I've had some time to spend installing the 19990322 snapshot on a PB 145
B,
and on an SE/30.

PB 145 B has 8 meg of ram, and an IBM DHAS-2540 internal hard drive.

SE/30 has 20 meg ram,  a Quantum LPS540S internal hard drive,  and
   Quantum TRB850S and HP C2233-300 external hard drives.  The
    only extension running when I use Mkfs, the installer, and the
booter
     is Mode32.

 All installations were carried out using Mkfs_1.45, NetBSD/mac68k
  installation utility 1.1g, and Booter 1.11.4a1.  For the installations

  described here, the hard drives were set up (counting from low
  numbered sectors) with a small MacOS partition first, the swap
  partition second, and a Root & Usr partition last.  The initial
  partitioning was done by APS 4.01 -- I used this to create
  mac partitions, and subsequently reformatted using Mkfs.

 The installation onto the PB went smoothly.  I installed all of the
non-X
 base distribution sets.  To judge from dmesg, X support for the PB 145
B
  is not in this kernel (netbsd.GENERIC-104.tar.gz), but I've installed
  the screen package (screen-3.7.6.tgz) and I seem to have a usable
   machine.  I need to get battery support into the kernel yet.

  The installation onto the SE/30 has been wildly problematic.

     --first attempt:  the GENERIC-104 kernel onto the Quantum
           TRB850S at scsi 1.  Upon booting, I gave
                 vt220
            in response to the initial query for terminal type.  This
caused
            a seg fault, leaving tset.core in the / directory.   I am
not
             sure that any further data is useful here -- but, for what
it
             is worth, ls -l also triggered a seg fault.  I got a 'swap
device
             busy' error as well.


      --second attempt:  the GENERICSBC-104 kernel onto the
            Quantum TRB850S, again at scsi 1.  The same problems
occurred.

       --third attempt:  the GENERICSBC-104 kernel onto the
             Quantum LPS540S, at scsi 0.  Everything looked good:  I got

               to multi-user mode, looked at some man pages, tried
                editing a text file, got twm up -- ....  and then vi
froze.  The only out was
                to hit the programmer's switch on the side.  This
damaged
                the file system.  I used fsck_ffs, as per the boot
suggestion,
                 but it reported back to me that it had modified the
file
                 system.  I reformatted, and reinstalled -- but the
result
                  of this was that the booter hung at apparently random
                  spots in the boot process.

          --fourth attempt:  the GENERIC-104 kernel onto the
                HP C2233-300 at scsi 2.  Typing 'vt220' in the course of

                  the boot process caused a seg fault, as in my first
attempt.

            --the fifith attempt is ongoing:  the GENERICSBC-104 kernel
                  onto HP C2233-300 at scsi 2.  So far so good -- I am
now
                  installing the X-sets.  This will be strictly
experimental, as
                  the drive is too small for a useful installation.



This has been fun, but I would surely appreciate some advice!  What is
the best kernel to use for my SE/30?  Does the order of the partitions
on
 the hard drive matter?  Is it a good idea to use a monolithic
partitioning
scheme (swap, and then root & usr).  Am I making some other bone headed
mistakes?


Thanks in advance,


D. Troeger
City College