Date: Sat, 20 Oct 2001 09:32:16 -0500 From: Bob Willcox <bob@immure.com> To: "David W. Chapman Jr." <dwcjr@inethouston.net> Cc: stable list <freebsd-stable@freebsd.org> Subject: Re: dirpref gives massive performance boost Message-ID: <20011020093216.A69602@luke.immure.com> In-Reply-To: <01f701c157fc$61626ce0$fe0c4042@inethouston.net>; from dwcjr@inethouston.net on Thu, Oct 18, 2001 at 12:43:21PM -0500 References: <20011002214353.A653@student.uu.se> <20011018083713.A20403@polands.org> <20011018144850.A1943@shikima.mine.nu> <20011018105736.A43973@leviathan.inethouston.net> <p05101005b7f4bb89498e@[128.113.24.47]> <01c101c157f9$87868da0$fe0c4042@inethouston.net> <20011018133100.A4472@palomine.net> <01d101c157fb$1fab1c80$fe0c4042@inethouston.net> <20011018134140.A4720@palomine.net> <01f701c157fc$61626ce0$fe0c4042@inethouston.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Oct 18, 2001 at 12:43:21PM -0500, David W. Chapman Jr. wrote:
> On Thu, Oct 18, 2001 at 12:34:22PM -0500, David W. Chapman Jr. wrote:
> > And that's exactly what I'm trying to figure out. Did I build my kernel
> > with then capability and then newfs my drive, or did I newfs my drive
> before
> > this capability. I can't remember. So once
> > again.......................Does anyone know how to check to see if a
> slice
> > is utilizing this feature or has it enabled or whatever term you would
> like
> > to use. Please don't respond with meaningless spam if you dont' know how
> to
> > check because that's all I am looking for.
>
> It's not meaningless spam. You are just failing to understand what people
> are
> trying to tell you.
>
> You don't have to newfs a file system after the code was added for that
> filesystem to be dirpref "enabled." If your kernel has the code in it, then
> all
> of your UFS file systems, even if you newfs-ed them years ago, are
> dirpref-enabled.
>
> Perhaps you should actually read the answers you've been given, instead of
> dismissing them as "meaningless spam."
>
> -------
> If I had received an email like that previously I would not have dismissed
> it. Most everyone I have talked to is under the impression that you have to
> newfs a drive with the kernel to enable dirpref, you are the only one I have
> talked to who doesn't share this view.
The new dirpref code (as I understand it) is simply a change in the
algorithm used to physically locate the subdirectories on the disk
(keeps them closer to their parents). Consequently, once you have
built a new kernel with this new algorithm it will try to locate the
subdirectories accordingly.
However, for mature filesystems with little update activity, there
won't be much opportunity for the new dirpref code to do it's thing
(you aren't creating new subdirectories). So you will see little to no
improvment for these filesystems (depending upon update activity). If,
however, you dump, newfs and restore those same mature filesystems, the
new dirpref code will then restructure the filesystems (as they are
restored) according to the new, improved, layout algorthim.
Note: If I off-base here, someone please correct me.
Bob
>
>
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-stable" in the body of the message
--
Bob Willcox Putt's Law:
bob@vieo.com Technology is dominated by two types of people:
Austin, TX Those who understand what they do not manage.
Those who manage what they do not understand.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011020093216.A69602>