Subject: That ole uVax II again
To: None <port-vax@NetBSD.ORG>
From: Dan Williamson <email@example.com>
Date: 08/04/1997 13:09:45
The continuing saga continues ...
My tribulations so far:
It took 3 tape writes (2 of version 1.2G as I started with 1.2C) to get a
boot tape that worked sometimes from the TK50. Took me a while to work out
how to get a 512K blocked file before writing to tape. Once I finally got
it to boot I went into VMS and did a copy/contig of the tape to my DUA1:
disk mounted foreign. Now I could boot at will quite quickly using the
command b/3 dua1: at the chevrons.
My 2 Disks are -
DUA1: /dev/ra0 MAXTOR XT-2190 (MSCP?)
311200 1221 15 17 512
DUB0: /dev/ra1 Fujitsu M2249E (ESDI?)
632400 1240 15 34 512
edlabeled as follows:
Drive Partition Size Offset Used_as
DUA1: ra0a 25000 0 /tmp
ra0b 35000 20000 (Emergency_Miniroot)
ra0c 632400 0 (Total_Disk)
ra0d 572400 60000 /mnt
DUB0: ra1a 20000 0 /
ra1b 25000 20000 swap
ra1c 311200 0 (Total_Disk)
ra1d 0 0 (Not_Used)
ra1e 266200 45000 /usr
After innumerable attempts with 'COPYING' miniroot from mt(0,0) to
ra(0,1,1) and having it go into never-neverland I decided to allow it to
proceed for a couple of hours and after that time simply stop the TK50
from some kind of endless shuffle that was happening by pushing the red
button. That brought the system alive and it appeared to complete the
'Reading' and was ready to write. I wrote the miniroot hoping that I had
enough depth on the disk to at least give me some start up diagnostics
before it hit 'end-of-story'. As it turned out I had managed to transfer
the miniroot to ra0b and could bring it up fully.
I newfs'd the partitions (left ra0b completely alone as a precaution).
I dragged down the 'snapshot' via binary ftp from ftp.bitcon.no to a PC
and brought up the miniroot. I ifconfig'd the qe0 interface and did an ftp
to the PC. I pulled the gzipped files onto the ra0d partition. From here
on my problems were weird. Over the course of the weekend I gunzip'd and
untar'd all the files in the distribution to the structure as above. The
system would just freeze up (almost as if the drive decided to just stop)
and ^C had no effect so a reboot was required. I'd get 2 or 3 operations
done such as a gunzip or an untar and then the freeze. When the freeze
occurred I know processing still occurred because if I brought up the net
to ftp I'd get over-ring - skipping errors on the interface but no action
related to the gunzipping etc. I set up the fstab accordingly. I set up
the basic files like mygate etc.. I did a disklabel -B to ra1 and fired
up using b dub0:. The bootup froze while executing instructions in
netstart. I rebooted miniroot using b/3 and selected ra1a for root with
ra1b for swap. I re-mounted in write mode and rewrote the netstart to
include only the ifconfig of lo0 and qe0 with associated 'route' entries
and a hosts file. I set name resolution to file first. I again rebooted
using b dub0: and this time it choked in the rc routines after netstart.
After hours of the above trying to tweak a booting system I discovered
that the freezing was a wee bit independent of the files. Sometimes it
would get further than other times with the same bootup sequences.
sorry about the huge preamble but now comes my questions.
1./ Does my setup look reasonably correct (have I screwed up a drive
edlabelling or partition selection)? I can send the list any lines from
the post I get at NetBSD bootup if it helps.
2./ What is the simplest boot up sequence to get to interactive mode (past
single-user) with the qe0 interface running and the standard inet daemons
such as telnetd, ftpd, and others (which are absolutely necessary lines
for networking etc in rc, netstart, etc).
3./ Has anyone else run into this type of fun? If I bring up miniroot and
don't do anything I think it would not die and would respond to console
for days but the installation activity seems to take its toll.
4./ The VAX ran VMS 5.2_2 flawlessly for months of up time and did not
give errors on any of the equipment prior to my loading NetBSD and the
initial diagnostic sequence does fine.
Thanks again for help in the past. I now have a microvax that will run
NetBSD but to make it a practical tool I need to be able to manipulate
files and compile code. This intermittent problem would foil any attempt
to compile a kernel I'm sure.