Re: Rather large LA

From: Andres Freund <andres(at)anarazel(dot)de>
To: pgsql-performance(at)postgresql(dot)org
Cc: Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>, Richard Shaw <richard(at)aggress(dot)net>, Andy Colson <andy(at)squeakycode(dot)net>
Subject: Re: Rather large LA
Date: 2011-09-06 14:51:17
Message-ID: 201109061651.18111.andres@anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Monday 05 Sep 2011 22:23:32 Scott Marlowe wrote:
> On Mon, Sep 5, 2011 at 11:24 AM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> > On Monday, September 05, 2011 14:57:43 Richard Shaw wrote:
> >> Autovacuum has been disabled and set to run manually via cron during a
> >> quiet period and fsync has recently been turned off to gauge any real
> >> world performance increase, there is battery backup on the raid card
> >> providing some level of resilience.
> >
> > That doesn't help you against a failure due to fsync() off as the BBU can
> > only protect data that actually has been written to disk. Without
> > fsync=on no guarantee about that exists.
>
> Further, if you've got a bbu cache on the RAID card the gains from
> fsync=off wll be low / nonexistent.
Thats not necessarily true. If you have a mixed load of many small writes and
some parallel huge writes (especially in combination with big indexes)
fsync=off still can give you quite big performance increases. Even in the
presenence of synchronous_commit=off.

Andres

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Tomas Vondra 2011-09-06 16:50:19 Re: Sudden drop in DBb performance
Previous Message Richard Shaw 2011-09-06 14:19:34 Re: Rather large LA