> > Here's how I built a new fs on just the usr partition:
> 
> Thanks.
> 
> > 1) ran mkfs, picked my drive's scsi id
> > 2) mkfs offered three partitions: root, usr, swap. I chose usr
> > 	these are the correct partitions.
> > 	root previously had mkfs run on it
> > 	root had devices built
> > 	root had distribution installed & booted successfully
> > 
> > In other words, the drive was already partitioned & mkfs'd (on root & 
> > usr). I re-ran mkfs on just usr for the reasons indicated below. At no 
> > point did I bother running mkfs on the swap partition.
> 
> And MKFS just trashed your root partition, right?  Hmm, not to imply I
> don't believe your story, but it's kind of hard to believe...  I
> certainly haven't encountered this kind of problem.  I'd suspect the
> SCSI bus cables and terminators, but other than that, I haven't much
> idea.

Actually, I'm a little suspicious of the active terminatino on my New APS
drive, but the consistency of the problem leads me to believe that it's
not hardware: Everything's great until I run mkfs on the /usr partition.
Then it won't boot anymore.

> 
> BTW, you mentioned mount complained to you to run fsck on the usr
> partition.  Did you run fsck after unmounting the usr partition?
> Running fsck on a mounted file system is extremely dangerous.

Actually, I was never able to mount the usr partition, so no, it wasn't
mounted when I *tried* to fsck it. However, the fsck simply quit saying
"BAD SUPER BLOCK: VALUES IN SUPER BLOCK DISAGREE WITH ALTERNATE
floating point exception"

I can mount this partition in the installer, and put directories on it,
and they stay there. I can ls it, etc. But I can neither mount nor fsck
it when I'm actually in BSD. What can I do!

-Avram