On Mon, 25 Oct 1999 Paul.Sawyer@tellabs.com wrote:

> I am trying to install 1.4.1 onto a 750 from TU81 -
> 
> The machine has 6.5 MB of memory, and can currently
> run BSD4.2 from hp style drives.
> 
> It also has 2 ra81 drives (that I want to install 1.4.1
> onto.)

Congratulations!  You're probably the first person to ever attempt to do a
sysinst install on an 11/750.  Very likely you're a brave person and will
be frustrated before you're done.

In theory, it should work, but obviously it's not for you.  I'll work on
some notes for a manual install.  Do you have a DEUNA or DELUA ethernet
module, or just the TU81 tape drive?  And for the benefit of my ignorance,
how much data can tape media for the TU81 hold?

> I have managed to manually (from the 1.4.1 miniroot single user
> prompt) disklabel both of the ra drives, copy the miniroot.fs from the
> tape to ra0a, install the boot block using disklabel -B, and even cold
> boot the miniroot from ra0a ....

Well, if you can then manually newfs the labeled disk partitions, you're
almost home free.  You just need to get the distribution sets onto the new
disk using either a net connection or trucking them over on tape.  From
there you can untar the sets onto the new disk, and perform a few
configuration steps (Run the MAKEDEV script, edit the /etc/rc.conf, and
some other minor tweaks).

If you can provide me with your intended layout of filesystem accross the
RA81 (or RA81s), then I'll work on some install notes.
 
> Sysinst will not work for some reason - it gets to the point where it
> wants to disklabel the install disk, and returns to the first screen
> without reporting any errors.
>
> I have turned on the logging, and no errors are reported, just the
> final /sbin/umount /mnt2. I have turned on the scripting, and the
> script file shows where sysinst was trying to cat the disk geometry
> params onto diskpart in /etc, and then the next thing is the
> /sbin/umount /mnt2. No errors reported.  The thing never goes beyond
> this point.
> 
> Is there a manual install proceedure somewhere that I can follow ? (I
> tried reverse engineering sysinst code and got tired of that pretty
> quick.) 

:-)  It's very ugly.

> Is there some hidden de-bug switch I can turn on to get more
> info about where sysinst is going wrong ? verbose mode or the like ?
> 



--- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----
                   Do not fold, mutilate, or spindle.