kde beta installer crash - now can't boot hdd

[arjay]

I originally posted this in the general announcements area specifically for this beta but didn't get any reply. Sorry if I have cross-posted but maybe someone here will at least take an interest. This is what I posted:

Very disappointed. Live DVD runs fine but cannot install - numerous installer crashes at partition stage. I have a bog standard current installation of Elyssa gnome on a single HDD partitioned as root, swap and home. Nothing unusual and everything runs fine. Decided to try the kde beta.

First install attempt, the partition manager starts up, searches HDD, gets to 46% then hangs for 5-6 minutes.. So I cancel install and try again. This time I get a Code 10 whatever that is. So I try again and this time I get to the partition screen but it shows /dev/sda as empty (I know it isn't because it has my gnome install on it). It wants me to create new partitions but I don't want to do that - I want to use the existing partitions, so I cancel the install.

I reboot and try again. This time I get an "Installer Crashed" dialogue with an invite to report the bug to ubiquity, which I do, including the partman and syslog files. I reboot and try again and get the same crash report. Now cannot get even the installer to start.

Very frustrating. I spent hours downloading the dvd for nothing it seems. I have rechecked the sum and it checks out ok.

EDIT: Just in case it helps, I rebooted and tried again. Same error report but this time I captured the details:

"Installer Crashed"

Traceback (most recent call last):
File "/usr/lib/ubiquity/ubiquity/frontend/kde_ui.py", line 2319, in index
childItem = parentItem.child(row)
File "/usr/lib/ubiquity/ubiquity/frontend/kde_ui.py", line 2359, in child
return self.childItems[row]
IndexError: list index out of range

SECOND EDIT: I find that Installer/PartMan crash has well and truly trashed my HDD. Now I can't boot from the HDD. If I use the live dvd and try sudo fdisk it can't open /dev/sda or sdb, hda/hdb etc. Sudo qtparted says "no device found"??

Anyone any ideas about how I might get access to the HDD again, and get it to boot the previously installed Elyssa gnome version if there is anything left?

Thanks

[Husse]

I originally posted this in the general announcements area

What does the text in red on top of every page in the general announcements area state?

Forum rules
Section reserved for the team. You can reply to announcements here but please do not post new topics.

A good rule for yourself is to obey by forum rules in a forum - right?
If you boot the live DVD and open "Computer" do you see any disks?
And what does

Code: Select all

 ls /dev/sd*

give?
This is hard to diagnose especially remotely...

[arjay]

Husse - I respect your enormous experience and you have helped me more than once in my early days with Mint - you have my thanks for that. However, I beg to differ on this occasion. If you look through the announcement thread you will see numerous similar posts in which people report one bug or another. I don't see much difference between my post and others which are critical. Save that mine was more critical than most, but then I think I have reason to be. I have had bugs here and there in most betas I have tried but they are usually hardware related (screwed up xorg.conf file or no sound or whatever). But I don't expect to have a partitioner fail on me to the extent of leaving me with an unreadable disk. Also, I did look for a forum for this specific beta but could not find one - if there is I apologise. If there is not, then why not?

As to your specific question, your request for more info gave this:

Code: Select all

mint@mint ~ $  ls /dev/sd*
ls: cannot access /dev/sd*: No such file or directory
mint@mint ~ $ ls /dev/hd*
ls: cannot access /dev/hd*: No such file or directory
mint@mint ~ $ 
 

I have tried creating a temp dir and then mounting the HDD but same problem.

I have tried to reinstall elyssa gnome, but the partitioner can't find the HDD. Neither can Knoppix or Fedora FC6's rescue mode - my two normal avenues if I have a hard disk problem. BIOS reports it correctly and Seagate's Diagnostic Tools don't report any physical problems with the disk. There shouldn't be - it is only a few months old. I have a couple of other ideas to try - but it doesn't look good. Fortunately I backed up everything of value before the install, so it is only a matter of finding some way to reformat the HDD and reinstalling an O/S - not mint KDE you may be sure!

EDIT - OK fixed this. Managed to download a bootable linux rescue disk that got the HDD booted. Found that the kde PartMan had screwed up the menu.lst settings for the HDD, locating it as hd1 instead of hd0. Rather surprising since I only have one HDD! There were a few other surprises too, but easy enough to fix once the disk was bootable. For example it had added the following to the menu.lst settings:

Code: Select all

# This is a divider, added to separate the menu items below from the Debian
# ones.
title		Other operating systems:
root


# This entry automatically added by the Debian installer for an existing
# linux installation on /dev/sda1.
title		Ubuntu hardy (development branch), kernel 2.6.24-16-generic (on /dev/sda1)
root		(hd0,0)
kernel		/boot/vmlinuz-2.6.24-16-generic root=UUID=7fa4ec52-7d70-47de-88b3-8be47f8970b0 ro 
initrd		/boot/initrd.img-2.6.24-16-generic
savedefault
boot


# This entry automatically added by the Debian installer for an existing
# linux installation on /dev/sda1.
title		Ubuntu hardy (development branch), kernel 2.6.24-16-generic (recovery mode) (on /dev/sda1)
root		(hd0,0)
kernel		/boot/vmlinuz-2.6.24-16-generic root=UUID=7fa4ec52-7d70-47de-88b3-8be47f8970b0 ro single 
initrd		/boot/initrd.img-2.6.24-16-generic
savedefault
boot


# This entry automatically added by the Debian installer for an existing
# linux installation on /dev/sda1.
title		Ubuntu hardy (development branch), memtest86+ (on /dev/sda1)
root		(hd0,0)
kernel		/boot/memtest86+.bin  
savedefault
boot


# This entry automatically added by the Debian installer for an existing
# linux installation on /dev/sda1.
title		Ubuntu hardy (development branch), kernel 2.6.24-12-generic (on /dev/sda1)
root		(hd0,0)
kernel		/boot/vmlinuz-2.6.24-12-generic root=UUID=7fa4ec52-7d70-47de-88b3-8be47f8970b0 ro 
initrd		/boot/initrd.img-2.6.24-12-generic
savedefault
boot


# This entry automatically added by the Debian installer for an existing
# linux installation on /dev/sda1.
title		Ubuntu hardy (development branch), kernel 2.6.24-12-generic (recovery mode) (on /dev/sda1)
root		(hd0,0)
kernel		/boot/vmlinuz-2.6.24-12-generic root=UUID=7fa4ec52-7d70-47de-88b3-8be47f8970b0 ro single 
initrd		/boot/initrd.img-2.6.24-12-generic
savedefault
boot


# This entry automatically added by the Debian installer for an existing
# linux installation on /dev/sda1.
title		Ubuntu hardy (development branch), memtest86+ (on /dev/sda1)
root		(hd0,0)
kernel		/boot/memtest86+.bin  
savedefault
boot

Just deleted the lot and rebooted.
Anyway, I am going to install a different OS, so it doesn't make much difference now that any new OS install can at least find the HDD. Just a waste of two days of fiddling and fooling. All good experience at the end of the day.

[Husse]

You may answer topics but not create one in the General Announcements forum. Topics not by the team are moved as soon as I discover them
What else did you have on that disk?
This seems like a bug in Ubiquity, a serious one at that
I'm glad the problem was solved and without data loss
My thoughts were going in another direction.....

[arjay]

Husse - consider my wrist slapped. I was in such a foul mood, it could have got itself posted anywhere. Is there a specific forum for the release?

As to the disk, I had a bog standard install of elyssa 5 gnome which was running fine. I was using it to tidy up a near new web site I am working on. All the material for that is on my ISP's server so nothing else on the hard disk except filezilla for FTP purposes. It had been running just fine for months.

I agree there was a serious bug in Ubiquity and/or the partitioner on the kde version. I reported it to the ubiquity bug address including all relevant files (from the crash message).

All that happened was the unhelpful dev emailed me to ask me to run a series of backtraces and who knows what else. Because my HDD was trashed, I couldn't even try for a reinstall. I told him that, so he closed the bug report!! I had already sent him the syslog, partman log and everything else could get my hands on!

So much for trying to help with bugs.

Now I am running debian etch stable and guess what - it is..

Regards

Richard

[Husse]

Now I am running debian etch stable and guess what - it is..

Yes
But if you run the testing (or whatever it's called - I've forgotten) the opposite is true...
Welcome back to Mint

[arjay]

Now I am running debian etch stable and guess what - it is..

Yes
But if you run the testing (or whatever it's called - I've forgotten) the opposite is true...
Welcome back to Mint

If debian testing caused a problem, I'm afraid I would run SimplyMepis before Mint now. Not nearly as advanced but never given me a minute's problem on several PCs from an old twin PIII to something pretty modern. That is the thing with distros - you have to have confidence in it in the first place. One bad experience, a moment of mistrust, and there are 500 other distros to try. And even then there is stil vista - are now I am really fantisising. Aren't we linux folks a lucky lot