[LUGSB] Kernel failures
James Crasta
jcrasta at gmail.com
Fri Feb 5 20:09:25 EST 2010
When I was having issues with my nvidia drivers doing the same, I was
still able to SSH into my machine and subsequently kill X that way,
even though input had seemed to lock up.
Do you have sshd running, and can you still get in, or is the machine
more locked up than just in the input/output part? If you could log
in, you'd be able to run "dmesg" and see if anything is printed in the
kernel ring buffer. It could be a number of things, such as a funky
memory error, kernel panic (though in those cases, it's not usually
possible to still SSH in), OOM error that just couldn't cleanly
resolve, other issues.
For example, when I was playing around with no swap partition mounted,
and compiling chromium in tmpfs one day, with virtualbox running too,
I managed to cause my system to completely hose itself, though I was
able to (slowly) ssh in and fix it. Anyway, lesson learned, and now I
have a swap partition.
Unless it's a true crash, you'd be better off trying to find the
source of your issue than just backtracking to 9.04, as it might be
something unrelated like the graphics card, not necessarily the
kernel.
On Fri, Feb 5, 2010 at 2:04 PM, Peter Beierle <pjbeierle at yahoo.com> wrote:
> hey. after I updated to ubuntu 9.10 from 9.04 the GUI has "froze" multiple
> times, which at first seems to be a graphics card issue, but because the
> keyboard is also unresponsive (i can't refresh the GUI w/ ctrl-alt-backspace
> and the caps/num lock is also unresponsive, although alt-sysrq-REISUB does
> still work) it is likely a kernel failure of some sort. It seems like i
> should therefore either revert back to 9.04, just downgrade the kernel, or
> something else i am not thinking of. any suggestions?
More information about the lugsb
mailing list