In v5 & v6 an improper shutdown always
resulted in data loss. Lately, v7, v8,
data have not been lost (maybe I've been
lucky.) However, in the case of filesystems
mounted on mountpoints, their bsdlabel often
becomes obliterated, so one must
Code:
mount -o async -o force /dev/ad10.journal /mnt
to access the data (I forget the
precise other filesystem-types examples, that
specific command is why I've become accustomed
to using gjournal on attached disks, the data
is (usually) still there although the drive cannot
be mounted without the "force" parameter. And
since fast writes to the external drives are often
the cause of the data/bsdlabel corruption/panic,
I've used the "bwlimit" parameter to rsync to
slow the copy down to where it is reliable. (See
other posts).
.....
Probably irrelevant mostly to permanently
attached disks.
.....