Subject: Newbie problems installing NetBSD 1.6.1 on an SS10
To: None <port-sparc@NetBSD.org>
From: Richard Liang <squixy@hotmail.com>
List: port-sparc
Date: 07/25/2003 09:24:34
Hi everyone,

I'm not sure if this is the place for a NetBSD newbie to be asking 
questions, but I've been having some big problems installing NetBSD 1.6.1 on 
an old Sparc 10.  The system has 112MB of RAM, a cgsix framebuffer, one 
424MB internal HD, one 4GB external HD, and a Sun CD-ROM drive.  I am 
installing from the multi-cd1 ISO image I downloaded from 
ftp://ftp.cs.umn.edu/pub/NetBSD/.

The installer boots correctly from the CD-ROM, and on startup everything 
seems to work okay; it identifies everything correctly (most of the time).  
I have two problems.  The major problem is that whenever I try to install, 
at the partitioning stage, once I've specified partition sizes, the 
installer invariably fails at either the disklabel stage or when creating a 
new filesystem on rsd1a (my root directory partition, on the internal HD).  
The frustrating thing is that it gives different messages at different 
times.

The most common failure I get is a message that says
"(esp0:3:0): selection failed; 0 left in FIFO [intr 18, stat 87, step 3]"
followed by many
"esp0: invalid state: 6"

This is often what I get when trying to partition the drive, and sometimes 
even during the initial stage of installation, when it is detecting all the 
hardware.  If this happens, if I try to partition again the installer fails 
to see the hard drives.

When I do a "probe-scsi" in the PROM shell, everything seems to work, but 
when I do a "test-all" it seems to hang on a line:
"Testing /obio/zs@0,100000"

Anyway, the problem could be as simple as old, failed hardware.  When I dug 
the computer out of mothballs it had two internal hard drives, both bootable 
with Solaris installed.  However, somehow the second one died, and I was 
forced to remove it.  The fact that I get different error messages every 
time also makes me suspect that it's bad hardware.  However I have read 
about people getting similar errors on their Sparcstations, due to something 
about tagged queueing (which I don't pretend to understand).  I also don't 
have a proper SCSI terminator, but as the same setup was working under 
Solaris I figured that the Sun drives have internal termination.

Has anyone had similar problems?  And can anyone help me?

Richard

_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE*  
http://join.msn.com/?page=features/virus