DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=message-id:date:from:user-agent:mime-version:to:cc:subject
	:references:in-reply-to:content-type:content-transfer-encoding;
	b=Z4vlT46bDl+AReqDJu+MqTyi/kApmRHLYK2ZEJvxVV3nq0y/tbvJ7YlLN4vYKl20Dn
	fDpRqaLP7VysdAJ9eUG6PvE6DxRaIHxdz+p2TfbezyLO4rr63gwZrFbMwupLuN5KxLmM
	5JSzSFtJop8yVD4a/+gw6wGFUVrS8ZsRAGzXY=
Message-ID: <488460ED.8090001@gmail.com>
Date: Mon, 21 Jul 2008 12:11:57 +0200
From: Stephane Bailliez <sbailliez@gmail.com>
User-Agent: Thunderbird 2.0.0.14 (Windows/20080421)
MIME-Version: 1.0
To: Greg Smith <gsmith@gregsmith.com>
CC: pgsql-performance@postgresql.org
Subject: Re: Performance on Sun Fire X4150 x64 (dd, bonnie++, pgbench)
References: <4881E9EF.5000805@gmail.com>
	<Pine.GSO.4.64.0807201820320.479@westnet.com>
In-Reply-To: <Pine.GSO.4.64.0807201820320.479@westnet.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Greg Smith wrote:
>
> Note that I've had some issues with the desktop Ubuntu giving slower 
> results in tests like this than the same kernel release using the 
> stock kernel parameters.  Haven't had a chance yet to see how the 
> server Ubuntu kernel fits into that or exactly what the desktop one is 
> doing wrong yet. Could be worse--if you were running any 8.04 I expect 
> your pgbench results would be downright awful.

Ah interesting. Isn't it a scheduler problem, I thought CFQ was the 
default for desktop ?
I doublechecked the 7.10 server on this box and it's really the deadline 
one that is used:

cat /sys/block/sdb/queue/scheduler
noop anticipatory [deadline] cfq

Do you have some more pointers on the 8.04 issues you mentioned ? 
(that's deemed to be the next upgrade from ops)

>> postgresql 8.2.9 with data and xlog as mentioned above
> There are so many known performance issues in 8.2 that are improved in 
> 8.3 that I'd suggest you really should be considering it for a new 
> install at this point.

Yes I'd definitely prefer to go 8.3 as well but there are a couple 
reasons for now I have to suck it up:
- 8.2 is the one in the 7.10 repository.
- I need plr as well and 8.3-plr debian package does not exist yet.

(I know in both cases we could recompile and install it from there, but ...)

> In general, you'll want to use a couple of clients per CPU core for 
> pgbench tests to get a true look at the scalability.  Unfortunately, 
> the way the pgbench client runs means that it tends to top out at 20 
> or 30 thousand TPS on read-only tests no matter how many cores you 
> have around. But you may find operations where peak throughput comes 
> at closer to 32 clients here rather than just 8.
ok. Make sense.

> As far as the rest of your results go, Luke's comment that you may 
> need more than one process to truly see the upper limit of your disk 
> performance is right on target.  More useful commentary on that issue 
> I'd recomend is near the end of 
> http://www.commandprompt.com/blogs/joshua_drake/2008/04/is_that_performance_i_smell_ext2_vs_ext3_on_50_spindles_testing_for_postgresql/ 
>
Yeah I was looking at that url as well. Very useful.

Thanks for all the info Greg.

-- stephane