Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 25 Jan 1999 08:33:37 +1100
From:      jonathan michaels <jon@caamora.com.au>
To:        freebsd-advocacy@FreeBSD.ORG
Subject:   Re: FreeBSD - A User's Point of View
Message-ID:  <19990125083337.B18834@caamora.com.au>
In-Reply-To: <XFMail.990124115638.asmodai@wxs.nl>; from Jeroen Ruigrok/Asmodai on Sun, Jan 24, 1999 at 11:56:38AM %2B0100
References:  <19990124212533.B17658@caamora.com.au> <XFMail.990124115638.asmodai@wxs.nl>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Jan 24, 1999 at 11:56:38AM +0100, Jeroen Ruigrok/Asmodai wrote:
> On 24-Jan-99 jonathan michaels wrote:
> > On Sun, Jan 24, 1999 at 01:49:14AM -0800, David Greenman wrote:
> 
> >>    Actually, prior to softupdates, FreeBSD's filesystem performance
> >>    wasn't very good compared to ext2fs for the very reason that ext2fs
> >> is "fast and loose" by defering metadata writes. This has the downside
> >> of making ext2fs filesystem integrity unreliable in the face of a
> >> system crash or power failure. FFS does not have this problem, but is
> >> much slower as a result.
> > 
> > is this softupdates teh same as a journaling filesystem, if not is
> > freebsd going to evolve such a creature ?
> 
> Nay, it's _not_ the same as a journaling filesystem. Well, I was actually
> looking into that the other day (looking at VxFS, JFS and the Novell FS)
> and afaik FreeBSD has no such thing.

yup, i can see that. i asked my question 'cause i;m looking to "sell" freebsd 
to a freind about to dump a significant investment in good novell technology 
for a ms win nt shop approach. for him "the novell journalling filesystem" is 
everything and he is sorta convinced nt will give him "the same", brochures 
are marvellious, read enough of them and you will believe anything.
 
> Journaling differs from SoftUpdates (metadata enhancers) in that SU caches
> metadata and repositions the cached writes to allow for more contigious
> writes. A journaling filesystem keeps a logfile in which every transaction
> to the filesystem gets logged/journaled (d'oh =) and in case of a serious
> crash it can back out every transaction if needed.

thier are a few other tricks depending on the skill of teh implementors but as 
i understand it that is the basic argument.

> Some people who deal exclusively with filesystems will surely modify my
> answer to reflect their better understanding of the above, but this is in
> how far I know how they work.

that would be good to see .. also provide some ammunition to use.
 
> > what would teh arguments be one way ot the other, please.
> 
> Dunno for sure as I am not an adept in the intricities of filesystems.

its times like these i wish  had my old library, to many moves and to many 
forced cleanouts .. shame, but thats life.

regards

jonathan
 
-- 
===============================================================================
Jonathan Michaels
PO Box 144, Rosebery, NSW 1445 Australia
===========================================================<jon@caamora.com.au>


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



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990125083337.B18834>