Re: postgresql random io test with 2 SSD Kingston V+100 500GB in (software) Raid1

From: Laurent Laborde <kerdezixe(at)gmail(dot)com>
To: pgsql-performance(at)postgresql(dot)org
Subject: Re: postgresql random io test with 2 SSD Kingston V+100 500GB in (software) Raid1
Date: 2011-04-19 12:36:42
Message-ID: BANLkTikfsbiMNP2hOKZ2OEHjdixFzk1=EA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Tue, Apr 19, 2011 at 2:07 PM, Greg Smith <greg(at)2ndquadrant(dot)com> wrote:
> On 04/19/2011 05:15 AM, Laurent Laborde wrote:
>>
>> 2 kingston V+100 500GB

4x250GB in Raid10 (see my 2nd post)

> Thanks for the performance report.  The V+100 is based on a Toshiba T6UG1XBG
> controller, and it doesn't have any durable cache from either a battery or
> capacitor.  As such, putting a database on that drive is very risky.  You
> can expect the database to be corrupted during an unusual power outage
> event.  See http://wiki.postgresql.org/wiki/Reliable_Writes for more
> information.
>
> At this point most people considering one of Kingston's drives for a
> database would be better off getting an Intel 320 series drive, which is
> around the same price but doesn't have this issue.

If we use them (unlikely), recovery in case of power outage isn't a
problem, as we will use it on slave database (using Slony-I) that can
be created/destroyed at will.
And, anyway, our slave have fsync=off so the battery won't change
anything in case of power outage :)

i am currently testing on a single V+100 250GB (without raid).
Report will follow soon :)

--
Laurent "ker2x" Laborde
Sysadmin & DBA at http://www.over-blog.com/

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Laurent Laborde 2011-04-19 12:49:32 Re: postgresql random io test with 2 SSD Kingston V+100 500GB in (software) Raid1
Previous Message Greg Smith 2011-04-19 12:07:41 Re: postgresql random io test with 2 SSD Kingston V+100 500GB in (software) Raid1