*BSD News Article 92205


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!feed1.news.erols.com!news.maxwell.syr.edu!news.mathworks.com!enews.sgi.com!news.be.com!news1.crl.com!nexp.crl.com!usenet
From: "Jordan K. Hubbard" <jkh@FreeBSD.org>
Newsgroups: comp.unix.bsd.freebsd.misc,comp.unix.bsd.bsdi.misc,comp.sys.sgi.misc
Subject: Re: no such thing as a "general user community"
Date: Fri, 28 Mar 1997 16:18:39 -0800
Organization: Walnut Creek CDROM
Lines: 30
Message-ID: <333C5FDF.15FB7483@FreeBSD.org>
References: <331BB7DD.28EC@net5.net> <5hfh2l$i13@flea.best.net> <5hfl3n$a3t@fido.asd.sgi.com> <5hh5n2$9q8@flea.best.net> <5hhi67$1gl@fido.asd.sgi.com>
NNTP-Posting-Host: time.cdrom.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 3.01 (X11; I; FreeBSD 2.2.1-RELEASE i386)
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:37952 comp.unix.bsd.bsdi.misc:6489 comp.sys.sgi.misc:29489

Ray Chen wrote:

> The commit records get batched and written out asynchronously.  If you
> crash before the record is written out, you lose the file creation but
> your filesystem is still consistent.
> 
> There's no way an FFS will ever be able to do that.  Sure you can run
> in async mode.  But if you crash, even if the async writes are properly
> ordered, unless all of them went out, you've probably got an inconsistent
> filesystem.

Now I think you may be unfairly slamming Kirk McKusick's work on soft
updates - it makes FFS do exactly that, using a complex internal state
machine (I wish Kirk read USENET, but I doubt he has the time).  He's
still chasing a few errant bugs, last I heard, but the concept seems to
work and we (in the *BSD camps) are all just basically waiting around
for him to release something we can BETA.

I've no doubt that XFS also provides some substantial benefits of its
own, but then it wouldn't exactly be the only proprietary high-speed
filesystem solution available, either.  Rather than going on bended
knee, checkbook in hand, to SGI or Transarc for a very expensive or
outright unavailable filesystem port, it's far more my preference to
support Kirk's work and get something in the end which can be freely
applied to a wide range of operating systems.  Just judging from my last
casual census, at least, FFS does appear to be quite successful in that
role and I think we may be comparing apples and oranges here.
-- 
- Jordan Hubbard
  FreeBSD core team / Walnut Creek CDROM.