Message-ID: <44BE547D.1000602@joeconway.com>
Date: Wed, 19 Jul 2006 08:49:17 -0700
From: Joe Conway <mail@joeconway.com>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US;
	rv:1.7.13) Gecko/20060501 Fedora/1.7.13-1.1.fc5
MIME-Version: 1.0
To: Stephen Frost <sfrost@snowman.net>
CC: Guoping Zhang <guoping.zhang@nec.com.au>,
	'Florian Weimer' <fweimer@bfk.de>, pgsql-performance@postgresql.org
Subject: Re: Performance penalty for remote access of postgresql
References: <82irlu2jk9.fsf@mid.bfk.de> <003501c6ab05$a4978f20$74304c93@eddy>
	<20060719140135.GE20016@kenobi.snowman.net>
In-Reply-To: <20060719140135.GE20016@kenobi.snowman.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Stephen Frost wrote:
> * Guoping Zhang (guoping.zhang@nec.com.au) wrote:
> 
>>Obviously, if there is no better solution, the TCP round trip penalty will
>>stop us doing so as we do have performance requirement.
> 
> Actually, can't you stick multiple inserts into a given 'statement'?
> ie: insert into abc (123); insert into abc (234);
> 
> I'm not 100% sure if that solves the round-trip issue, but it might..
> Also, it looks like we might have multi-value insert support in 8.2 (I
> truely hope so anyway), so you could do something like this:
> insert into abc (123),(234);

Yeah, see my post from last night on PATCHES. Something like "insert 
into abc (123); insert into abc (234); ..." actually seems to work 
pretty well as long as you don't drive the machine into swapping. If 
you're doing a very large number of INSERTs, break it up into bite-sized 
chunks and you should be fine.

Joe