Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 20 May 1998 09:58:28 -0700 (PDT)
From:      Tom <tom@sdf.com>
To:        The Hermit Hacker <scrappy@hub.org>
Cc:        hackers@FreeBSD.ORG
Subject:   Re: [HACKERS] sorting big tables :(
Message-ID:  <Pine.BSF.3.95q.980520095444.12309D-100000@misery.sdf.com>
In-Reply-To: <Pine.BSF.3.96.980520131553.14056W-100000@hub.org>

next in thread | previous in thread | raw e-mail | index | archive | help

On Wed, 20 May 1998, The Hermit Hacker wrote:

> On Wed, 20 May 1998, Tom wrote:
> 
> > 
> > On Wed, 20 May 1998, The Hermit Hacker wrote:
> > 
> > > 	One of the things that the Unix FS does is auto-defragmenting, at
> > > least the UFS one does.  Whenever the system is idle (from my
> > > understanding), the kernel uses that time to clean up the file systems, to
> > > reduce the file system fragmentation.
> > 
> >   No, that doesn't happen.  The only way to eliminate fragmentation is a
> > dump/newfs/restore cycle.  UFS does do fragmentation avoidance (which is
> > reason UFS filesystems have a 10% reserve).
> 
> 	Okay, then we have two different understandings of this.  My
> understanding was that the 10% reserve gave the OS a 'temp area' in which
> to move blocks to/from so that it could defrag on the fly...

  The 10% reserve is use to avoid fragmentation and excessive CPU usage
from fragmenting files all over the place.

  Besides fragmentation is not really much of performance problem, unless
the fragmentation is excessive.  This is not DOS, so the filesystem may be
doing lots more than reading one file at a time.

Tom


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



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95q.980520095444.12309D-100000>