Date: Mon, 12 Aug 2013 14:18:17 -0400
From: "David F. Skoll" <dfs@roaringpenguin.com>
To: pgsql-admin@postgresql.org
Cc: Scott Whitney <swhitney@journyx.com>
Subject: Re: Opinions on SSDs
Message-ID: <20130812141817.6377101d@hydrogen.roaringpenguin.com>
In-Reply-To: 
 <14817358.266767.1376323269109.JavaMail.root@mail.int.journyx.com>
References: 
 <CAFwQ8rd7cg1QpuQnSdtcL0mJYuC9eDJ26wUzjWa4_NFJxD+R=A@mail.gmail.com>
 <14817358.266767.1376323269109.JavaMail.root@mail.int.journyx.com>
Organization: Roaring Penguin Software Inc.
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

On Mon, 12 Aug 2013 11:01:09 -0500 (CDT)
Scott Whitney <scott@journyx.com> wrote:

> When you say "16 10K drives," do you mean:

I mean 8 RAID-1 pairs with data striped across the pairs.  The Linux
software RAID "offset" scheme is described here:

http://www.ilsistemista.net/index.php/linux-a-unix/35-linux-software-raid-10-layouts-performance-near-far-and-offset-benchmark-analysis.html?start=1

> The SSD solution I put in has shown significant speed improvements,
> to say the very least.

OK; thanks.

> Basically, just assume that you're getting 130 iops per drive. Well,
> 16 drives in a RAID 0 is going to max you out at 2100ish iops,

With our RAID-10 array, we're looking at about 1050 iops.  Our server
currently is holding up OK with a decidedly non-optimal arrangment
(four RAID-1 volumes with pg_xlog on one, most DB files on a second,
and a couple of tablespaces with hand-placed tables and indexes on the
other two, and only 7200 RPM disks.)

So I think 16 10Krpm spinning disks will probably suffice; the failure mode
of SSDs makes me nervous.

Regards,

David.


-- 
Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin