Subject: RE: sysinst test (unsuccessful)
To: Roger Fischer <rogerf1@mac.com>
From: Bob Nestor <rnestor@augustmail.com>
List: port-mac68k
Date: 12/17/2000 00:12:48
Roger Fischer wrote:

>>RFischer@PanAmSat.com wrote:
>>
>>>Standard Apple keyboard (not extended though).
>>>MouseSystems 3 button mouse (Which is recognized and I've been using
>>>under NetBSD 1.4 for quite a while.
>>>Centris 610 with full 040 processor.
>>>Quantum Fireball 3.2GB drive.
>
>>  I'd suggest using a standard Apple
>>mouse to start with.  I do know I had problems with the A3 mouse that we
>>thought had been fixed, at least it now works fine for me and a couple of
>>others.  Maybe there is still a problem on fast systems like yours though.
>
>Well Bob, you're right again.  I swapped out my MouseSystems A3 Mouse,
>and sysinst booted fine.  I managed to wipe out my disk with the sysinst,
>but there was nothing important on it anyway.  A lot of software I've
>compiled in the past, but nothing I can't do again.
>
Ah! I know what's going on then.  The A3 Mouse problem _IS_ fixed in the 
1.5 Release but is not in the 1.5 Installation Kernel.  You'll notice 
that the Installation Kernel identifies itself as an ALPHA or BETA 1.5 
version.  Basically the A3 mouse problem was fixed very late in the 1.5 
Release cycle and was verified to be working on some problem systems, but 
the Installation Kernels weren't rebuilt to incorporate the change.

On the sysinst front the news is similar.  There were two versions of the 
sysinst code. The original was used in the 1.4.2, 1.4.3 and 1.5 
Installation Kernels.  It has problems with disk size and never checked 
to make sure the in-core and on-disk Maps were the same before 
initializing the disk.  This in most cases causes disk corruption.  The 
newer version of sysinst code is in -current and it verifies the two maps 
are identical before proceeding.  Unfortunately that code didn't make it 
into the 1.5 Release due to another problem.

There was some discussion about not including Installation Kernels in 
1.4.3 and 1.5 because of this.  I didn't want them theree until the 
problems were resolved, but finally agreed it would be OK if a warning 
was included in the Release Notes.  Others volunteered to field the 
questions that would come up from users should they attempt to use 
sysinst too.  (Who remembers this promise?)

>The installer didn't like the partition map currently on my disk, and
>says there are all kinds of errors.  It has two sd0g partitions, and it
>says that some partitions are past the edge of the disk.  I'll have to
>dig up something external to boot off of so that I can have a look at
>it with MacOS side tools (HD toolkit) and see if it reports the same
>errors before I choose to reformat completely.  I'd like to keep my
>Mac OS partitons.

Yes, this is consistent with the first version of sysinst code.  I'd 
suggest you get your disk partitioned with a good 3rd party tool (like 
Apple SC Setup), and only use sysinst to populate the filesystems.  Don't 
use it to partition the disk at this time!  Hopefully the problems with 
disk partitioning will be resolved for the 1.5.1 Patch Release.

-bob