From: Andres Freund <andres@2ndquadrant.com>
Organization: 2ndquadrant.com
To: Julien Cigar <jcigar@ulb.ac.be>
Subject: Re: Postgres becoming slow, only full vacuum fixes it
Date: Mon, 24 Sep 2012 15:34:31 +0200
User-Agent: KMail/1.13.7 (Linux/3.6.0-rc6-andres-00066-g2682cc4-dirty;
	KDE/4.8.4; x86_64; ; )
Cc: pgsql-performance@postgresql.org
References: <D15ECA12-8D6E-4E53-A294-5E175564CA61@gmail.com>
	<201209241434.56760.andres@2ndquadrant.com>
	<506057E7.5090008@ulb.ac.be>
In-Reply-To: <506057E7.5090008@ulb.ac.be>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-15"
Content-Transfer-Encoding: 7bit
Message-Id: <201209241534.31974.andres@2ndquadrant.com>

On Monday, September 24, 2012 02:53:59 PM Julien Cigar wrote:
> On 09/24/2012 14:34, Andres Freund wrote:
> > On Monday, September 24, 2012 02:21:09 PM Julien Cigar wrote:
> >> 5) synchronous_commit = off should only be used if you have a
> >> battery-backed write cache.
> > 
> > Huh? Are you possibly confusing this with full_page_writes?
> 
> indeed...! sorry for that
> (note that you still have a (very) small chance of loosing data with
> synchronous_commit = off if your server crashes between two "commit
> chunks")
Sure, you have a chance of loosing the last some transactions, but you won't 
corrupt anything. Thats the entire point of the setting ;)

Greetings,

Andres
-- 
 Andres Freund	                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services