Ubtuntu 6.06 flight 7 livecd Installer crashed (using existing partitions)

[hob4bit]

The error report is given below. I tried the Ubuntu
6.06 flight 7 livecd installer using the following
partitions:

hda1: FAT16
hda2: FAT32
hda3: FAT32
hda4: Extended
hda5: ext3 (knoppix)
hda6: swap (knoppix)
hda7: ext3 (intended for Ubuntu)
hda8: swap (for Ubuntu)
hda9: ext3 (data)
hda10: FAT32
hda11: FAT32

I have a AMD Athlon64 3500+ with 1GB RAM,
nForce 4mummyboard, 6600GT graphics and
using the 32bit Ubuntu livecd

I choose the custom install as I want to keep
my existing partition (hda7, hda8) are my testing
partition while hda5,hda6 are for stable Linux.

Anyway, I choose keep partition in the custom
option. The I choose hda7 for "/" and hda8
for "swap". Please forward and forward again
for the summary page.

***BANG*** the installer crashed with error message
below. I abandon install and tried to reboot. The
Ubuntu installer has already screwed up my MBR
even before it tried to copy stuff to hda7. This
is not very nice.

I repaired the MBR using knoopix livecd from a
backup:

dd if=virgin.mbr count=63 of=/dev/hda

Then in Knoppix I got errrors with hda9. I checked
usign Partition Manger 2.40 (FREE Dos opartitoner).
I found out that Ubuntu has deleted all partitions
from hda7 to hda11.

This is just not acceptable!!! If I want to keep
existing partitions the Ubuntu installer should not
touch my partition at all. The MBR should be
update after copying all the stuff from CDROM to
hda8 with a message that this is happening.

Sorry about my problems. I like to be hopefully.
In fact I am almost ready to switch to Ubuntu.
Ubuntu 5.04 and 5.10 were very buggy. I am
hoping that Ubutnu 6.06 will be better based
on Ubuntu promise to support this release for
five years on servers and a nice delay to stabilize
the code.

This bug is a secuirity problem bnecasue it screws
up the partition table.

===============

We're sorry; the installer crashed. Please file a bug report at https://launchpad.net/distros/ubuntu/+source/ubiquity/+filebug and a developer will attend to the problem as soon as possible. To help the developers understand what went wrong, include the following detail in your bug report:

Traceback (most recent call last):
  File "/usr/lib/python2.4/site-packages/ubiquity/frontend/gtkui.py", line 1084, in watch_debconf_fd_helper
    return callback(source, debconf_condition)
  File "/usr/lib/python2.4/site-packages/ubiquity/filteredcommand.py", line 143, in process_input
    if not self.process_line():
  File "/usr/lib/python2.4/site-packages/ubiquity/filteredcommand.py", line 81, in process_line
    return self.dbfilter.process_line()
  File "/usr/lib/python2.4/site-packages/ubiquity/debconffilter.py", line 219, in process_line
    if not input_widgets[0].run(priority, question):
  File "/usr/lib/python2.4/site-packages/ubiquity/components/partman_commit.py", line 123, in run
    ('ubiquity/text/go_back', 'ubiquity/text/continue'))
  File "/usr/lib/python2.4/site-packages/ubiquity/frontend/gtkui.py", line 1463, in question_dialog
    dialog = gtk.Dialog(title, transient, gtk.DIALOG_MODAL, tuple(buttons))
RuntimeError: first member of each text/response id pair must be a string

[hob4bit (hob4bit)]

Oh, I did recover my partitions as I use Partition
Manager 2.40 to dump the sector addresses and
printed it out. I just needed to re-enter the
sector addresses to recover all my deleted partitions

[Blujuice (blujuice)]

I just tried flight 7 and experienced almost the same bug (I just reported it, so please see bug #45771). The error report is exactly the same as yours, but on my computer, the installer fortunately only crashed and touched neither the MBR nor any other partitions.

Before flight 7, I tried beta 1 & 2 and had a very similar (which was reported as bug #39734 and bug #40819), with the only difference that with beta 1 & 2, the installer crashed without even giving an error report. According to the ubuntu-annouce mailing list ( https://lists.ubuntu.com/archives/ubuntu-announce/2006-May/000080.html ), this bug was fixed in flight 7, but on my computer, as just mentioned, the only difference to flight 7 is that the installer now shows an error report when crashing.

[Blujuice (blujuice)]

As my bug (bug #45771) gives exactly the same error report as this bug, I think it's OK to mark both bugs as "Confirmed".

[Colin Watson (cjwatson)]

Thanks for your report. The crash is also bug 43012, and is fixed in current daily builds. Other bugs are filed about the unreliability in the partitioner, and I'll be addressing those today.