Debian Squeeze won't boot anymore (after updates) SOLVED

[guy13]

Dear all,

Once again I humbly turn to you for guidance and assistence.

As of yesterday I've lost access to my Debian Squeeze installation. Somewhere during booting the system hangs and all I get is a black screen. The same happens wether I try a normal boot or a recovery boot and then continue from there. What I can do is login as root in a command line environment. My guess is that something went wrong with one of the updates that I allowed to install two days ago. (Don't remember which ones. Sorry.)

The last lines in /var/log/syslog refer to something called mt-daapd. Among those ten or so lines I get the following error messages:

1. mt-daapd: Error loading plugin /usr/lib/mt-daapd/plugins/scc-script.so: plugin declined to load
2. mt-daapd: Error scanning MP3 files: No such file or directory
3. mt-daapd: Error starting web server

I've created /home/media/music which took care of error no.2.
Error no.1 may not be fatal as the system continues to do stuff after that.
Apparently the problem always arrises after error no.3.

Any suggestions are welcome. E.g. can I safely remove this mt-daapd business? It was installed, it seems, three years ago and never caused problems but as it has to do with iTunes I don't think I ever consciously used it. (Unless of course Rhythmbox depends on it somehow.)

Anyway, thanks in advance (as usual),

guy13

[roseway]

mt-daapd has been dropped from Debian Testing because it no longer has active development. It's been replaced by forked-daapd (which is also available in Stable), so I think your best option is to purge mt-daapd from your system, and then, if you find you need it, install forked-daapd in its place.

[guy13]

Hi Eric,

Thanks for the suggestion. I've purged mt-daapd and those error messages are now gone but that didn't solve my problem. Unfortunately there was another error message that I overlooked yesterday. It says:

kernel: PM: Error -22 checking image file
kernel: PM: Resume from disk failed

After that, I get another 23 screens of messages. The last lines of the penultimate screen read:

shutdown: shutting down for system halt
init: Switching to runlevel: 0

And the last screen is a bunch of services exiting and shutting down (ntpd, avahi-daemon, mysqld,bluetoothd, Networkmanager).

Seems like some configuration file got messed up somehow and I'm pretty sure I didn't fiddle with those any time recently. Can something like that happen during a package update?

G.

[nelz]

Hi Eric,

Thanks for the suggestion. I've purged mt-daapd and those error messages are now gone but that didn't solve my problem. Unfortunately there was another error message that I overlooked yesterday. It says:

kernel: PM: Error -22 checking image file
kernel: PM: Resume from disk failed

Is that an error or a warning? It's the sort of thing you get when the system gets confused abut whether it should be booting cold or resuming from hibernation.

Add noresume to the kernel boot options in GRUB (not in the config, just for one boot) to ignore the resume image.

[roseway]

I would be very surprised if an update of Debian Stable had such a result, so long as you haven't installed something from a suspect source recently.

It sounds rather more likely that some kind of hardware problem has corrupted your installation. As you can still get command line access, I suggest that you back up all your personal data while you can, and then run some memory and disk checks to see if this reveals anything.

[Edit] I've just seen nelz's reply, and that sounds good as a first approach.

[guy13]

Again thanks,

Tried the noresume option: added it to the end of the GRUB line that says linux /boot/vm-linuz-something but still no result.

I also ran some diagnostic program from a System Rescue CD that took about 3 hours to decide that all cylinders are ok. That at least sounds good to me.

Then I ran fsck on /home, /boot and / (different partitions). That restored some inodes on / that were blocked or something. I don't remember the exact terms but it sounded like a good idea that might solve my problem. Unfortunately I still get a black screen. Maybe I should have mentioned that the machine doesn't shut down completely. After getting to the black screen there's still some stuff going on on the hard disk for a while (judging by the flickering indicator) and then the led on my mouse starts to slowly pulsate so it looks like the computer is hibernating. Only there's no way to wake it up.

Suggestions are still welcome as they may save me from reinstalling.

G.

[Dutch_Master]

Although much can be learned from healing such a sick system, it all comes down to the effort v/s cost issue: how much time do you want to spend on finding out exactly what went wrong to fully restore normal operation? As Roseway stated: it's highly unlikely Debian Stable will crash if you've installed updates from Stable only. So, what's in the sources.list file (in /etc/apt/)? Also, your recent update may have pulled in a package that caused this hickup, you should examine the list of installed packages to see if there's anything suspicious. If in doubt, check the Debian package webpage: http://www.debian.org/distrib/packages

In order to restore normalcy quickly, consider re-installing the lot (after the aforementioned backups of data have been made!)...

[guy13]

Sounds like reinstalling is going to be my only option. Probably would've done that already but with Debian being so reliable it's been a while since I needed to.

Anyway I'd still like to figure out where exactly the command to go to runlevel 0 comes from. I checked grub.cfg and files in init.d and none of those were altered recently, nor did they look funny. Default runlevel is set to 2 so why does the box suddenly decide to go to runlevel 0?

If someone could tell me where else to look I'd give it one last shot before resorting to reinstalling.

G.

[guy13]

Thanks again to everyone who made an effort.

It turned out that I was barking up the wrong tree. Just before the whole return to runlevel 0 business there was a whole lot of messages alternating like so:

Mar 25 10:16:08 guynux2 gdm3[1600]: WARNING: GdmDisplay: display lasted 0,336475 seconds
Mar 25 10:16:08 guynux2 acpid: client 3944[0:0] has disconnected
Mar 25 10:16:08 guynux2 acpid: client connected from 3949[0:0]
Mar 25 10:16:08 guynux2 acpid: 1 client rule loaded
Mar 25 10:16:08 guynux2 gdm3[1600]: WARNING: GdmDisplay: display lasted 0,337462 seconds
Mar 25 10:16:08 guynux2 acpid: client 3949[0:0] has disconnected
Mar 25 10:16:08 guynux2 acpid: client connected from 3954[0:0]
Mar 25 10:16:08 guynux2 acpid: 1 client rule loaded

Which made me look into gdm. (And which also made me realise that these messages probably came from the disk drive activity I witnessed after the screen went blank, ànd that the return to runlevel 0 command probably came from silly old me cutting the power manually. My bad!)

To keep a long story short: I ended up (re)installing fglrx and everything seems to be fine again.

Humble greetz,

G.