Re: access time performance problem

From: "Louis-Marie Croisez" <louis-marie(dot)croisez(at)etca(dot)alcatel(dot)be>
To: <pgsql-general(at)postgresql(dot)org>
Subject: Re: access time performance problem
Date: 2002-10-11 07:24:02
Message-ID: 002601c270f7$2c5b9e30$3970cb8a@telchoi.alcatel.be
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

I have tried to modify checkpoint_to and checkpoint_segment, commit_delay and commit_sibblings parameters, but without any effect.
Touching the different postgres parameters, I was not able to influence the height of the peaks (see captures1.jpg), nor the delay
between the peaks.
I wonder if that peaks are independent from postgres or not ? I have executed bonnie++, good results as expected.
Should I upgrade my kernel to the last one (2.4.19), and/or install the last postgres version?

----- Original Message -----
From: "Erwan DUROSELLE" <EDuroselle(at)seafrance(dot)fr>
To: <pgsql-general(at)postgresql(dot)org>
Sent: Thursday, October 10, 2002 5:58 PM
Subject: Rép. : Re: [GENERAL] access time performance problem

Just to be sure:

Could it be a checkpoint?
Checkpoint is supposed to flush all data to disks at regular intervals, isn't it?

-------------------------------------------------------------------------------
Erwan DUROSELLE // SEAFRANCE DSI
Responsable Bases de Données // Databases Manager
Tel: +33 (0)1 55 31 59 70 // Fax: +33 (0)1 55 31 85 28
email: eduroselle(at)seafrance(dot)fr
-------------------------------------------------------------------------------

>>> "scott.marlowe" <scott(dot)marlowe(at)ihs(dot)com> 10/10/2002 17:34 >>>
On Thu, 10 Oct 2002, Louis-Marie Croisez wrote:

> Sorry, bad url:
> use this instead:
> http://louis.croisez.free.fr/capture1.jpg

I'm gonna make a guess here that it's WAL activity, or log activity.

So first off, are you saving postgresql's output to a log? if so, try
logging to a different drive than your data drive or turning off logging
for a quick test.

If it's WAL writes that are causing the problem, you can either try to
change the settings for commit_delay and commit_siblings or symlinking the
pg_xlog directory somewhere else. If you aren't sure how to do that,
don't learn how to on a live production server with real data.

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo(at)postgresql(dot)org)

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/users-lounge/docs/faq.html

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Vince Vielhaber 2002-10-11 09:58:54 Re: news.postgresql.org?
Previous Message Bruce Momjian 2002-10-11 05:19:14 Re: [ADMIN] Multiple backends on a single physical database