Subject: disappearing disklabels :(
To: None <port-hp300@NetBSD.ORG>
From: Mark F Willey <willey@ecn.purdue.edu>
List: port-hp300
Date: 12/27/1994 19:41:55
Hi, All.

The HP300 system I was playing with last week turned out to be more trouble
than it was worth, so it's off to the parts heap.  I'm going to use the
processor/coprocessor, but let me know if anyone wants the other parts.
It has a bad frame buffer and bad (?) Lan card.  I'm including the output
of the boot at bottom for anyone who's interested.  It was a 68020 + the
frame buffer and LAN card reported errors (maybe because no cable hooked up
to it?).

Anyhow, now I am having a new problem.  I put the disks from that system on
the 375, and my disklabels seem to disappear when I reboot!  I also
relabeled the disk that I used to install (to correct partitions not on
cyl. boundries) and it loses the partition also.

Here's how I disklabel them:
disklabel -W rd1
disklabel -w -B rd1 rd7958A-camel

where the rd7958A-camel entry from /etc/disktab is:
rd7958A-camel|rd7958A-camel|HP7958A-camel:\
        :ty=winchester:ns#36:nt#7:nc#1013:\
        :pa#20160:oa#252:ta=4.2BSD:ba#8192:fa#1024:\
        :pb#32256:ob#20412:tb=swap:\ 
        :pc#255024:oc#252:bc#4096:fc#1024:\
        :pd#255276:od#0:\ 
        :pe#202608:oe#52668:te=4.2BSD:be#4096:fe#512:

At next boot, the fsck complains about BAD MAGIC NUMBER.  So I
re-disklabel, and fsck, same error.  :(  Need to re-newfs.  Then newfs
complains that it cannot rewrite disklabel.  I re-newfs, and it works.
This is the same story for all disks.  The disklabel that it reverts to is
the original one that I dumped on it when installing - on the other disks,
it goes back to no label.  :(  Here's the /etc/disklabel for the two new
disks:
rd7958A-cameldata|rd7958A-cameldata|HP7958A-cameldata:\
        :ty=winchester:ns#36:nt#7:nc#1013:\
        :pa#252:oa#252:ta=4.2BSD:ba#8192:fa#1024:\
        :pb#252:ob#504:tb=swap:\ 
        :pc#255024:oc#252:bc#4096:fc#1024:\
        :pd#255276:od#0:\
        :pe#254520:oe#756:te=4.2BSD:be#4096:fe#512:

Thank you to anyone who can shed some light on this!

The booting, for those interested:
>> NetBSD BOOT HP9000/320 CPU [$Revision: 1.1 $]
Boot: [[[rd16a:]/netbsd][-s][-a][-d]] :- 

Booting rd16a:/netbsd @ 0x97c
526732+24184+66152 [34536+34035] total=0xff9a8cc3  start 0xff90197c
Copyright (c) 1982, 1986, 1989, 1991, 1993
        The Regents of the University of California.  All rights reserved.

NetBSD 1.0 (CAMEL) #0: Tue Nov 29 23:37:46 PST 1994
    root@camel.sc.intel.com:/usr/src/sys/arch/hp300/compile/CAMEL
HP9000/320 (16.67MHz MC68020 CPU, HP MMU, 16.67MHz MC68881 FPU, 16K
virtual-addr
ess cache)
real mem = 7331840
avail mem = 4636672
using 179 buffers containing 733184 bytes of memory
dma: 98620B with 2 channels, 16 bit DMA
hpib0 at sc7, ipl 3
dca0 at sc9, ipl 5, flags 0x1
hpib1 at sc14, ipl 4
ccd0: component sd1c open failed, error = 6
panic: cannot mount root
syncing disks... done
Cannot set battery backed clock

dumping to dev 201, offset 8
dump device bad