To: Andreas Pflug <pgadmin@pse-consulting.de>
Cc: Bob Ippolito <bob@redivi.com>, Mark Cotner <mcotner@yahoo.com>,
	pgsql-performance@postgresql.org
Subject: Re: sustained update load of 1-2k/sec 
In-reply-to: <4305E57E.1010604@pse-consulting.de> 
References: <20050819082404.53751.qmail@web32915.mail.mud.yahoo.com>
	<67788A4A-4011-49AB-B329-683FD9532661@redivi.com>
	<17690.1124458553@sss.pgh.pa.us>
	<4305E57E.1010604@pse-consulting.de>
Comments: In-reply-to Andreas Pflug <pgadmin@pse-consulting.de>
	message dated "Fri, 19 Aug 2005 15:58:22 +0200"
Date: Fri, 19 Aug 2005 10:09:05 -0400
Message-ID: <18000.1124460545@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

Andreas Pflug <pgadmin@pse-consulting.de> writes:
> Tom Lane wrote:
>> As far as the question "can PG do 1-2k xact/sec", the answer is "yes
>> if you throw enough hardware at it".  Spending enough money on the
>> disk subsystem is the key ...
>> 
> The 1-2k xact/sec for MySQL seems suspicious, sounds very much like 
> write-back cached, not write-through, esp. considering that heavy 
> concurrent write access isn't said to be MySQLs strength...

> I wonder if preserving the database after a fatal crash is really 
> necessary, since the data stored sounds quite volatile; in this case, 
> fsync=false might be sufficient.

Yeah, that's something to think about.  If you do need full transaction
safety, then you *must* have a decent battery-backed-write-cache setup,
else your transaction commit rate will be limited by disk rotation
speed --- for instance, a single connection can commit at most 250 xacts
per second if the WAL log is on a 15000RPM drive.  (You can improve this
to the extent that you can spread activity across multiple connections,
but I'm not sure you can expect to reliably have 8 or more connections
ready to commit each time the disk goes 'round.)

			regards, tom lane