Skip site navigation (1) Skip section navigation (2)

Re: Hardware/OS recommendations for large databases (

From: Matthew Nuzum <mattnuzum(at)gmail(dot)com>
To: David Boreham <david_list(at)boreham(dot)org>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Hardware/OS recommendations for large databases (
Date: 2005-11-16 18:51:25
Message-ID: (view raw, whole thread or download thread mbox)
Lists: pgsql-performance
On 11/16/05, David Boreham <david_list(at)boreham(dot)org> wrote:
>  >Spend a fortune on dual core CPUs and then buy crappy disks...  I bet
>  >for most applications this system will be IO bound, and you will see a
>  >nice lot of drive failures in the first year of operation with
>  >consumer grade drives.
> I guess I've never bought into the vendor story that there are
> two reliability grades. Why would they bother making two
> different kinds of bearing, motor etc ? Seems like it's more
> likely an excuse to justify higher prices. In my experience the
> expensive SCSI drives I own break frequently while the cheapo
> desktop drives just keep chunking along (modulo certain products
> that have a specific known reliability problem).
> I'd expect that a larger number of hotter drives will give a less reliable
> system than a smaller number of cooler ones.

Of all the SCSI and IDE drives I've used, and I've used a lot, there
is a definite difference in quality. The SCSI drives primarily use
higher quality components that are intended to last longer under 24/7
work loads. I've taken several SCSI and IDE drives apart and you can
tell from the guts that the SCSI drives are built with sturdier

I haven't gotten my hands on the Raptor line of ATA drives yet, but
I've heard they share this in common with the SCSI drives - they are
built with components made to be used day and night for years straight
without ending.

That doesn't mean they will last longer than IDE drives, that just
means they've been designed to withstand higher amounts of heat and
sustained activity. I've got some IDE drives that have lasted years++
and I've got some IDE drives that have lasted months. However, my SCSI
drives I've had over the years all lasted longer than the server they
were installed in.

I will say that in the last 10 years, the MTBF of IDE/ATA drives has
improved dramatically, so I regularly use them in servers, however I
have also shifted my ideology so that a server should be replaced
after 3 years, where before I aimed for 5.

It seems to me that the least reliable components in servers these
days are the fans.

Matthew Nuzum

In response to

pgsql-performance by date

Next:From: Steinar H. GundersonDate: 2005-11-16 18:51:38
Subject: Re: Hardware/OS recommendations for large databases (
Previous:From: Scott MarloweDate: 2005-11-16 17:54:34
Subject: Re: Hardware/OS recommendations for large databases (

Privacy Policy | About PostgreSQL
Copyright © 1996-2017 The PostgreSQL Global Development Group