On Thu, Nov 14, 2002 at 06:53:08PM +0100, Bert Kiers wrote:
=> On Thu, Nov 14, 2002 at 11:29:59AM -0500, Paul Mather wrote:
=> 
=> > # fsck_lfs -d -y /dev/rsd1c
=> 
=> You cannot use the c partition, it is the whole disk by definition.
=> 
=> Use disklabel to make (for example) an e partition that is the
=> whole disk and make that the 4.4LFS type. Newfs that. The fstype of
=> the c partition should be 'unused'.

LFS doesn't use component labels, so I don't see why using partition c
should present any serious problem.  It hasn't before, and it didn't
cause a problem when I had the c partition as an FFS filesystem
before.  (Also, I don't understand why making an e partition covering
the whole disk [using the same start and end sectors as the c
partition] would be any different---other than stylistically---than
just using c.)

Anyway, I think I discovered the problem.  Looking closer at the
fsck_lfs output, there is this at the end:

# fsck_lfs -d -y /dev/rsd1c
** /dev/rsd1c
sb0 sn=12711, sb1 sn=12711
[...]
25581 files, 254780 used, 0 free
The following duplicate blocks remain: 2231,
[...]


So, it seems that my LFS file system was simply corrupted beyond
the ability of fsck_lfs to repair. :-(

(Sorry for the false alarm, folks!)

I guess I'll be switching back to FFS for that file system.  (Is
anyone using LFS long term, here?  I've always had reliability or
performance problems whenever I've used it.)

Cheers,

Paul.

e-mail: paul@gromit.dlib.vt.edu

"Without music to decorate it, time is just a bunch of boring production
 deadlines or dates by which bills must be paid."
        --- Frank Vincent Zappa