From: Josh Berkus <josh@agliodbs.com>
Organization: Aglio Database Solutions
To: pgsql-performance@postgresql.org
Subject: Re: [SPAM?] Re: PG8 Tuning
Date: Tue, 16 Aug 2005 09:12:31 -0700
User-Agent: KMail/1.8
Cc: "Jeffrey W. Baker" <jwbaker@acm.org>, Steve Poe <spoe@sfnet.cc>,
	paul@oxton.com
References: <1536.217.45.209.171.1123763001.squirrel@www.gradwell.com>
	<1123836428.19976.26.camel@amd64-laptop-spoe>
	<1123779524.6664.1.camel@toonses.gghcwest.com>
In-Reply-To: <1123779524.6664.1.camel@toonses.gghcwest.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Message-Id: <200508160912.32263.josh@agliodbs.com>

Jeff,

> > 4) pg_xlog: If you're pg_xlog is on a spindle is *only* for pg_xlog
> > you're better off.
>
> Like Mr. Stone said earlier, this is pure dogma. =C2=A0In my experience,
> xlogs on the same volume with data is much faster if both are on
> battery-backed write-back RAID controller memory. =C2=A0Moving from this
> situation to xlogs on a single normal disk is going to be much slower in
> most cases.

The advice on separate drives for xlog (as is all advice on that web page) =
is=20
based on numerous, repeatable tests at OSDL.   =20

However, you are absolutely correct in that it's *relative* advice, not=20
absolute advice.   If, for example, you're using a $100,000 EMC SAN as your=
=20
storage you'll probably be better off giving it everything and letting its=
=20
controller and cache handle disk allocation etc.   On the other hand, if=20
you're dealing with the 5 drives in a single Dell 6650, I've yet to encount=
er=20
a case where a separate xlog disk did not benefit an OLTP application.

=46or Solaris, the advantage of using a separate disk or partition is that =
the=20
mount options you want for the xlog (including forcedirectio) are=20
considerably different from what you'd use with the main database.

=2D-=20
Josh Berkus
Aglio Database Solutions
San Francisco