a wag basically, but it would seem to explain the
> difference. In both cases, the indices were identical. Five on each.

Let's test the shared_buffers theory.

Would you mind loading only 5M rows per table, but load the same amount
of data overall? That should keep us within the comparable zone overall.

Best Regards, Simon Riggs




view thread (44+ messages)  latest in thread

Message-ID: <1112774998.16721.1000.camel@localhost.localdomain>
Permalink:  ../1112774998.16721.1000.camel@localhost.localdomain/
Also on:    postgresql.org/message-id/1112774998.16721.1000.camel@localhost.localdomain

 · 

reply

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-performance@postgresql.org
  Cc: simon@2ndquadrant.com, petrilli@gmail.com, tgl@sss.pgh.pa.us, scrappy@postgresql.org
  Subject: Re: Sustained inserts per sec ... ?
  In-Reply-To: <1112774998.16721.1000.camel@localhost.localdomain>

* 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