Re: reiserfs vs ext3fs
Hi,
"Hugo S.Carrer" <hscarrer@efn.unc.edu.ar> writes:
> On Mon, 6 Oct 2003 15:17:35 -0700
> Tony Godshall <togo@of.net> wrote:
>
> > According to Hugo S.Carrer,
> > > On Mon, 6 Oct 2003 14:06:21 +0200
> > > "PF" <nobody@tiscali.it> wrote:
> > >
> > > > Hi,
> > > >
> > > > I found no faq or thread about comparing ReiserFS and Ext3FS
> > > > for use on
> > > > laptops.
> > > >
> > > > Can someone tell me her/his impressions? I'd like to install a
> > > > Debian 3.1 on
> > > > a HP Omnibook and I wish using journaled fs principally to avoid
> > > > fschk's...
> > > >
> > >
> > > I'm using ext3 on my toshiba, no fsck's for almost a year, and
> > > with several
> > > power failures during heavy FS usage. For now its working great, the only
> > > problem that I can't resolve is the commit interval. the HD spins down
> > > after some time to save battery power but spins right back because of the
> > > journal update or something , anyway it didn't happen with ext2 only.
> >
> > ...
> >
> > Check out noflushd.
> >
> I have nolushd running and it dows spin the disk down.
Noflushd does not work (properly) for journaled filesystems. From
/usr/share/doc/noflushd/BUGS
"Journaling filesystems like ext3 or reiserfs bypasse the kernel's delayed
write mechanisms and write straight to disk. Therefore noflushd is unable to
postpone writing of journaling data. As a result, expect lousy spindown
behaviour when working off ext3/reiserfs/jfs/xfs/... partitions. Fixing
the issue requires a) a kernel change to make the journal's update interval
configurable, and b) a new noflushd release to make use of the new feature.
a) is on the ext3 folks' todo list, b) will have to wait for a)."
That said, you can apply a the "laptop-mode" patch, see
http://www.ussg.iu.edu/hypermail/linux/kernel/0305.3/0096.html
which allows to change the journal expire time for ext3, and therefore
lets the disk spin down (at least, that is what it claims). WARNING: I
never used it.
>
> > Also you can mount with option noatime to reduce writes
> > (atime updating means you have a write for every read!).
> >
>
> Doesn't this option reduce the recovery capability of the FS?
I can not comment on this, but, using ANY delayed write mechanism means
that buffers in memory and disk are out of sync for some time, and you
will lose whatever is in memory and not on disk if the computer
crashes. It's a tradeoff: either you keep your disk in sync and no (or
few) data is lost, or you let your disk spin down and out of sync. So I
don't see why one should care about noatime reducing recovery
capability.
Hope this helps,
Jaume
--
Please answer to the group/list. Don't CC me.
Reply to: