[comp.os.mach] Installation of Mach on IBM RT

karp-brad@CS.YALE.EDU (Brad Karp) (10/03/90)

We recently obtained the IBM RT distribution of Mach, including full sources
and the IBM RT Common Lisp implementation.

While following the instructions in the installation guide for Mach on the RT,
I met with only partial success. I booted the distribution floppies just fine,
and mounted the miniroot successfully.

When I ran install.tape, I entered the parameters for our machine's hard disks
(we have an ESDI controller in this box, and three drives: two E70's and an
E114), and answered "yes" when asked whether the partitions should be created
for installation of the tapes. The mkfs commands for the Root/Sys0 tape and
the Usr0 tape went fine (these filesystems were all created on hd0). However,
the mkfs command for /dev/rhd1g failed each of the many times I tried it.
(install.tape attempts to create this filesystem for the Usr1/Sources tape.)
I gave up on having install.tape create the filesystem, and went ahead and
loaded the Root/Sys0 and Usr0 tapes, figuring I could create the filesystem
myself and _then_ load the Usr1/Sources tape onto the system.

I used the CMU disk management utility (DKW or is it DWK) to create a
filesystem on hd1, and when I accepted the defaults I was presented with, it
successfully created a filesystem on drive 1, called /dev/rhd1c. When I tried
to name the filesystem hd1g, it refused to make the filesystem. The problem
here is that install.tape expects the filesystem on drive hd1 to be named
hd1g.

Can anyone offer any advice as to what exactly might be causing this
predicament? I can't see any reason myself why the filesystem can't be built.
Mach appears to be functioning flawlessly other than this (rather significant)
detail.

Thanks.
-- 
Brad Karp, (203) 436-3064 (voice)          | The views expressed in the text
my 386: karp%softshop.uucp@cs.yale.edu     | you have just perused are not my
via Yale CS Dept: karp-brad@cs.yale.edu    | own; rather, they are those of the 
Box 2443 Yale Station, New Haven, CT 06520 | heavenly muse who sings in me. -JM 

bijan@zildjian.chorus.fr (Bijan Forutanpour) (10/08/90)

   When running install.tape (which is on the miniroot floppy),
there is a /etc/disktab or something to that effect, which has disk layouts
for every kind of disk. (Sort of like a /etc/termcap file).  Anyway, when you
run install.tape it calls /etc/newfs which reads from /etc/disktab the disk
information such as disk partition size, begin and end points, block size,
track size, etc to create the file systems.   This process would fail if there
is ANOTHER, different layout of the disk defined ON the disk, via a disk utility
program for example. This second layout causes problems for /etc/newfs.  

SO, you can either:

1) Use the disk utility program to define a layout, AND edit the /etc/disktab
   file on the miniroot floppy to conform to this layout, OR

2) Define an empty or NO layout using the disk utility program and go with
   the configuration on /etc/disktab.

werner@nikhefk.UUCP (Werner Vogels) (10/09/90)

----- News saved at 8 Oct 90 23:06:36 GMT
In article <1990Oct8.114538@zildjian.chorus.fr> bijan@zildjian.chorus.fr (Bijan Forutanpour) writes:
>

.......

>SO, you can either:
>
>1) Use the disk utility program to define a layout, AND edit the /etc/disktab
>   file on the miniroot floppy to conform to this layout, OR
>
.......

Give the entry in disktab a new name and refer to it when the installation
procedure ask for the type of disk. Do not destroy the original entries
in /etc/disktab, they may be very handy at an other momement.

You can still use the minidisk program on your SAUTIL disk from AOS 4.3
to create the diskpartitions if you feel more comfortable with that utility.


Werner H.P. Vogels

Software Expertise Centrum                      
Haagse Hogeschool, Intersector Informatica     tel: +31 70 618419
Louis Couperusplein 2-19, 2514 HP Den Haag     E-mail: werner@nikhefk.nikhef.nl
The Netherlands                                     or werner@hhinsi.uucp