*BSD News Article 62890


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.ysu.edu!usenet.ins.cwru.edu!pravda.aa.msen.com!nntp.coast.net!zombie.ncsc.mil!news.mathworks.com!newsfeed.internetmci.com!in1.uu.net!news.artisoft.com!not-for-mail
From: mday@elbereth.org (Matt Day)
Newsgroups: comp.unix.bsd.freebsd.misc,comp.os.linux.development.system
Subject: Re: The better (more suitable)Unix?? FreeBSD or Linux
Date: 7 Mar 1996 12:41:59 -0700
Organization: none
Lines: 35
Message-ID: <4hne67$qq@coyote.Artisoft.COM>
References: <Dnu8FD.CK2@pe1chl.ampr.org> <4hl00v$7it@coyote.Artisoft.COM> <Dnw3qn.G29@pe1chl.ampr.org>
NNTP-Posting-Host: coyote.artisoft.com
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:14943 comp.os.linux.development.system:18748

In article <Dnw3qn.G29@pe1chl.ampr.org> pe1chl@wab-tis.rabobank.nl writes:
>The examples in your summary only show the well-known cases of files
>or blocks being lost when the system crashes.  The problem can only
>occur with files that were being modified at the time of the crash.
>As we know by now, the content of such files is not reliably known
>anyway (because of delayed data writes).

I agree, after a system crash you may not know exactly what data is in
the files you were writing to at the time of the crash.  But, if your
file system sequences its metadata updates, then you will at least know
that either the data is old (the delayed data write or the associated
sequenced metadata changes never made it to disk) or new (the delayed
data write made it to disk, along with the associated sequenced
metadata changes).  If your file system does not sequence its metadata
updates, then a third possibility presents itself: the data is
garbage.  This third possibility is worth eliminating, because it is
good to know that your files will never contain garbage, and it also
closes a security hole since the garbage might be sensitive data from
someone else's deleted file.

If anyone does not believe what I have said here is true, then please
go read the technical reports on the matter.  This one is a good
start:
	http://www.pdos.lcs.mit.edu/~ganger/papers/CSE-TR-254-95/

>None of your examples shows how "you could loose *every* file on an
>active filesystem", and I still think this is FUD.

I agree, I can't think of a scenario where unsequenced metadata updates
could cause the loss of the entire file system.  (Assuming your fsck
doesn't grossly screw up the cases that it can't fix.)  The summary was
directed at the general confusion on the matter.  (Sorry if I didn't
make that clear.)

Matt Day <mday@elbereth.org>