Subject: 1.5.2 installation anomalies
To: None <port-alpha@netbsd.org>
From: Ted Spradley <tsprad@spradley.org>
List: port-alpha
Date: 01/05/2002 21:35:24
I just installed 1.5.2 on an AS200 and I ran into two little glitches
that I thought I should report.

1.  Everything went normally in sysinst until it tried to newfs sd0a. 
Just reported that it failed.  I picked "exit" from the main menu and
was gratified to get a # prompt rather than a halt or reboot.  That's
what I like about NetBSD, it gives you a chance to fix things when it
fails.

I tried the newfs /dev/rsd0a by hand and it told me I had to change the
bsize and fsize because I have 16056 (or something like that) sectors
per cylinder!  I took a closer look at the disklabel output and that
wasn't the only number that looked weird.

I'm converting this machine back to NetBSD from FreeBSD, and apparently
the FreeBSD disklabel isn't quite compatible, but it's close enough to
confuse sysinst or disklabel.  I just dd'ed 10 Megabytes of /dev/zero
onto /dev/rsd0c and started over.

2.  When I got to the point of starting the ftp, I got a message that my
network wasn't configured right because ping failed.  Back to the #
prompt.  Ifconfig looked OK, so I tried the ping by hand, and got
"Illegal instruction" and a core dump.  After a bit of head scratching I
mv'ed ping to ping-x and created a new ping (#!/bin/sh; exit 0).  Once I
remembered to chmod +x the install completed just fine.

This is an AS200 4/233 with a Quantum SCSI disk on the built-in
controller, an old S3 video board, and an extra ethernet board (vr0). 
The network was configured to use vr0 for the ftp (and the ping).

Thanks guys, it's good to be back!

-- 
Remember, more computing power was thrown away last week than existed in
the world in 1982.  -- http://www.tom.womack.net/computing/prices.html