Message-Id: <4.3.2.7.2.20010905112116.033227e0@pop.xs4all.nl>
Date: Wed, 05 Sep 2001 11:25:12 +0200
To: Peter =?iso-8859-1?Q?W=E4chtler?= <pwaechtler@loewe-komp.de>
From: Seth Mos <knuffie@xs4all.nl>
Subject: Re: System lock while accessing files causes file corruption
Cc: XFS mailing list <linux-xfs@oss.sgi.com>
In-Reply-To: <3B95EA01.15350CE@loewe-komp.de>
References: <3B94F726.E978C299@sgi.com>
 <85063BBE668FD411944400D0B744267A888526@AUSMAIL>
 <3B94F726.E978C299@sgi.com>
 <4.3.2.7.2.20010904193203.032499a0@pop.xs4all.nl>
Mime-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"; format=flowed
Content-Transfer-Encoding: 8bit
Sender: owner-linux-xfs@oss.sgi.com
Precedence: bulk

At 11:01 5-9-2001 +0200, Peter Wächtler wrote:
>Seth Mos wrote:
> > If you use a decent layout fopr your data it does not matter.
> > If you have a separate /usr /var /tmp /home like most servers do you could
> > just mount your / fs O_SYNC since it would only have a _very_ slight
> > performance loss since you almost never write to the root fs. :-)
> >
>
>What about the access time? If you mount your / with sync,noatime
>then it only gets changed when users are logging in
>(chown user.group /dev/pty and alike).
>
>But yes, it's still acceptable ;)

Remember, these operations are metadata operations and these _do_ get 
journaled. It's just to protect the the data files from magically getting 
emptied if you had just touched them.

Cheers

--
Seth
Every program has two purposes one for which
it was written and another for which it wasn't
I use the last kind.