From owner-cvs-all  Mon Dec 10 14: 9: 0 2001
Delivered-To: cvs-all@freebsd.org
Received: from alcatraz.iptelecom.net.ua (alcatraz.iptelecom.net.ua [212.9.224.15])
	by hub.freebsd.org (Postfix) with ESMTP
	id 16B7137B405; Mon, 10 Dec 2001 14:08:47 -0800 (PST)
Received: from ipcard.iptcom.net (ipcard.iptcom.net [212.9.224.5])
	by alcatraz.iptelecom.net.ua (8.9.3/8.9.3) with ESMTP id AAA53591;
	Tue, 11 Dec 2001 00:08:45 +0200 (EET)
	(envelope-from sobomax@FreeBSD.org)
Received: from notebook.vega.com (h195.229.dialup.iptcom.net [212.9.229.195])
	by ipcard.iptcom.net (8.9.3/8.9.3) with ESMTP id AAA18580;
	Tue, 11 Dec 2001 00:08:42 +0200 (EET)
	(envelope-from sobomax@FreeBSD.org)
Message-Id: <200112102208.AAA18580@ipcard.iptcom.net>
To: dillon@apollo.backplane.com
Cc: sheldonh@FreeBSD.org, cvs-committers@FreeBSD.org,
	cvs-all@FreeBSD.org
From: Maxim Sobolev <sobomax@FreeBSD.org>
Subject: Re: cvs commit: src/sbin/newfs newfs.8
X-Mailer: Pygmy (v0.5.13)
Date: Tue, 11 Dec 2001 00:08:39 EET
In-Reply-To: <200112101846.fBAIkee47815@apollo.backplane.com>
Sender: owner-cvs-all@FreeBSD.ORG
Precedence: bulk
List-ID: <cvs-all.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20cvs-all>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20cvs-all>
X-Loop: FreeBSD.ORG

On Mon, 10 Dec 2001 10:46:40 -0800 (PST), Matthew Dillon wrote:
> :This is very questionable topic, at least from the space efficience
> :POV. For example, I found that on my development system using 16:2
> :(5-CURRENT, /usr/ports + /usr/src) some 300MB are wasted in fragments
> :compared with the same partition formatted with 8:1. I would suggest
> :to extend this para explaining that 16:2 could result in a huge space
> :wastage when there are large number of small files on the fs (such as
> :/usr/ports or /usr/src) and 8:1 is better suited for such cases.
> :
> :
> :-Maxim
> 
>     Interesting.  I just ran a test using VN for /usr/ports:
> 
> Filesystem  1K-blocks     Used    Avail Capacity  Mounted on	(512/4K)
> /dev/vn0c     3943361   114700  3513193     3%    /mnt
> 
> Filesystem  1K-blocks     Used    Avail Capacity  Mounted on	(1K/8K)
> /dev/vn0c     4065694   154837  3585602     4%    /mnt
> 
> Filesystem  1K-blocks     Used    Avail Capacity  Mounted on	(2K/16K)
> /dev/vn0c     4129310   251300  3547666     7%    /mnt
> 
>     So /usr/ports alone will eat 100MB of additional disk space on a
>     2K/16K filesystem, which is quite significant.
> 
> 
>     /usr/src has these characteristics:
> 
> Filesystem  1K-blocks     Used    Avail Capacity  Mounted on	(1K/8K)
> /dev/vn0c     4065694  1002597  2737842    27%    /mnt
> Filesystem  1K-blocks     Used    Avail Capacity  Mounted on	(2K/16K)
> /dev/vn0c     4129310  1046374  2752592    28%    /mnt
> 
>     /usr/src isn't so bad in my test, only 44 MBytes larger, which makes
>     sense because the average file size is larger then 1K anyway (so
>     2K/16K wins over 1K/8K for files > 1K and < 2K in size).
> 
>     I didn't think /usr/ports would lose so badly.  It is definitely food
>     for thought.  But I also think that /usr/src validates our prior
>     discussion for 'general purpose' filesystems.

Heh, now imagine that you have checkouted from our repo version of
/usr/ports - due to cvs metadata it has twice as much small files as
normal one, so amount of space wasted in 2K/16K case is also doubles,
resulting in some 200M wasted in fragments. This was exactly the case
here.

-Maxim

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