Re: Tuning Postgres 9.1 on Windows - Mailing list pgsql-performance

From Merlin Moncure
Subject Re: Tuning Postgres 9.1 on Windows
Date
Msg-id CAHyXU0wAvR_vGa6AmZSgcZVo7foBoL4EoNCM1h3BCTBCn+dDvg@mail.gmail.com
Whole thread Raw
In response to Re: Tuning Postgres 9.1 on Windows  ("Walker, James Les" <JAWalker@cantor.com>)
List pgsql-performance
On Tue, May 1, 2012 at 9:44 AM, Walker, James Les <JAWalker@cantor.com> wrote:
> I installed the enterprisedb distribution and immediately saw a 400% performance increase. Turning off fsck made it
anorder of magnitude better. I'm now peaking at over 400 commits per second. Does that sound right? 

yeah -- well it's hard to say but that sounds plausible based on what
i know.  it would be helpful to see the queries you're running to get
apples to apples idea of what's going on.

> If I understand what you're saying, then to sustain this high rate I'm going to need a controller that can defer
fsyncrequests from the host because it has some sort of battery backup that guarantees the full write. 

yes --  historically, they way to get your tps rate up was to get a
battery backed cache.  this can give you burst (although not
necessarily sustained) tps rates well above what the drive can handle.
 lately, a few of the better built ssd also have on board capacitors
which provide a similar function and allow the drives to safely hit
high tps rates as well.  take a good look at the intel 320 and 710
drives.

merlin

pgsql-performance by date:

Previous
From: Clemens Eisserer
Date:
Subject: Re: Any disadvantages of using =ANY(ARRAY()) instead of IN?
Next
From: Thomas Kellerer
Date:
Subject: Re: Tuning Postgres 9.1 on Windows