On Mon, Jun 7, 2010 at 1:31 PM, Patrick O'Callaghan
<pocallaghan(a)gmail.com>wrote:
On Mon, 2010-06-07 at 12:15 -0400, Gregory Woodbury wrote:
>
>
> On Fri, Jun 4, 2010 at 10:33 PM, Patrick O'Callaghan
> <pocallaghan(a)gmail.com> wrote:
>
> On Sun, 2010-05-30 at 22:05 -0400, Sam Varshavchik wrote:
> > Patrick O'Callaghan writes:
> >
> > > The only recourse seems to be a hard reset as there's no
> reaction to the
> > > keyboard or mouse. I haven't yet tried ssh from another
> machine to poke
> > > around. If anyone tells me what to look for I'll be happy
> to do it next
> > > time.
> >
> > Try the ssh. This sounds like a kernel crash.
>
>
> It happened again so I tried the ssh. I was able to log in and
> everything seemed to be still running, including my usual GUI
> apps (I
> use KDE). Killing kdm, killing X and doing 'init 3' all had no
> effect.
> It's as if the monitor is physically disconnected (or the
> video driver
> is dead). I had to reboot. I'm going to report it to BZ.
>
> I'm having similar problems, what is the Bugzilla number?
I've been waiting for it to happen again so I can back up my report with
actual data, including /var/log/messages, /var/log/Xorg.*, dmesg output,
lsmod output etc. However the bug doesn't want to collaborate and is now
in hiding. I've turned off power-saving and the screen-saver, and it
hasn't happened again, which may be a clue in itself, but if you can
provide the above data, create the BZ entry and report it here so others
can add comments.
poc
It's not a kernel crash, It's an X server crash/hang. It depends on an
interaction between gnoe-screensaver modules, the gnome-power manager and
the X server. The workaround is to select a screensaver module that doesn't
crash/hang and leave it alone.
I have not exhaustively tested all the screensaver modules but have found
two that definitely elicit the bad behaviours on my x86_64 box. (I've not
had the problem on i686 boxes.)
"Ant Inspect" will crash the X server, showing a gdm-greeter when waking up
the monitor.
"Apple II" will hang the X server, requiring a ssh into the box and a manual
kill -9 of the X server to recover without a reboot.
I have filed a bugzilla report (BZ #606136) and invite y'all to add comments
thereunto.
--
G.Wolfe Woodbury