Re: Need help with 8.4 Performance Testing

From: "Merlin Moncure" <mmoncure(at)gmail(dot)com>
To: jd(at)commandprompt(dot)com
Cc: "Scott Marlowe" <scott(dot)marlowe(at)gmail(dot)com>, "Scott Carey" <scott(at)richrelevance(dot)com>, "Jean-David Beyer" <jeandavid8(at)verizon(dot)net>, "pgsql-performance(at)postgresql(dot)org" <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Need help with 8.4 Performance Testing
Date: 2008-12-09 20:07:04
Message-ID: b42b73150812091207n46b0f68fg413c4adee9eba44c@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Tue, Dec 9, 2008 at 1:11 PM, Joshua D. Drake <jd(at)commandprompt(dot)com> wrote:
> Hard drives work, their cheap and fast. I can get 25 spindles, 15k in a
> 3U with controller and battery backed cache for <$10k.

While I agree with your general sentiments about early adoption, etc
(the intel ssd products are the first flash drives that appear to have
real promise in the enterprise), the numbers tell a different story.
A *single* X25-E will give similar sustained write IOPS as your tray
for far less price and a much better worst case read latency. All
this without the 25 sets of whizzing ball bearings, painful spin-up
times, fanning, RAID controller firmware, and various other sundry
technologies to make the whole thing work.

The main issue that I see with flash SSD is if the promised wear
lifetimes are believable in high load environments and the mechanism
of failure (slowly degrade into read only) is accurate.

So, at least in relative terms, 15k sas drives are not 'fast'. They
are terribly, awfully, painfully slow. They are also not cheap in
terms of $/IOPS. The end is near.

merlin

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Joshua D. Drake 2008-12-09 20:20:42 Re: Need help with 8.4 Performance Testing
Previous Message Richard Yen 2008-12-09 19:56:45 query plan with index having a btrim is different for strings of different length