On Wed, 19 Nov 1997 21:10:07 -0800  Jonathan Stone wrote:

> Oh... The original intent was that you could dd the diskimage onto the
> raw disk, boot it as the `a' partition, and use sysinst to install
> _into_ the root you're running out of.  I tested this on a machine at
> home; it worked fine for me.  

I just tried this. dd the diskimage on to rz0, boot 3/rz0/netbsd, mount
-u /dev/rz0a /, sysinst.  Select custom partition layout, `root' = 65536
sectory, `swap' = 131072, `/usr' = 819200, `/foo' the rest.  Then:

> Writing the NetBSD disklabel ...
> Creating the new file systems on all 4.2BSD partitions...
> newfs: /dev/rrz0a is mounted on /
>  
> The program "/sbin/newfs /dev/rrz0a" failed unexpectedly with return code Unknown error: 256.
>  
> This is probably due to choosing the incorrect top-level install
> option---like trying to do an Upgrade on a bare disk, or doing a fresh
> Install on an already-running system.  Or it might be due to a
> mispackaged miniroot.  Whatever the cause, sysinst did not expect any
> errors here and the installation has almost certainly failed.
> 
> Check the error messages above and proceed with *extreme* caution.
> Press <return> to continue.
> exec: mount_ffs /dev/rz0a /mnt
> mount_ffs: /dev/rz0a on /mnt: Device busy
> Warning: 1144 sector(s) in last cylinder unallocated
> /dev/rrz0d:     819200 sectors in 514 cylinders of 19 tracks, 84 sectors
>         400.0MB in 33 cyl groups (16 c/g, 12.47MB/g, 3008 i/g)
> super-block backups (for fsck -b #) at:

I also tried a sysinst build from up-to-date sources with the same
result - am I doing something wrong here?

Simon.