Xref: utzoo comp.sys.att:7137 unix-pc.general:3434
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!egrunix!nucleus!doc
From: doc@nucleus.UUCP (Dave Mundhenk)
Newsgroups: comp.sys.att,unix-pc.general
Subject: UNIXpc weirdness on bootup - into shell
Keywords: boot,disk,shell,idontknow
Message-ID: <5436@nucleus.UUCP>
Date: 31 Jul 89 20:22:42 GMT
Distribution: na
Organization: The Nucleus Public Access Unix, Clarkston, MI
Lines: 79

[Pardon me if this posting appears more than once. Netnews was
acting screwy on holin and I didn't think it ever made it out.]

HERE WE GO AGAIN! %-{

About 2 weeks ago I posted a plea for help on my 7300 system.
(Actually, I think I only posted to comp.sys.att and unix-pc.general).
It had been difficult to boot and when it did the hard disk was
hosed. I reformated the disk and did a full restore. I had to
use 'afio' to restore some trashed cpio archives (thanks to all
who helped).

After another crazy week or so, I finally got to start my system
up again, and walked back downstairs to eat my dinner. When I come 
back, what does it tell me? "Panic, init died. Press hardware
reset."  Lovely.  I reboot and watch this time.

The loader seems OK. Then it says:

"Version 3.5.1.4
Real memory      = 1048576
Available memory = 696320
Main board is P3...P5you have mail
#                                 "

I got the '#' prompt within about 3 seconds after the memory
info appeared. Most commands don't work - e.g. "ls: cannot execute"
Same with mount, fsck, etc. On top of that, "cd /etc" yields
"/etc: bad directory".

Ok, so it looks like part of my filesystem is hosed. Especially bad
is that it seems like /etc is completely gone, along with init, rc,
and all the really handy stuff. Cpio doesn't work, so I can't restore
my backup of /. I have backups of any other important stuff.
I ran diagnostics on the disk and main board, with no problems.
BTW, the disk is a CDC 94155-36 of unknown history...and it's being
powered by an external 150W PC power supply, since a while back the
built in power supply decided not to spin it up any more.

My questions are: 
1. Do I have any options other than a full restore?
2. Does it sound like this disk should be retired to a quiet
   green pasture, like Messy-DOS?
3. Does it sound like this machine should be retired, period?
4. Might it be a good idea to have an extra copy of /etc somewhere
   else in the filesystem in case the first one gets trashed?
5. Does anyone have any ideas for what I can do to avoid this again?
   I *like* this funky machine, but I only get to use it about once
   a week, and don't want to have to do a full restore every time ;-}
6. Will either of the following drives work in this beast: ST225 or
   ST412. (I have these lying about.)

*UPDATE*: Since I didn't get any replies and thought my posting never
made it out to the net, I decided to take matters in hand and did a
full restore. Using the advanced diagnostics to reformat the disk,
it found another bad block. Foundation set is installed and works fine.

A QUESTION: If you decide to use multiple filesystems [the adv.diags.
let you define multiple partitions] how do you keep the installation
from creating one big filesystem across partitions? I defined two
partitions: fp002=9000 blocks and fp003=~9000 blocks. After installing
the Foundation and logging in, 'df -t' shows ~11000 blocks free, total
of 18000. 

It appears that the install of the Foundation set ignored the
partitioning and made its own 18000 block filesystem. Forgive me if
these questions seem naive, but I am more familiar with 3B2's and it 
took me a while to figure out "iv -t /dev/rfp000" is roughly equivalent to
"prtvtoc /dev/rSA/disk1" for example. 

Where in the Foundation set install does the mkfs take place (floppy
boot?) and what can I do to change the defaults?
[postnews was hosed on holin and I had to post from nucleus.
 please respond to me on one of the systems below if possible.]


Thanks,
Dave Mundhenk
...!att!{lc15a,holin}!doc