Subject: Re: LFS partition limitations
To: Tracy J. Di Marco White <gendalia@iastate.edu>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: current-users
Date: 10/03/2000 11:09:36
In message <200010031800.NAA07160@entilzha.cc.iastate.edu>"Tracy J. Di Marco Wh
ite" writes
>
>}On Tue, Oct 03, 2000 at 11:15:56AM -0500, Tracy J. Di Marco White wrote:

>}
>}Why would a journal/log based filesystem be more reliable ?
>
>My husband thought it would be better.  He's started using one of the
>Linux journaling filesystems on his laptop, and his reasoning was such
>that given the testing he's doing, he ends up power cycling it semi
>regularly, he figured that the journaling file system would leave him
>much less vulnerable to corruption.

the line of reasoning here seems to goo something like this:
   * Linux ext2fs  is crap
   * On Linux, log-structured filesystems avoid the ex2fs braindamage.
   * Thus, log-structured filesystems are a win on Linux.
   * Therefore, log-structured filesystems will be a win on NetBSD.

But since FFS with softdeps doesn't suffer from the brain-damage of
totally-async ext2fs, the last step isn't really justified :).

Oh yes, the BSD fsck times are *much* faster, too :-).