Date: Tue, 17 Jul 2007 20:04:28 +0200
To: "Marc Mamin" <M.Mamin@intershop.de>, pgsql-performance@postgresql.org
Subject: Re: Postgres configuration for 64 CPUs, 128 GB RAM...
From: PFC <lists@peufeu.com>
Content-Type: text/plain; format=flowed; delsp=yes; charset=utf-8
MIME-Version: 1.0
References: 
 <CA896D7906BF224F8A6D74A1B7E54AB301750B12@JENMAIL01.ad.intershop.net>
Content-Transfer-Encoding: 7bit
Message-ID: <op.tvme5qb2cigqcu@apollo13>
In-Reply-To: 
 <CA896D7906BF224F8A6D74A1B7E54AB301750B12@JENMAIL01.ad.intershop.net>
User-Agent: Opera Mail/9.21 (Linux)

On Tue, 17 > We have the oppotunity to benchmark our application on a  
large server. I
> have to prepare the Postgres configuration and I'd appreciate some
> comments on it as I am not experienced with servers of such a scale.
> Moreover the configuration should be fail-proof as I won't be able to
> attend the tests.
>
> Our application (java + perl) and Postgres will run on the same server,
> whereas the application activity is low when Postgres has large
> transactions to process.

	Please, can you be more specific about your application :

	- what does it do ?
	- what kind of workload does it generate ?
	[ie: many concurrent small queries (website) ; few huge queries,  
reporting, warehousing, all of the above, something else ?]
	- percentage and size of update queries ?
	- how many concurrent threads / connections / clients do you serve on a  
busy day ?
	(I don't mean online users on a website, but ACTIVE concurrent database  
connections)

	I assume you find your current server is too slow or foresee it will  
become too slow soon and want to upgrade, so :

	- what makes the current server's performance inadequate ? is it IO, CPU,  
RAM, a mix ? which proportions in the mix ?

	This is very important. If you go to the dealer and ask "I need a better  
vehicle", he'll sell you a Porsche. But if you say "I need a better vehcle  
to carry two tons of cinderblocks" he'll sell you something else I guess.  
Same with database servers. You could need some humongous CPU power, but  
you might as well not. Depends.

> There is a large gap between our current produtcion server (Linux, 4GB
> RAM, 4 cpus) and the benchmark server; one of the target of this
> benchmark is to verify the scalability of our application.

	Define scalability. (no this isn't a joke, I mean, you know your  
application, how would you like it to "scale" ? How do you think it will  
scale ? Why ? What did you do so it would scale well ? etc.)