[rfc] use ext3 instead of ext2

Ken Moffat ken at kenmoffat.uklinux.net
Fri Jul 16 14:05:39 PDT 2004

On Fri, 16 Jul 2004, Matthias B. wrote:

> On Thu, 15 Jul 2004 22:30:31 +0100 Matthew Burgess
> <matthew at linuxfromscratch.org> wrote:
> > That's my understanding too.  With a journalled FS there is no need to
> > fsck the drive _at all_ as everything is journalled.
> That is a common misconception. Journaling filesystems can still get
> corrupted, e.g. by bugs in the fs code. So there always needs to be a tool
> to do a complete integrity check and possible repairs. But you don't
> usually need to run this after every unclean shutdown. But you should run
> it occasionally.

 Too true.  I've unknowingly trashed an ext3 rootfs before now (DMA
problems), and carried on until fsck ran.  After the fsck, there wasn't
very much left.  Partly, this is the difference between journalling for
quick recovery (keep the meta data intact, ignore the user data) and
journalling the data, although in my case journalling the corrupted data
would not have helped.  Could have been worse, the platform I was
running is known not to be stable.  The things that scare me are old
versions of reiser3 (should be ok now, but conceptually I give up on
tail-packing)  and current xfs (threads this week in l-k on nulls
appearing in files that had been edited a day or two before an unclean

 das eine Mal als Tragödie, das andere Mal als Farce

More information about the lfs-dev mailing list