[comp.bugs.4bsd] iinit panic in 4.3 installation

whm@megaron.UUCP (12/03/86)

A friend of mine is trying to bring up 4.3 on a 750 and he's getting an
iinit panic upon bringing up the kernel on the mini-root.  He copies the
mini-root to the swap partition using standalone copy, and then boots
the kernel.  The autoconfig stuff seems to go ok and find all the devices
and gets as far as asking for the root device, but when he replies "ra0*"
to this, it gets an iinit panic.  The 4.3 installation instructions say
this panic is due to failure in mounting the root filesystem and basically
imply that it can't happen unless the user has done something wrong,
but I can't see what he's done wrong.

I wondered if the mini-root might have been corrupted somehow, so we tried
copying it to an empty partition using dd and then successfully fsck'd and
mounted it, so it would appear to be ok.

Anybody have any ideas or things to try?  I think the current plan is to
bring up the 4.2 mini-root kernel and try to restore the root dump on
the 4.3 tape with it.  Anybody know if this wouldn't work?
					
					Bill Mitchell
					whm@arizona.edu
					{allegra,cmcl2,ihnp4,noao}!arizona!whm

chris@mimsy.UUCP (Chris Torek) (12/04/86)

In article <1331@megaron.UUCP> whm@megaron.UUCP writes:
>A friend of mine is trying to bring up 4.3 on a 750 and he's getting an
>iinit panic upon bringing up the kernel on the mini-root.

iinit panics are always due to one of three things:

	broken hardware (unlikely);
	a corrupted root file system; or
	the dreaded `user error'.

>I wondered if the mini-root might have been corrupted somehow, so we tried
>copying it to an empty partition using dd and then successfully fsck'd and
>mounted it, so it would appear to be ok.

Chances are your 4.2 swap area is not in the same place as 4.3's `b'
partition on your particular drive.  If you copy the mini-root with
the standalone copy, that should not be a problem, since the standalone
copy and the mini-root were (presumably) made from the same sources
and therefore have the same partition tables.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690)
UUCP:	seismo!mimsy!chris	ARPA/CSNet:	chris@mimsy.umd.edu

kwlalonde@watmath.UUCP (12/04/86)

We had the same problem the first time we booted 4.3 on
a VAX 750 with a RA81 disk: the system prints the usual
autoconfig messages, asks for the root device, then panics "iinit".

This is because we used the boot and copy programs on the 4.2
cassette.  You can't do that with an RA81 disk, because the partition
sizes are different under 4.3.  In particular, the swap partition
starts in a different place.  The effect is that /vmunix is read into
memory intact, but can't make sense of the miniroot file system.

To get around this, extract the /boot and /copy programs from the
4.3 miniroot to /43boot and /43copy, then:

	# halt
	>>>B/3
	Boot
	: ra(0,0)43copy
  ...
	From: ts(0,1)	(or whatever your tape device is)
	To: ra(0,1)
  ...
	>>>B/3
	Boot
	: ra(0,0)43boot
  ...
	Boot
	: ra(0,1)vmunix

That should do it.

--
Ken Lalonde					kwlalonde@math.waterloo.edu
U of Waterloo					ihnp4!watmath!kwlalonde

bzs@bu-cs.BU.EDU (Barry Shein) (12/05/86)

In article <1331@megaron.UUCP> whm@megaron.UUCP writes:
>A friend of mine is trying to bring up 4.3 on a 750 and he's getting an
>iinit panic upon bringing up the kernel on the mini-root.


From: chris@mimsy.UUCP (Chris Torek)
>Chances are your 4.2 swap area is not in the same place as 4.3's `b'
>partition on your particular drive.

Yes, but if this were the case would he have gotten that far into
the boot? The generic kernel is up and asking him for the root partition.

Is it possible it this is not unit 0 he is trying this on? Even if you
use a different drive it might be a good idea to temporarily set the
unit plug to 0 (you said it was an ra) while you do this first step.

It's almost certainly something like that, I've been thru this message
and it always was something like that.

	-Barry Shein, Boston University