AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20090202113454.54f3d4df@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:exch1.onstor.net
NSV:
SSH:
R:<patrick.haverty@onstor.com>,<dl-LeopardCoreTeam@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/andys@onstor.net@exch1.onstor.net/INBOX	0	2779531E7C760D4491C96305019FEEB51851E6532D@exch1.onstor.net
X-Sylpheed-End-Special-Headers: 1
Date: Mon, 2 Feb 2009 11:35:01 -0800
From: Andrew Sharp <andy.sharp@onstor.com>
To: Patrick Haverty <patrick.haverty@onstor.com>
Cc: dl-Leopard Core Team <dl-LeopardCoreTeam@onstor.com>
Subject: Re: ZFS Default Record Size Compare
Message-ID: <20090202113501.6aee09f4@ripper.onstor.net>
In-Reply-To: <2779531E7C760D4491C96305019FEEB51851E6532D@exch1.onstor.net>
References: <20090202104619.17580b2b@ripper.onstor.net>
	<2779531E7C760D4491C96305019FEEB51851E6532D@exch1.onstor.net>
Organization: Onstor
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

I guess what I'm saying is that when multiple clients are specified,
it's implied that they are running concurrently, so specifying
concurrently or individually is redundant, hence my confusion.

On Mon, 2 Feb 2009 11:01:11 -0800 Patrick Haverty
<patrick.haverty@onstor.com> wrote:

> I wanted to see what each client could do individually, then see what
> the affect of running them all concurrently did to each one's
> throughput (you know, see if they share politely).  So, I'm not sure
> what you are trying to point out. By sequential I meant the disk
> access was sequential, not that the test were run one after the
> other.  Is that where I got a little fuzzy?
> 
> -----Original Message-----
> From: Andy Sharp 
> Sent: Monday, February 02, 2009 10:46 AM
> To: Patrick Haverty
> Cc: dl-Leopard Core Team
> Subject: Re: ZFS Default Record Size Compare
> 
> Hmm, well, except for Read-concurrent test, setting 128K record size
> is a no-brainer.  I'm slightly curious what it means to run the
> clients "individually".  If the clients run their tests one at a
> time, it would not seem to make a difference how many there are.
> 
> On Sun, 1 Feb 2009 14:32:58 -0800 Patrick Haverty
> <patrick.haverty@onstor.com> wrote:
> 
> > I created two volumes (raidz 5+1), each with one folder, on the 
> > Leopard system.  One of the folders was created with the default 
> > record size set to 16K and the other was left at the default record 
> > size of 128K.  Each folder was mounted by two clients and the tests 
> > were ran with all clients running concurrent and also each running 
> > individually.  I ran the tests three times and the averages of
> > those runs are charted here.  The test are all sequential using 16k
> > blocks running dd.
> > 
> > I plan to chart the sio random access test results as time allows,
> > but for now....Enjoy the Super Bowl.
