12.04 no longer boots after apt-get upgrade

[Steve the Pocket]

Hi. I'm one of those who decided to give Linux a go when Valve released their Steam client for it a few months back. I've played around with Linux before in the past (Caldera, RedHat, OpenSUSE, Ubuntu) but never anything serious and never for very long. I think I've learned more about how Linux is put together and how to bend it to my will from this than in the last whole decade combined. Mostly out of necessity. And frankly, I'm surprised I lasted this long before wrecking something. See, despite having a self-updater, Steam can't actually update certain parts of itself. The first time I realized I needed to update it manually, I just uninstalled it and installed a fresh copy from the command line. The second time, someone pointed out that I could just exit the program and run

sudo apt-get update && sudo apt-get upgrade

to automatically update it. What they didn't bother to tell me is that this updates everything on the system, a process that for me took almost an hour. But whatever; I hadn't yet figured out how to update apps from KDE (turned out the update center thing is separate from Muon itself; clicking a popup I got toward the end of the process brought it up) and I was desperate to get GIMP up to 2.8 anyway, so I figured it was two birds with one stone.

And then I complied with its request to restart and... the computer failed to boot up. It made it to the Kubuntu logo with the five blinking dots, and at the point where it would normally go to the login screen, there was a blip and it just froze. Here's what I have gleaned from my searching for a solution so far and asking on UbuntuForums:

• Attempting a package restore from the recovery mode menu did download and install some stuff, but it did not fix the issue.
• Sometimes I would get an ANSI-mode version of the splash screen instead, which goes to an empty black screen when it crashes. This seems to be related in some vague way to what I did the previous time, but the graphical version always comes back the following time.
• Trying to boot to the older version of the kernel listed in GRUB produces the same results.
• It doesn't seem to be a graphics driver issue; "lshw -C display" and "lspci -nnk | grep -iA3 vga" produce what I am told are favorable results.
• "dpkg --audit" returns no results.
• Typing "startx" gives me the error "Could not create lock file in /tmp/.tx0-lock", which is the most promising lead so far, but that's the point where people stopped responding to my thread. Which is why I am now here.

Interestingly, googling the problem led me to a blog entry in which someone had posted the a comment that, in his experience with Kubuntu,

Any time an upgrade to core system packages like kde-workspace/desktop or plasma occurs, the entire system breaks. It removes essential packages and doesn't reinstall them. One has to do this through a recovery prompt.

Sadly the commenter didn't leave any useful contact info, so I can't hunt him down and ask him what packages he restores and how, but since this is a forum of Kubuntu users, hopefully someone knows what he was talking about.

[Snowhog]

Hi.

From the command line type:

sudo apt-get update && sudo apt-get dist-upgrade

[vinnywright]

go back to recovery mode do

sudo apt-get -f install

then

sudo apt-get update && sudo apt-get dist-upgrade

reboot try to log in , if no post any error you got running thos commands .

if no one has responded by tomarow (I gota go to bead now ) I will start trying to debug this with you .

VINNY

[SteveRiley]

Please show the output of the following commands:

cat /etc/fstab

cat /proc/mounts

df

ls -al /

ls -al /tmp

[SteveRiley]

Ah, I see Snowhog and Vinny chimed in while I was typing. You'll note they provided an alternate command. apt-get dist-upgrade is preferred over apt-get upgrade because the former will perform a more complete dependency resolution. If a package update requires installing a new package not currently on your system, upgrade won't do this, while dist-upgrade will. I'm not sure if this is the actual cause of your problem, but I would concur with my colleagues that it's worth a shot.

My troubleshooting suggestions stem from a thought that you might have a permissions problem, you might have a full partition, or your root partition might be mounting as read-only for some reason.

[Steve the Pocket]

OK, this is never going to work. After painstakingly entering the results of the first command and previewing it to make sure the columns line up correctly, I realized to my horror that the monospace font on this tablet isn't even actually monospaced! Congratulatons, Amazon, you've just convinced me to never buy a Kindle Fire of my own. Now if you'll excuse me, I clearly need to go buy a laptop before we can continue this conversation. :mad:

Oh, and for the record, the dist-upgrade thing didn't help. It found some stuff and installed it, but I still ended up with a crash-to-black when I rebooted.

[SteveRiley]

Take photos. For each command:

* type clear
* type the command
* take a pic
* post it somewhere online like Imgur
* paste the URL here

[Steve the Pocket]

Yeah, I realized it would be easy enough to copy off a photo. Here's what happened:

cat /etc/fstab:

<file system>	<mount point>	<type>	<options>	<dump>	<pass>
/ was on /dev/sdb6 during installation
UUID=38c6f79b-ac4f-426d-86fb-1b6aab54709a	/	ext4	errors=remount-ro	0	1
/boot was on /dev/sdb2 during installation
UUID=f971b5d3-0d6e-4f70-8b17-13d682f77b4e	/boot	ext4	defaults	0	2
swap was on /dev/sdb5 during installation
UUID=3f3f3ac7-ce07-49ec-9be8-7a5367536a9a	none	swap	sw	0	0

cat /proc/mounts:

rootfs / rootfs rw 0 0
sysfs /sys sysfs rw,nosuid,nodev,noexec,relatime 0 0
proc /proc proc rw,nosuid,nodev,noexec,relatime 0 0
udev /dev devtmpfs rw,relatime,size=2012416k,nr_inodes=503104,mode=75 5 0 0
devpts /dev/pts devpts rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode= 000 0 0
tmpfs /run tmpfs rw,nosuid,relatime,size=809856k,mode=755 0 0
/dev/disk/by-uuid/38c6f79b-ac4f-426d-86fb-1b6aab54709a / ext4 rw,relatime,errors=remount-ro,user_xattr,barrier=1,data=ordered 0 0
none /sys/fs/fuse/connections fusectl rw,relatime 0 0
none /sys/kernel/debug debugfs rw,relatime 0 0
none /sys/kernel/security securityfs rw,relatime 0 0
none /run/lock tmpfs rw,nosuid,nodev,noexec,relatime,size=5120k 0 0
none /run/shm tmpfs rw,nosuid,nodev,relatime 0 0
/dev/sdb2 /boot ext4 rw,relatime,user_xattr,barrier=1,stripe=4,data=ord ered 0 0

df:

Filesystem	1K-blocks	Used	Available	Use%	Mounted on
/dev/sdb6	492152848	38865700	428287212	9%	/
udev	2012416	4	2013412	1%	/dev
tmpfs	809856	904	808952	1%	/run
none	5120	0	5120	0%	/run/lock
none	2024632	0	2024632	0%	/run/shm
/dev/sdb2	243951	98428	132928	43%	/boot

ls -al:

total 44

drwx------	7	root	root	4096	Mar 15	16:55	.
drwx-xr-xr-x	26	root	root	4096	Mar 21	01:05	..
-rw-------	1	root	root	182	Mar 21	01:06	.bash_history
-rw-r--r--	1	root	root	3106	Apr 19	2012	.bashrc
dxwxr-xr-x	2	root	root	4096	Dec 19	14:42	.config
drwx------	3	root	root	4096	Dec 19	14:42	.dbus
drwxr-xr-x	2	root	root	4096	Feb 10	03:07	.gstreamer-0.10
drwxr-xr-x	3	root	root	4096	Jan 25	12:46	.kde
-rw-r--r--	1	root	root	140	Apr 19	2012	.profile
drwx------	2	root	root	4096	Mar 21	01:09	.pulse
-rw-------	1	root	root	256	Dec 19	13:38	.pulse-cookie

ls -al /tmp:

total 24

drwxrwxrwt	5	root	root	4096	Mar 21	01:09	.
drwxr-xr-x	26	root	root	4096	Mar 21	01:05	..
drwxrwxrwt	2	root	root	4096	Mar 21	01:09	.ICE-unix
drwx------	2	root	root	4096	Mar 21	01:09	pulse-PKdhtXMmr18n
-r--r--r--	1	root	root	11	Mar 21	01:09	.X0-lock
drwxrwxrwt	2	root	root	4096	Mar 21	01:09	.X11-unix

[SteveRiley]

Well, looks like you aren't having any mount or space issues. Thanks for taking the time to supply all that info.

Not sure what to do right next. Need to think about this for a bit.

[dibl]

If it's getting to the point of launching X, and then failing, odds are it is a video issue of some kind. To show us the GPU and driver, run

lspci -vnn

(We only need the "VGA" stanza)Then there is a log file at /var/log/Xorg.0.log that might give some clue as to what is failing.

[Steve the Pocket]

The thread on ubuntuforums told me to run the same command, but with the -nnk parameter instead. I'll just tack on the grep they told me to use, since I assume that's how you say "only show the part with this in it."

01:00.0 VGA compatible controller [0300]: Advanced Micro Devices [AMD] nee ATI M98L [Mobility Radeon HD 4850] [1002:944a] (prog-if 00 [VGA controller])
Subsystem: XFX Pine Group Inc. Device [1682:3000]
Flags: bus master, fast devsel, latency 0, IRQ 16
Memory at d0000000 (64-bit, prefetchable) [size=256M]

And for good measure, the -nnk param also gave me the following:

Kernel driver in use: fglrx_pci
Kernel modules: fglrx, radeon

For what it's worth, that's not entirely accurate; the card isn't a laptop card, and it's one gigabyte, not 256M. Though if Linux thinks it is, that might explain the poor performance I've been getting from it. Also the fglrx driver had to be installed with the -force param because nothing I did was able to remove the old version thorouhly enough for its satisfaction. Just in case any of that is relevant.

[SteveRiley]

Do you receive any errors if you try this:

touch /tmp/blah

[Steve the Pocket]

No. It doesn't do anything, in fact.

[SteveRiley]

Did you check to see if it created the empty file blah in the /tmp directory?

[Steve the Pocket]

Ah. Yes, it did.