Message-ID: <437CEA9B.6020401@commandprompt.com>
Date: Thu, 17 Nov 2005 12:39:55 -0800
From: "Joshua D. Drake" <jd@commandprompt.com>
User-Agent: Thunderbird 1.5 (Windows/20051025)
MIME-Version: 1.0
To: Alex Turner <armtuk@gmail.com>
CC: William Yu <wyu@talisys.com>, pgsql-performance@postgresql.org
Subject: Re: Hardware/OS recommendations for large databases (
References: <3E37B936B592014B978C4415F90D662D01995F7B@MI8NYCMAIL06.Mi8.com>	
	<33c6269f0511152115v1a34ab81j30f1f5df5a6c6956@mail.gmail.com>	
	<dlfjgf$41b$1@news.hub.org> <437B5859.9040805@commandprompt.com>
	<33c6269f0511171154t22e0bce0p803a1803f8b1a539@mail.gmail.com>
In-Reply-To: <33c6269f0511171154t22e0bce0p803a1803f8b1a539@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


>> So what happens is that under reasonable load we are actually waiting
>> for the CPU to process the code.
>>
>>     
>
> This is the performance profile for PHP, not for Postgresql.  This is
> the post
And your point? PostgreSQL benefits directly from what I am speaking 
about as well.

>> Performance of PHP, not postgresql.
>>
>>     
Actually both.

> [snip]
>
> Running postgresql on a single drive RAID 1 with PHP on the same
> machine is not a typical installation.
>   
Want to bet? What do you think the majority of people hosting at 
rackshack, rackspace,
superrack etc... are doing? Or how about all those virtual hosts?

> 300ms for PHP in CPU time?  wow dude - that's quite a page.  PHP
> typical can handle up to 30-50 pages per second for a typical OLTP
> application on a single CPU box.  Something is really wrong with that
> system if it takes 300ms per page.
>   
There is wait time associated with that because we are hitting it with 
50-100 connections at a time.

Joshua D. Drake

> Alex.
>
> ---------------------------(end of broadcast)---------------------------
> TIP 3: Have you checked our extensive FAQ?
>
>                http://www.postgresql.org/docs/faq
>