I always keep a small "rescue" CLI-only GNU/Linux system in a small
partition. AFTER a successful SL 4.0 install I proceeded to install such a
rescue system (Debian stable) in a 1GB partition on /dev/hda1.

After that SL 4.0, which I keep all in /dev/hda7, with SELinux active,
didn't boot as fsck.ext3 found "bad magic number in superblock" of
/dev/hda1, a partition that was not supposed to be mounted to start with.  I
remembered I ran into a similar problem as I tried to install SL on a system
where Debian was already present.  Then, I noticed that the "/" line in the
SL /etc/fstab was

	LABEL=/	/	ext3	defaults	1	1

What happens is that

1) Debian labels BY DEFAULT all its partitions with the corresponding
mountpoints, even if a Debian system does not by default identify partitions
by label but by device name.

2) Similarly, SL also labels partitions in (exactly?) the same way, but then
proceeds to actually use the labels in building /etc/fstab.

3) The live SL system probably scans /dev/hdxx from lower to higher numbered
partition at boot, locks onto the first one with the needed label (im my
case the wrong one), and security rightly sounds the alarm.


CLEAN FIX
When you install Debian or another labelistically hyperactive distro MAKE
SURE YOU DON'T LABEL PARTITIONS unless needed, or REMOVE/CHANGE labels
before installing SL.


DIRTY FIX
Keep a copy of the automatically generated SL /etc/fstab.  Rewrite
/etc/fstab calling partitions by device instead of label, but DO NOT TOUCH
the entries for removable devices, which are marked "managed" and are indeed
managed by fstab-sync for hot-plugging and automounting purposes.  Remember
that if you delete/add/move non-SL partitions, a SL system with an original
fstab won't even notice, while one with a thus deviceized (is it a word?)
fstab may get confused and not boot.