strange kernel messages, problems installing system updates

I'm not sure what the problem is here. When i try to install certain packages, or mainly when installing updates it will fail part way through.

I just started using this system again so I will have to start paying more atttention to what is going on. Like i said, this system was set up a while ago, and i've messed around with it under the influence but i haven't used it for a few weeks because of a power supply issue (this is besides the point, i know that has nothing to do with this)

i had noticed strange messages before but never took too much note of it but this is what i get by running dmesg -k -T | tail right after a package failed to install:

acenyc@ubuntu:~$ dmesg -k -T | tail
[Fri Jun 20 17:43:50 2014] <6>[fglrx] Firegl kernel thread PID: 1886
[Fri Jun 20 17:43:50 2014] <6>[fglrx] Firegl kernel thread PID: 1887
[Fri Jun 20 17:43:50 2014] <6>[fglrx] IRQ 100 Enabled
[Fri Jun 20 17:43:50 2014] <6>[fglrx] Reserved FB block: Shared offset:0, size:1000000
[Fri Jun 20 17:43:50 2014] <6>[fglrx] Reserved FB block: Unshared offset:f87c000, size:484000
[Fri Jun 20 17:43:50 2014] <6>[fglrx] Reserved FB block: Unshared offset:fff8000, size:8000
[Fri Jun 20 17:43:50 2014] <6>[fglrx] Reserved FB block: Unshared offset:ffff4000, size:c000
[Fri Jun 20 17:43:54 2014] Adding 8282108k swap on /dev/mapper/cryptswap1. Priority:-1 extents:1 across:8282108k FS
[Fri Jun 20 17:58:57 2014] audit_printk_skb: 117 callbacks suppressed
[Fri Jun 20 17:58:57 2014] type=1400 audit(1403301538.346:69): apparmor="DENIED" operation="capable" parent=1 profile="/usr/sbin/cupsd" pid=1320 comm="cupsd" pid=1320 comm="cupsd" capability=36 capname="block_suspend"

the stuff in bold is the stuff i'm interested in. i'm hoping lode, or someone else could help me figure this out

Also, ran ~$ dmesg -T | grep -i 'cups' | less

and got these msgs, and looks interesting:

[Fri Jun 20 17:58:57 2014] type=1400 audit(1403301538.346:69): apparmor="DENIED" operation="capable" parent=1 profile="/usr/sbin/cupsd" pid=1320 comm="cupsd" pid=1320 comm="cupsd" capability=36 capname="block_suspend"
[Fri Jun 20 18:12:58 2014] type=1400 audit(1403302378.378:70): apparmor="DENIED" operation="capable" parent=1 profile="/usr/sbin/cupsd" pid=1320 comm="cupsd" pid=1320 comm="cupsd" capability=36 capname="block_suspend"

it looks like it has something to do with CUPS but i have no idea why or what i should do about it. and how should i investigate this

i know HF probably isn't the place to ask stuff like this, but i know we have a few people knowledgeable of computers here, and i like HF and maybe we can all learn something

 

do you run kde? try upgrading cups

 

this ones running unity. i tried reinstalling cups but it failed

got the same type kernel messages too right after the failure:

 

That's a permissions issue. Cups isn't running as nobody and apparmor isn't letting it run properly in the background because of it. Realistically it should be configured by default, and you should only change cups permissions when needed.
try:

if it doesn't work, post a log of /etc/apparmor/usr.sbin.cupsd
A word of caution, apparmor is very deep and necessary, do not touch it with remove or purge.

 

^there you are buddy

apt-get gave me the following output when i attempted to purge

and for ../usr.sbin.cupsd which doesnt seem to exist

^that's all that's there, hmm what to do about this?

i guess it got rid of that file before it failed huh?

now i got this dpkg error when i tried apt-get remove on cups

Edit: i updated some packages, i'm gonna restart the system, try some stuff again, RTFM, etc. hope to hear a reply from you lode. i think there's a couple issues here maybe that are somewhat intertwined. fixing this will probably demand 100% of my attention for a good part of a day probably

 

run

That'll tell you the kernel verion.

Find /var/log/dpkg and post up the log.

try:

You've got a few kernel issues too. After you clean dpkg it should be

That'll tell you the available kernel versions. It's likely you need to upgrade it, which will be easy once you get dpkg going.

 

I had kernel issues too but doing a shoot dump and soft reset of var log cleaned it up great.



edit...had to remove that gif it was just a little too much

 

okay i'm going to read the contents of dpkg.log & dpkg.log.1, i sent 40 lines of each to a file so i could compare them and make sure they're not identical, and they're not so here

dpkg.log :

dpkg.log.1:

any more lines than that would fill this whole page i think, should i grep it for something specific?

okay, now i'm gonna cleanup packages, and attempt to fix dependencies ............. and that exited with the same error (apt-get install -f) , same kernel issue

reconfigured dpkg and tried to fix dependency issues again still failed the same way

oh man!

BBAD: I really wish you hadn't posted that

edit: Linux version 3.11.0-20-generic (buildd@allspice) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) #35~precise1-Ubuntu SMP Fri May 2 21:32:55 UTC 2014

^ that's what im currently running

 

Try the end of dpkg status

That should be the broken packages. After you remove those packages with apt, you should be able to clean apt with the above method.

 

okay, there's a bunch of crap in dpkg status, but why exactly do you think those packages are broken? i need to understand why. i don't think i've ever had this happen before. system seems to mostly work but somethings off and i need to figure out how to fix it.

i think it would be wise for me to get some sleep before i did this, because im pretty sleep deprived and i keep having to remind myself where i started from.

 

Not all the packages are broken, but the ones that are will be listed in there. grep it for unknown and broken. You said it yourself.

Dpkg, Debian's package manager is broken because it has dependencies which weren't met. It happened to me last week installing Apache and Nginx on the same computer.

Usually you can just clean it and reboot. If not the files that are broken will be listed in /var/lib/dpkg/status. You'll have to read and just delete the ones that say broken.

And if that doesn't work out, you have backups in /var/backups. You can run ls -lta to see the dates of the backups, and restore dpkg to a known good state.

Here's something from ubuntu's forums.
http://askubuntu.com/questions/3436...r-parsing-file-var-lib-dpkg-status-what-to-do

 

i couldn't find anything that looked broken in /var/lib/dpkg/status

it also wouldn't let me touch those backups, even after changing the permissions with chmod, so i changed the permissions back to the way they were

all of this has something to do with the kernel images

:/

 

well, haven't been to this thread in a while but the issues do seem to be resolved

i did all of the things that seemed logical to me, but was still getting errors.

i won't go into the detail of how i came to find this out, basically something caught my eye as walls of text flooded my terminal. i then started grepping the outputs of several different operations that were failing, and reading the output pretty much fully, starting at the beginning, instead of the end or where it appeared things started going wrong.

turns out it was one line in /etc/default/grub

^that was the problem, idk what "to" is there for, plus it looks redundant except for the second one is quotted, which isn't necessary anyway ... the quote characters are wrong too, but i dont think think the got that far, i think it started tripping at "to"

i have no clue how it got that way, i certainly didn't put it that way but after editing that file and commenting out that part and running apt-get build-dep for a bunch of different packages again, apt-get clean and apt-get autoclean again i tried running apt-get update again and it finished without error

you can't modify /etc/default/grub w/o running as root, .. idk man, idk how that happened ..