Installing avahi-daemon crashes kernel, corrupts filesystem
My system is a Helios single-board computer, with specs:
CPU: Vortex86 SoC (800MHz) - I *think* this is pretty much a 486, I
could be wrong
RAM: 256MB
Swap: partitioned, about 236MB (just what the installer recommended)
Kernel: 2.6.32-5-486
Debian: Squeeze (from installer daily build, updated after install;
this also happens with Emdebian)
Storage is a 4GB flash card (I don't have sdparm or hdparm, and I
can't install either to get more info. If there's another way, let me
know.) It was formatted ext2.
My problem started when I installed avahi-daemon - the system crashed,
and I could not recover to a useable state. So I reinstalled the
system and approached it a little more carefully.
>From a fresh Squeeze install, I installed every dependency of
avahi-daemon (no recommends), and nothing bad happened. When I
installed avahi-daemon itself (aptitude -R install avahi-daemon), it
got to the stage of setting up the package when my screen filled with
what looked like a stack trace:
http://heeris.id.au/stuff/SBCError.jpg
Anything else went past too fast to see, sorry. I also got the
numlock/capslock blinking lights.
I can start in recovery mode, but when I log in as root I see the message
::: No such file or directory
(This is un-Google-able!)
First thing, I run fsck on the root FS, and one of the first problems is:
Unconnected directory inode 98170 (/var/lib/dpkg/???)
I hit "y" a few times when prompted, then Ctrl+C and run "fsck -y"
instead. There's a lot of cleanup, and I reboot.
Now if I run aptitude, I get:
E: 7/.aptitude - mkdir (2: No such file or directory)
E: dpkg was interrupted, you must manually run 'dpkg --configure
-a' to correct the problem.
...and then a segfault from aptitude. Funny thing is, when I run the
dpkg command, it just says
Setting up avahi-daemon (0.6.27-2)
...and back to the prompt. Next aptitude run, same mkdir error and segfault.
Anyway, suffice to say it's all pretty damaged. Another example is, as root:
# groups
groups: cannot find name for group ID 0
I'm not too interested in saving the system, it's a dev environment
and is completely expendable (because I use source control and make
backups, of course). So I can reinstall the system, but then I *would*
like to (a) install avahi-daemon if possible, and (b) figure out what
is causing this, because I'd rather not install some other package in
a few weeks and hose the system again... if I can help it.
I could use some help getting to the bottom of this! I can boot off my
installer USB into rescue mode, and I can do some very limited things
from recovery mode on the partition itself. I'm not averse to
recompiling packages (including the kernel, if there's room). I'll
keep the damaged system around for diagnostics until I can't get any
more info from it.
So if anyone can help, even with a "have you tried <this>..." or
"maybe it's the <something>..." remark, I'd appreciate it.
Thanks,
Jason
Reply to: