roy@prism.gatech.EDU (Roy Mongiovi) (05/15/89)
(I'm not 100% sure of this scenario, but this is the way I reconstruct
the events. There may be other things involved, but I don't think so.)
This last weekend I upgraded from 0.8 to 0.9.
In preparation, I bought a blank od, and initialized it with
"disk -i -t omd-1-all". I backed up all the files I had changed
on my 0.8 hard disk.
I then booted 0.9 and hit the hammer. After the hard disk had
rebuilt, I rebooted 0.9 off of it.
I then inserted my backup optical disk. It automounted under
the name "/me/INITIALIZED".
I thought "this is a pretty yucky name", so I used the disk program
under 0.9 to change it to "backup" (since that's what I plan to use
the od for). This apparently changed the disk type from omd-1-all
to omd-1. Then I unmounted the od and reinserted it to get it under
the new name.
Well, in the course of mounting the disk, it decided that the file
system was dirty and ran an "fsck -y" (-y!) on it.
Unfortunately (and apparently without warning), NeXT has changed the
/etc/disktab description for optical disks. In particular (I think
this is the culprit, anyway), they changed the number of sectors in
a cylinder group. Anyway, the good ol' fsck "fixed" my disk into
oblivion. After it got done mounting, the mount point was a file,
not a directory.
The upshot of all this is that everything I did under 0.8 is gone
(not that that's any big deal, but it is very depressing).
So be careful when using disks formatted by 0.8 under 0.9. There
are some circumstances under which it will eat them.
--
Roy J. Mongiovi Systems Support Specialist Office of Computing Services
Georgia Institute of Technology Atlanta, Georgia 30332-0275 (404) 894-4660
uucp: ...!{allegra,amd,hplabs,ut-ngp}!gatech!prism!roy
ARPA: roy@prism.gatech.edueht@cs.cmu.edu (Eric Thayer) (05/16/89)
I'd like to echo Roy Mongiovi's comment about being extremely cautious when using 0.8 disks under 0.9. Someone at CMU got hammered in a similar way. ...eric