Re: Performance advice

From: pgsql(at)newtopia(dot)com
To: Postgresql Performance <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Performance advice
Date: 2003-06-25 13:04:31
Message-ID: Pine.LNX.4.44.0306250830120.21655-100000@server1.techplant.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Wed, 25 Jun 2003, Achilleus Mantzios wrote:

> What i think would be ideal (helpful/feasible)
> is some kind of documentation of the algorithms involved
> in the planner/optimizer, along with some pointers
> to postgresql.conf parameters where applicable.
>
> This way we will know
> - Why something is happening
> - If it is the best plan
> - What tuning is possible

I agree. In combination with this, I would find case studies very useful.
Have the documentation team solicit a few volunteers with different setups
(w/r/t db size, db traffic, and hardware). Perhaps these folks are
running with the default postgresql.conf or have done little tuning. Via
the performance list, work through the tuning process with each volunteer:

1. Gathering information about your setup that affects tuning.
2. Measuring initial performance as a baseline.
3. Making initial adjustments based on your setup.
4. Identifying poorly-written SQL.
5. Identifying poorly-indexed tables.
6. Measuring effects of each adjustment, and tuning accordingly.

(Note: I am certainly no performance expert -- these steps are meant to be
examples only.)

Solicit a list member to monitor the discussion and document each case
study in a consistent fashion. Run completed case studies by the
performance and docs lists for review.

I would be happy to join the docs team to work on such a project.

michael

p.s. Should this discussion be moved to psgql-docs?

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Rod Taylor 2003-06-25 13:36:06 Re: How to optimize monstrous query, sorts instead of
Previous Message Michael Mattox 2003-06-25 12:48:15 Re: How to optimize monstrous query, sorts instead of