Date: Thu, 24 Nov 2005 22:26:44 +1300
From: Mark Kirkwood <markir@paradise.net.nz>
Subject: Re: Hardware/OS recommendations for large databases (
In-reply-to: <BFAAC63B.14618%llonergan@greenplum.com>
To: Luke Lonergan <llonergan@greenplum.com>
Cc: stange@rentec.com, Greg Stark <gsstark@mit.edu>,
	Dave Cramer <pg@fastcrypt.com>, Joshua Marsh <icub3d@gmail.com>,
	pgsql-performance@postgresql.org
Message-id: <43858754.4020300@paradise.net.nz>
MIME-version: 1.0
Content-type: text/plain; format=flowed; charset=ISO-8859-1
Content-transfer-encoding: 7bit
User-Agent: Mozilla Thunderbird 1.0.6 (X11/20051106)
References: <BFAAC63B.14618%llonergan@greenplum.com>

Luke Lonergan wrote:
> Mark,
> 
> 
>>>Time: 197870.105 ms
>>
>>So 198 seconds is the uncached read time with count (Just for clarity,
>>did you clear the Pg and filesystem caches or unmount / remount the
>>filesystem?)
> 
> 
> Nope - the longer time is due to the "second write" known issue with
> Postgres - it writes the data to the table, but all of the pages are marked
> dirty?  So, always on the first scan after loading they are written again.
> This is clear as you watch vmstat - the pattern on the first seq scan is
> half read / half write.
>

Ah - indeed - first access after a COPY no? I should have thought of 
that, sorry!