In message <200101230532.VAA16060@theory-lab2.Stanford.EDU>, Robert Kennedy wri
tes:
>Congratulations on your success. I would like to say that people's
>cautionary statements are not misplaced. Like you, I was successful
>the first few times I played with various disklabel tricks, and I got
>overconfident until one day when I tried something I figured would be
>innocuous, mistakenly applied the command to the wrong device, and
>discovered I had not made sufficient notes to recover the partition
>table for the drive I mistakenly labeled. Nor did I have a recent
>backup. It was most frustrating, so even though your recent experience
>went smoothly, please do use extra caution. There is not a lot of
>redundancy in the structures you are messing with.

Everyone should be aware that it is possible to recreate most of a disk
label by scanning for UFS magic numbers.  If you don't know how to do
this, ask your local wizard.

If enough people ask me nicely, I would be happy to try to go about developing
a utility to try to recreate a whole partition table this way.  I'm not sure
it can *all* be done, because I'm not sure how complete the UFS superblock
is about disk geometry information, but I'd guess one could come very
close.

-s