Kubuntu upgrade from 8.04 to 9.1 failed, resulting in unbootable system

If you boot to liveCD you can chroot into the nonbooting system and run repair commands / reinstall grub etc etc

If you have sufficient backups you could simply clean install the new OS and restore data.

Thank you actionparsnip, you've offered rather general advice lacking in the sort of details that most computer users would need, but you did make a suggestion.

I hope more can be provided here. If not, then neither Microsoft nor Apple need worry about Ubuntu/Kubuntu. This level of help will not be even close to enough for most computer users.

Yes I do that so as to give direction rather than spoonfeed, "teach a man to fish" mentality. You can use my guideance to find appropriate guides online. Using hollow threas wont make anyone support you more or less, just makes you sound pathetic.

http://ubuntuforums.org/showthread.php?t=157250
http://developer.spikesource.com/wiki/index.php/Article:Ubuntu_how_to_re-install_grub_using_chroot

Your system may not boot but the files are still on the partition, you can manipulate tem using a chroot from the livecd to reinstall whatever is missing, you can even reinstall ubuntu-desktop to fix the OS at large as well as reinstall grub from the liveCD, Windows and Mac cannot do this without external programs yet Ubuntu and many other Linux distros manage it with ease.

I wasn't making any threats, hollow or otherwise, and to remain constructive that's all I'll say about that. In addition, I suppose I properly should have said I do not get a desktop. That situation is what most computer users would consider "unbootable." User perceptions count quite highly for their acceptance of a system. I do not like either proprietary software vendor I mentioned. I want Ubuntu to succeed.

Grub appears to still be working, but the system which was working very well before my "upgrade" is not working as it was now. I was aware "the files are still on the partition," I attached several of them to the previously mentioned bug which I had "naughtily" reopened.

If I reinstall the desktop, as opposed to upgrading it, won't that smash my Kontact data?

I would like to find a way of completing the process which failed. To do that I should make sense of the previously mentioned files I posted as well as 'apt-term.log.' Is there a a good guide to making sense of those files and will any harm come to changing 'apt-term.log' to be readable?

99% of .log files are simple text files so will be readable if you mount the internal partition it resides on in live cd (under the places menu) and opening the file. This will not change it or damage it so you are free o view the logs as you wish

If your Kontact data is important yuo should have it backed up. I suggest you take a fresh backup asap to preseve whatever data is there, if you enter your users home directory you will find lots of hidden folders which start with a period (.) These can be made visible in file browsers using ctrl+. or ctrl+h (i forget which) but you can also backup those to preserve settings too (again if these are important to you they should / will be backed up).

If you chroot to the nistalled system you can remove the proprietary nvidia driver, or even easier, edit etc/X11/xorg.conf (if yu have one) on the installed system (not the live cd environment) and set the driver to vesa rather than nvidia or nv so that failsafe settings are used.

You will need to use:

gksudo gedit

to get write access to the file

When I wrote about making those files "to be readable." I should have been more clear. The live CD was not letting me see some of them, I found this was because of ownership issues. I've since discovered "gksudo nautilus" to get around those restrictions and similar problems and have started rescuing some data which wasn't previously backed up via USB devices.

When I checked etc/X11/xorg.conf I think I may have found a significant problem. I'll copy the part here that looks suspicious to me:

Section "Device"
 Identifier "NVIDIA Corporation NV18 [GeForce4 MX - nForce GPU]"
 Boardname "nv"
 Busid "PCI:2:0:0"
# Driver "nvidia"
 Screen 0
 Option "NoLogo" "True"
EndSection

I'm not an X11 wizard, but it looks to me as though the driver has been commented out. I'm sort of tired. Was your suggestion to replace what appears above as:
# Driver "nvidia"

with this:
   Driver "VESA"
?

I have compared the file with the interesting name:
xorg.conf.dist-upgrade-201002170247
to the file:
xorg.conf
and find that, other than for devices which I do not have which have been commented out and some apparent simplification toward the end of the file after specification of:
Section "ServerLayout"
 Identifier "Default Layout"
also done via commenting out lines,
The files seem virtually the same with the exception of the previously mentioned comment removal of the nvidia driver.

Perhaps the upgrade process did this because it could not remove NVIDIA-GLX, just speculating...

I've been "fishing." I used the intervening time from the posting of my last note above to back-up all my data. I also replaced the driver line in xorg.conf with "vesa" as also noted above. I have some questions about that file as regards all the modes and modeline parameters and whether they'll work correctly with the "vesa" driver, but that can wait.

When I booted I discovered that the problems occur prior to the X11 stage. The system displayed the splash for a while, with the highlighted bar moving back and forth from one end of the progress bar to the other. It then appeared to hang with a black screen and no disk activity. This was similar to what I had seen before.

The machine is an HP Pavilion a320n. It doesn't have a reset button and the way to shutdown outside of software is to hold the power button for a few seconds. After pressing the power button text appeared and I released the button quickly enough that the machine did not power down. (In previous attempts, I had not noticed the text but it had probably been there for a very short time.)

Here is the text it displayed:
------------------------------------------------------
Gave up waiting for root device. Common problems
-Boot args (cat /proc/cmdline)
- check rootdelay= (did the system wait long enough?)
- check root= (did the system wait for the right device?)
-Missing Modules (cat /proc/modules; ls /dev)
ALERT! /dev/disk/by-uuid/cf89fcce-a6d9-4202-858e-f1e9739f37ff does not exist. Dropping to a shell!

Busybox v1.13.3 (Ubuntu 1:1.13.3-1ubuntu7) built-in shell (ash) Enter 'help' for a list of built-in commands

(initramfs)

I tried the first suggestion "cat /proc/cmdline" and got this result:
root=UUID=cf89fcce-a6d9-4202-858e-f1e9739f37ff ro quiet splash

I was perplexed by that and rebooted the live CD, clicked Places on the Jaunty GNOME menu and selected the computer's hard drive, that made the drive appear on the desktop where I brought up its properties panel and selected the "Volume" tab, which gave its UUID as the previously noted: cf89fcce-a6d9-4202-858e-f1e9739f37ff

Something is not working correctly. My guess is the step which identifies device UUIDs during the initialization process. What next?

I found Question #87718 which seems like it might be related to my problem, though I'm not as proficient with recovery mode and tiny footprint editors. I was wondering if I could rely on my hard drive's boot partition being sda1 and using the live CD to "fix" menu.lst and if that would "stick." Have you any advice or a better path to recommend?

After mostly recovering from a nasty cold and having a clearer head, I think my comment above about the "the step which identifies device UUIDs during the initialization process" being faulty or mashed because of the unsuccessful driver update "points the finger" at GRUB. Do you concur?

I'm reading up at https://wiki.ubuntu.com/Grub2 about grub. When I boot from the hard drive it doesn't seem to me as though I have a GUI GRUB, so perhaps I'm not running GRUB2. I will continue reading and working on this. I would like to fix all this without a reinstall of the entire system and also come out of this with a better understanding of what's going on. I have already gained some in that area. I would certainly appreciate your Linux advice since you are far more experienced with this OS than me.

I've been learning more. I no longer blame GRUB since I was able to use it to try several things. I tried using GRUB to add rootdelay=130 (as per http://ubuntuforums.org/archive/index.php/t-987243.html), but found no joy from that. However, it did take longer to come up with the same messages.

I am able to transfer control to several different kernels, none of which seem happy with connecting to the boot drive. But I don't see much point in replacing GRUB itself since its functions don't seem to be a source of problems.

I thought I saw some interesting messages flash by several times, but they go by so fast I haven't made sense of them. I tried using /etc/default/bootlogd to set BOOTLOGD_ENABLE=Yes, but that doesn't seem to work for me, or at least hasn't yet. I saw a mention of /var/log/dmesg in relation to that also, but I didn't find anything in there after my last try.

Any advice?

Some success!!!

I followed the instructions for modifying parameters on GRUB start to root=/dev/sda1 instead of UUID=####...### and it worked, not perfectly, but somewhat.

The startup executed a file system check on the root drive which took a while and looked a little stranger than it did in 8.04, but I suppose that might relate to the video driver. (?)

Once to the desktop, things looked much stranger, since for some reason KDE settled on 1600x1200 which did not work very well, but other options worked worse (and reset after I tried them).

I did manage to see some things sort of work, though I could not really deal with everything since the screen seemed maybe like a Picasso or actually more similar to a Georges Braques painting. I will need to fix a cleaner X11 configuration.

I have my system back working. I don't understand why the boot process does not work using UUID, but since I'm able to get it to work using "/dev/sda1" now, I will worry about that later, probably before adopting GRUB2.

Using VESA and 1024X768 I am not currently exploiting my nvidia card or my wide screen monitor, but that's something I figure in time I can fix myself. I will next attempt to fix my flash plug-in, which isn't working. That along with the nvidia driver and CUPS configuration caused the apparent problems during the upgrade. Other things appear to be working very well, though at lower resolutions than I'd like.

ok then run:

sudo apt-get --purge remove nvidia*; sudo apt-get --purge autoremove

This will remove all the nvidia guff. You can then run:

sudo add-apt-repository ppa:nvidia-vdpau/ppa; sudo apt-get update; sudo apt-get install nvidia-glx-96 nvidia-96-modaliases

This should make things good.

I seem to have a talent for mishaps. Before I got your message above I attempted to update Flash. That seemed to make things weird/bad again.

My sound went from not working in the browser, to not working at all in the sda1 system, except on start-up and shutdown. However, it works fine booted from CD.

In addition, my desktop no longer comes up... the splash screen (air bubbles?) after entering password (without the icons showing progress) is as far as it got last time, though Kontact did come up for a while. No menu bar, no context menus, no command line, nada...

It seems odd to me that Flas could have such a profound effect, so I figure that it must be interacting with something else on a lower level which is notr at is should be, maybe the nvidia driver. However, before I muck things up even worse (my confidence level isn't what it was before using KpackageKit to update Flash), do you believe that executing the apt-get commands you suggested above is still the right thing to do? (Is the PPA relatively stable?) If so, should I boot into recovery mode to do that? I will wait for your advice.

Okay, I waited, but while doing so, I thought about your prior recommendations and I could not figure why what was good advice then would not still be good advice.

So I did the apt-get commands in recovery mode and all seemed to go off well there... apt-get did mention that flash-nonfree would require flashplugin-installer and that nvidia-settings was recommended, but I did not address those issues yet.

When I rebooted after the apt-get I still did not get the normal sequence of the splash screen after logon for KDE where the icons show up for disk, and different levels of hardware and then the desktop. I got only the general splash screen and then after a while the wallet manager and a couple Kontact screens with the splash screen still in the background, no menus, no real desktop.

Something is obviously still wrong. I still have my cobbled together xorg.conf in place. I probably should have run dpkg-reconfigure xserver-xorg (which I just noticed in the xorg.conf) to build that, but didn't.

I will read up on dpkg-reconfigure xserver-xorg and also nvidia-settings to see if nvidia-settings might be required for dpkg-reconfigure xserver-xorg best operation. Thank you for your help so far. Do you have suggestions for a next step?

I installed nvidia-settings from recovery mode (read about it and it is a graphically oriented app, but figured it might be handy).

I also attempted dpkg-reconfigure xserver-xorg (both with and without -phigh). When I did that I got nothing (very slight pause) and a new command prompt... No character dialog box... nada...

This suggestion to remove and reinstall xserver-xorg looks interesting:
http://www.ubuntugeek.com/ubuntu-tiphow-to-removeinstall-and-reconfigure-xorg-without-reinstalling-ubuntu.html

After doing all the commands above as well as attempting the xserver-xorg removal, I started getting the impression that all these commands were not really changing anything. One thing that apt-get "seemed obsessed" about was flashplugin-nonfree, which it complained about after many commands. I revisited Bug #429841 and used the fix suggested there. That almost worked flawlessly and I posted this note on that bug.

-----------------------------------------------------------------------

I had a batch of problems after upgrading to 9.10, one of which was this 'flashplugin-nonfree' nightmare.
The first two commands worked very slick.
"rm /var/lib/dpkg/info/flashplugin-nonfree.prerm"
resulted in new command line
"dpkg --remove --force-remove-reinstreq flashplugin-nonfree"
complained about "very bad inconsistent state" but did it "because --force enabled"
"dpkg --purge --force-remove-reinstreq flashplugin-nonfree"
gave me a warning:
"dpkg: warning: while removing flashplugin-nonfree, directory '/var/cache/flashplugin-nonfree' not empty so not removed."
at the resulting command line I did
"ls /var/cache/flashplugin-nonfree"
which showed one file:
install_flash_player_9.tar.gz

I expect that removing this manually (and perhaps reissuing the "dpkg --purge") will complete the process started with the third command, but as I have spent a great deal of time on this problem I will wait for confirmation of that.

Thank you for your time and consideration.
-----------------------------------------------------------------------

I'm not sure of my conjecture at the end of the above, so I am hesitant to go ahead beyond this either removing manually or not until I get an answer. After that I will retry the commands you suggested. I think that might be all I will need (outside of figuring out the UUID problem).

I feel like I have learned a fair amount working on these problems. I think I may see the light at the end of the tunnel, and this time it probably isn't an approaching train. :)

Okay, I am free of flashplugin-nonfree, and I see how that was plugging my system up but good, and I've started on the list of commands recommended by you several notes bak up-thread. I did the "apt-get --purge remove nvidia*" which removed a large amount of stuff.

Now when I execute the
"apt-get --purge autoremove"
a very large amount of stuff completely unrelated to nvidia also appears in the list of things to be removed. Lots of lib* packages, python stuff, etc.

"After this operation, 1352MB disk space will be freed."

I aborted this as the list seemed mainly unrelated to nvidia... what is all this stuff to be autoremoved?

Yowser thats going to be most of the OS so avoid

I didn't execute the "apt-get --purge autoremove" because of the largeness and seemingly nvidia-unrelated nature to the list of things that it wanted to remove. I did the apt-get install nvidia-glx-96 nvidia-96-modaliases after adding the nvidia-vdpau ppa. I then tried the configuration with the xorg.conf file generated by the upgrade, but without the driver commented out.

It didn't work, got a screen with this:

Ubuntu is running in low-graphics mode
The following error was encountered. You may need to update your configuration to solve this.
(EE) Failed to load module "freetype" (module does not exist, 0)
(EE) Failed to load module "type1" (module does not exist, 0)
(EE) NVIDIA(0): Failed to load the NVIDIA kernel module!
(EE) NVIDIA(0): *** Aborting ***
(EE) Screen(s) found, but none have a usable configuration.

I have the Xorg log file associated with this saved from later dialogs than this one.

The NVIDIA Kernel seems to be missing.

Yeah, I think the --purge autoremove would have 'clipped too much.' I Googled it and found problem statements claiming it removed too much. Perhaps the apt-get --purge remove nvidia* also did.

I recall seeing something with notations in it looking like nvidia-kernel-common (20080825+1ubuntu2) or close to that, especially the stuff in parentheses (as at https://launchpad.net/ubuntu/+source/nvidia-kernel-common/20080825+1ubuntu2) using dpkg -l nvidia* (I'm learning some command line tricks) before --purge remove nvidia*.

After the apt-get --purge remove nvidia* I'm not sure that I've seen something comparable, though I'd've thought that if nvidia-glx-96 required it, then it would get picked up. However, maybe I've got that wrong. I'll recheck to see if it or something like it is there or not.

I issued dpkg -l nvidia* >/var/log/failsafe_list_dpkg-l-nvidia here's the file:

+----+----+----+----+----+----+----+----+----+----+
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Cfg-files/Unpacked/Failed-cfg/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version Description
+++-==========================================-==========================================-============================================
un nvidia-173-kernel-source <none> (no description available)
un nvidia-173-modaliases <none> (no description available)
un nvidia-177-kernel-source <none> (no description available)
un nvidia-180-kernel-source <none> (no description available)
un nvidia-185-kernel-source <none> (no description available)
un nvidia-185-modaliases <none> (no description available)
un nvidia-71-kernel-source <none> (no description available)
ii nvidia-96-kernel-source 96.43.13-0ubuntu6 NVIDIA binary kernel module source
ii nvidia-96-modaliases 96.43.13-0ubuntu6 Modaliases for the NVIDIA binary X.Org drive
un nvidia-common <none> (no description available)
rc nvidia-glx 1:96.43.05+2.6.24.18-27.5 NVIDIA binary XFree86 4.x/X.Org driver
un nvidia-glx-173 <none> (no description available)
un nvidia-glx-177 <none> (no description available)
un nvidia-glx-180 <none> (no description available)
un nvidia-glx-185 <none> (no description available)
un nvidia-glx-71 <none> (no description available)
ii nvidia-glx-96 96.43.13-0ubuntu6 NVIDIA binary Xorg driver
un nvidia-glx-envy <none> (no description available)
un nvidia-glx-legacy <none> (no description available)
un nvidia-glx-legacy-envy <none> (no description available)
un nvidia-glx-new <none> (no description available)
un nvidia-glx-new-envy <none> (no description available)
un nvidia-glx-src <none> (no description available)
un nvidia-kernel-1.0.7184 <none> (no description available)
un nvidia-kernel-1.0.7185 <none> (no description available)
un nvidia-kernel-1.0.8776 <none> (no description available)
un nvidia-kernel-1.0.9631 <none> (no description available)
un nvidia-kernel-1.0.9639 <none> (no description available)
un nvidia-kernel-1.0.9755 <none> (no description available)
un nvidia-kernel-100.14.19 <none> (no description available)
un nvidia-kernel-169.12 <none> (no description available)
un nvidia-kernel-71.86.04 <none> (no description available)
un nvidia-kernel-96.43.05 <none> (no description available)
un nvidia-kernel-common <none> (no description available)
un nvidia-kernel-source <none> (no description available)
un nvidia-kernel-source-envy <none> (no description available)
un nvidia-kernel-src <none> (no description available)
un nvidia-legacy-kernel-source <none> (no description available)
un nvidia-legacy-kernel-source-envy <none> (no description available)
un nvidia-new-kernel-source <none> (no description available)
un nvidia-new-kernel-source-envy <none> (no description available)
ii nvidia-settings 195.36.15-0ubuntu1~nvidiavdpauppa4 Tool of configuring the NVIDIA graphics driv
un nvidia-xconfig <none> (no description available)

+----+----+----+----+----+----+----+----+----+----+

It looks as though I have the nvidia kernel source (maybe because of the PPA), but not the modules. I don't really want to attempt compiling the nvidia kernel if I don't have to.... XD

I bet I need nvidia-kernel-common, but probably more also. Which ones?

Okay, I added 'nvidia-kernel-common' but as I expect that was not enough. I am next going to add nvidia-kernel-96.43.05 as it seems to be more related than the other options... but I somewhat in the dark here...

The large list of packages also continues to appear with the suggestion to autoremove them.

Strange, now apt-get says it can't find nvidia-kernel-96.43.05 to install, perhaps it is the wrong time of day to find it...

I just re-read the question. Did youupgrade from hardy to Intrepid to jaunty to karmic, or did you do a clean install of karmic, or did you skip from hardy straight to karmic using upgrades?

I went from "hardy straight to karmic using upgrades" only it was really only one upgrade process.

It was the only upgrade offered to me on my Hardy system and I had a number of applications that had prerequisites of later than Hardy which I wanted to run.

I have learned a lesson though about keeping upgrade targets closer and upgrading more often.

This method is not supported and why you are getting issues. I strongly suggest you reinstall

Okay, I will reinstall. I was beginning to think that's the conclusion to which this would come. However, that Hardy to Karmic upgrade path was offered to me.

In fact, as I said above, it was the *only* upgrade offered to me. That really should be fixed for others who may still be on your Hardy LTS systems.

'Unsupported' options should not be offered as the only upgrade path. I was on a LTS version and was not looking to be adventurous on that PC. I have other PCs for "adventures." I was not early into the Karmic upgrade either. I waited several months.

I'm sorry that it may seem to you as if I've wasted your time with something that is 'unsupported.' However, the offering of 'unsupported' options should be identified as a *major* bug in Hardy Kubuntu (btw, from what I've seen of KpackageKit, I liked Adept better).

I don't mean to sound scolding above. You helped me out when I needed it and I want make sure that you know I appreciate that and I thank you.

I have gained a fair amount of confidence with Ubuntu over the last month. I've never really liked Windows and had used other systems which I always considered better but which became essentially "abandoned."

I started using Ubuntu mainly because it wasn't Windows. However, Linux, and Ubuntu in particular, has impressed me, though there are still many rough edges for a casual user (like the one which got me).

I have been a 'casual user' of Linux. That has probably permanently changed, as a result of this experience, to a more intensive user. I've also learned to appreciate GNOME more than I did, but I still think I like KDE more. :)

Thanks again for all your help.