Subject: Re: sysinst, two more notes
To: None <jmarin@pyy.jmp.fi>
From: Wolfgang Solfrank <ws@tools.de>
List: current-users
Date: 02/17/1999 12:55:52
Hi,

> Right after the standard install of 1.3.3 (before erasing the first sector
> of the disk), the machine just locked up when it was supposed to start
> loading the kernel.  No messages were printed by the bootloader.

Obviously the BIOS liked the mbr, and the mbr code either was broken, or
it liked the first sector in the active partition, but the code in there
was broken.  Unfortunately, now we cannot determine any more what exactly
went wrong :-(.

> After erasing the first sector and doing installboot, the system didn't lock
> up, but I saw a request to insert a bootable floppy in the drive (this was
> actually a nice animation from the IBM BIOS :-)

That's to be expected, as the BIOS didn't find a bootable disk.

> > Did you erase the first sector of the disk, or of the NetBSD partition?
> 
> Hmm.. I may have used "/dev/wd0d".. I think I used "/dev/rwd0d" on the
> second time.  Anyway, erasing the disk before install _did_ help.

Both devices access the same part of the disk (namely the whole of it).
The difference is only in the method used inside the kernel (that's
irrelevant for the problem at hand).

> I guess I should be running fdisk all the time, not installboot, then?

Depending on the exact nature of your problem, it looks so. Note that
"fdisk -i ..." overwrites you partition table as well in 1.3.x.

Installboot is neccessary however, if, during the upgrade to 1.3.3 or later,
you are changing the NetBSD partition ID from 165 to 169 in order to get
rid of the warning about the old ID.

Ciao,
Wolfgang
-- 
ws@TooLs.DE     (Wolfgang Solfrank, TooLs GmbH) 	+49-228-985800