From owner-freebsd-hackers  Fri Jan 10 05:36:33 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.4/8.8.4) id FAA11480
          for hackers-outgoing; Fri, 10 Jan 1997 05:36:33 -0800 (PST)
Received: from smtp-gw01.ny.us.ibm.net (smtp-gw01.ny.us.ibm.net [165.87.194.252])
          by freefall.freebsd.org (8.8.4/8.8.4) with SMTP id FAA11475
          for <freebsd-hackers@freebsd.org>; Fri, 10 Jan 1997 05:36:31 -0800 (PST)
Received: (from uucp@localhost) by smtp-gw01.ny.us.ibm.net (8.6.9/8.6.9) id NAA140969; Fri, 10 Jan 1997 13:35:09 GMT
Message-Id: <199701101335.NAA140969@smtp-gw01.ny.us.ibm.net>
Received: from slip166-72-229-80.va.us.ibm.net(166.72.229.80) by smtp-gw01.ny.us.ibm.net via smap (V1.3mjr)
	id smaEIsDmb; Fri Jan 10 13:35:02 1997
Reply-To: <SimsS@IBM.Net>
From: "Steve Sims" <SimsS@IBM.Net>
To: "The Hermit Hacker" <scrappy@hub.org>,
        "Joe Greco" <jgreco@solaria.sol.net>
Cc: "Joe McGuckin" <joe@via.net>, <freebsd-hackers@freebsd.org>
Subject: Re: CCD questions (news server)
Date: Fri, 10 Jan 1997 08:34:14 -0500
X-MSMail-Priority: Normal
X-Priority: 3
X-Mailer: Microsoft Internet Mail 4.70.1155
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

The Hermit Hacker <scrappy@hub.org> writes:
> > > I arbitrarily set the interleave to 64. Is there a good method to
derive 
> > > the optimal interleave factor?
> > 
> > You are killing your performance with such a low number.
> >
> 	Now I'm curious, what interleave are you using?  I don't assume its
> possible to *change* the interleave once the device is created, is it?  I'm
> running even lower (32), since...well...I didn't know better :(
> 
> 	Should maybe add a note to the man page that states a recommended
> value for a news server, or something like that? :)
>  

For what it's worth, I've been playing (in the strictest sense of the word,
not actually *knowing anything* about ccd ;-) with CCD and various interleave
values.

Empirically (using `iozone`) I've decided that performance appears to
"balance peak" (i.e.: best tradeoff between optimized read and optimized
write speeds) when the interleave value specified will *not quite* fill the
on-drive cache.

So, with a farm of el-cheapo antique drives w/128K on-drive cache, I used an
interleave of 240. 

Tweaking the interleave value higher bumps up the write performance but
reading suffers.  Moving the interleave downward didn't significantly help
reading or writing performance. (Although some settings caused the array to
lose, big time!)

N.B.: This is in a plain-vanilla 4-drive, one controller, ccd configuration: 
No mirroring, no CCDF_SWAP or CCDF_UNIFORM flags.  I'd have to play some to
benchmark these.

I briefly tested a configuration with dual Seagate 12550N's, one with 512K
and one with 1M (!) on-drive cache and the result was similar; performance
"balanced" best with an interleave of around 1000 (which is just under the
size of the smaller cache).

YMMV, I learned a "lot" by just playing with the values and newfs'ing.

...sjs...