Re: Electricity Cutoffs, EXT3 and Filesystems
On Mon, Nov 03, 2008 at 11:47:40AM +0200, Volkan YAZICI wrote:
> On Sun, 2 Nov 2008, Andrei Popescu <firstname.lastname@example.org> writes:
> > Please post an /etc/fstab and the output of 'dumpe2fs -h' for one of
> > your filesystems (considering that they have been created with same
> > options).
> I've attached related /etc/fstab and "dumpe2fs -h" outputs for R&D and
> PREPROD servers.
> # /etc/fstab: static file system information.
> # <file system> <mount point> <type> <options> <dump> <pass>
> proc /proc proc defaults 0 0
> /dev/sda3 / ext3 noatime 1 1
> /dev/sda4 /home ext3 noatime,nodev,nosuid 0 2
> /dev/sda2 /boot ext2 noatime,nodev,nosuid 1 1
> /dev/sda1 none swap sw 0 0
> /dev/md0 /srv/fs ext3 noatime,nodev,nosuid 0 2
> /dev/scd0 /media/cdrom0 udf,iso9660 user,noauto 0 0
If you mount ext3 with the data=journal option you may have better
recovery. See the man page. data=journal is what ext3 used to do but
the default was changed to give better performance.
My experience with jfs was good. I never had the system die (not be
able to recover) and I don't think I ever lost data (I too don't have a
UPS and I used to live where the power was very unreliable). JFS
recovers fast (a design criteria). However, do a google search on
lists.debian.org for JFS and read the thread where it has been
discussed. In the midst of that thread, I received an email from the
maintainer of jfstools which forwarded an email from the one person at
IBM who is left to watch for bugs: he was recommended agains JFS for new
projects since IBM pulled the rest of his team off of JFS.