Bug#368711: installation report, Sparc32 with Etch debian installer
Package: installation-reports
Boot method: netboot
Image version: Etch beta2 from
http://ftp.nl.debian.org/debian/dists/testing/main/installer-sparc/beta2/images/
Date: 2006-05-13
Machine: Sun Sparcstation 10
Processor: 2x SM51
Memory: 160M
Partitions: 1=default small /boot, 2=most of drive as /, 4=143M swap
Output of lspci and lspci -n:
(not applicable to SS10, no PCI)
Unit has TGX video.
Base System Installation Checklist:
[O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it
Initial boot worked: [O]
Configure network HW: [O]
Config network: [O]
Detect CD: [ ]
Load installer modules: [O]
Detect hard drives: [O]
Partition hard drives: [O]
Create file systems: [O]
Mount partitions: [O]
Install base system: [E]
Install boot loader: [E]
Reboot: [E]
Comments/Problems:
The installer failed when it tried to download a kernel 2.4.27-3 sparc32
which did not exist in "testing". I was operating with the default debconf
priority which did not get set to low after the crash, so I was not given the
choice to retry using a specific kernel.
I entered a shell and manually installed a 2.4.27-2 kernel via apt-get. That
succeeded.
Exiting the shell and re-entering the installer menu, I could not get it to
let me configure silo with the kernel I had just installed. It kept wanting
to repartition the hard drive. I got a shell again and did a manual silo
configure which didn't make a bootable system until I hand-edited a silo.conf
file.
I tried re-entering the installer by rebooting using priority=low. That got
me lots of prompts, but the installer would not let me avoid repartitioning
and reinstalling, which would have wasted another several hours doing exactly
the same thing except for the last two steps - kernel and silo. I tried
manually making /target and mounting / and /boot there, which let me
eventually make a working silo.
(This had already been my second attempted install on this hardware on the
same day. I had used a netboot installer image that had worked previously on
an SS20 a couple of months ago. It took a couple of hours to go almost all
the way through in downloading and installing, but died when it could not
find something in the current Etch that it expected, perhaps a kernel.)
Even though my hand-installed 2.4.27-2-sparc32-smp kernel and silo do work,
the post-reboot configure step that sets up a root password, a user account,
and task software sets has not run. I'm not sure how to trigger that. I
installed a root password and user account and password, and did accustomed
apt-get installs to set up an xfce4 desktop.
With two complete downloads and configures taking up much of a day, this was
the most frustrating Debian install I have experienced, after about 40. The
previous sparc32 one a few months ago using Etch went very well as I have
reported. This one probably would have succeeded had Etch not contained a
broken "kernel-image-2.4-sparc32-smp" pseudo-package. It would have been
less annoying to recover from that had priority=low been reset after the
kernel install failure, or if there were a way for the installer to be
re-entered with a choice to mount /target and /target/boot instead of only
one to repartition and install again from square one.
Reply to: