Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 16 Mar 2000 11:20:37 -0800
From:      Greg Lehey <grog@lemis.com>
To:        fbsd-dave <fbsddave@mrcaffeine.com>
Cc:        Mark Smith <msmith@beta.tricity.wsu.edu>, freebsd-questions@FreeBSD.ORG
Subject:   Re: vinum stats and question
Message-ID:  <20000316112036.C333@mojave.worldwide.lemis.com>
In-Reply-To: <Pine.BSF.4.21.0003151426150.5782-100000@boris.netgate.net>; from fbsddave@mrcaffeine.com on Wed, Mar 15, 2000 at 03:29:08PM -0800
References:  <200003151658.IAA31925@beta.tricity.wsu.edu> <Pine.BSF.4.21.0003151426150.5782-100000@boris.netgate.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday, 15 March 2000 at 15:29:08 -0800, fbsd-dave wrote:
> On Wed, 15 Mar 2000, Mark Smith wrote:
>
>> Greetings,
>>
>> Here's some Bonnie stats from my 3.4R machine.  Plase read da1 and da2
>> instead of sd1 and sd2.  I benched the individual drives right after
>> installing 3.4 and was still thinking older device names. <G> Anway,
>> sd1 and sd2 are the drives striped together to get testc.
>>
>> Could somebody please explain why the sequentional block output is sooo
>> much slower on the striped partion than the individual sd1 and sd2
>> drives?
>>
>> #Bonnie -s 200
>>
>> /disk1 /dev/sd1
>>     -------Sequential Output-------- ---Sequential Input-- --Random--
>>     -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
>>  MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
>> 200  7510 93.4  8428 21.5  2856 15.8  7485 93.2 11164 27.4 227.0  6.3
>>
>> /disk2 /dev/sd2
>>     -------Sequential Output-------- ---Sequential Input-- --Random--
>>     -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
>>  MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
>> 200  7546 92.4  8403 21.6  2871 15.9  7512 93.0 11375 29.6 199.9  5.9
>>
>> /testc /dev/vinum/testc
>>     -------Sequential Output-------- ---Sequential Input-- --Random--
>>     -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
>>  MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
>> 200  7881 96.3  2886 25.9  5765 31.7  6297 77.0 16289 46.5 315.8  9.6
>
> Use rawio to test, at least for vinum storage. Bonnie tests through system
> buffers and fibs. (though usually it's optimistic)
>
> -- from rawio(1)'s manpage --
> rawio resembles bonnie in some of the things it does.  It differs strong-
> ly from bonnie by using a raw disk device, which bypasses buffer cache. As
> a result, some of the tests that bonnie performs are meaningless, for
> example character I/O.
> ---

Hmm.  That's not quite correct.  It should read:

  As a result, some of the tests that bonnie performs, for example
  character I/O, have no relation to storage device performance.

> Build it here: /usr/ports/benchmarks/rawio
>
> Still, those numbers do look strange. Can we see the output from "vinum
> printconfig"?

printconfig is a command to create config files.  It's not meant to be
human readable.  The output of 'vinum list' would be much more useful
here.

I'd hazard a guess that this test was done with too small a stripe
size.

Greg
--
When replying to this message, please copy the original recipients.
For more information, see http://www.lemis.com/questions.html
Finger grog@lemis.com for PGP public key
See complete headers for address and phone numbers


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000316112036.C333>