[GTALUG] Blank screen can't be refreshed

Giles Orr gilesorr at gmail.com
Tue Oct 19 09:18:41 EDT 2021


Hi folks.

My main machine is a Fedora 34 laptop that drives a 43" 4K Philips
monitor (it's actually a "monitor," not a "TV").  I've set
'xscreensaver' to blank the machine after five minutes.  The delay is
that short because the Philips (while otherwise a nice display) is
notorious for ghosting.  It's a Barrier client to another computer (a
NUC) - so the keyboard and mouse that control this machine are
actually attached to another machine.

Occasionally, the machine/screen can't be awoken from the blank screen
state.  The Philips backlight is on, I have a pointer I can move
(either by Barrier, or by the touchpad on the laptop itself) ... but
the screen remains blank.  The machine is up and running, and I can
SSH in and change stuff, but nothing non-disruptive I've found wakes
the screen.  I tried 'kill -HUP <openboxPID>' to no effect.  I tried
'openbox --restart' but it failed because it's a remote machine (I can
probably over-ride this with an appropriate environment variable(s),
but haven't got to that).  'kill <openboxPID>' - also had no effect.
But 'kill -9 <openboxPID>' brought me back to the login screen
(functioning correctly on the previously blank screen).  This (and
rebooting) are the only solutions I've found so far.  This problem has
usually occurred on wake from suspend-to-RAM, but has now happened a
couple times on wake from just screen blanking.

Having had to write that out, it makes me think that Openbox is the
problem.  Openbox is supposed to accept 'kill -HUP' and do something
sane with it, but it's clearly got itself in a state where it's not
listening to 'kill' at all (thus the need for '-9').

I've looked at the machine logs: I'm not great at reading them, but
I'm not seeing anything that looks like this problem, and I did spend
a fair amount of time looking ...  If it is Openbox, I should be
looking at the X logs, which I haven't done yet?  I've used Barrier
heavily for many years, and this problem hasn't manifested on any of
the several other computers that have been Barrier clients to the NUC
... but should I consider Barrier?  What about xscreensaver itself?

I could potentially compile Openbox rather than relying on the Fedora
package.  I don't want to do that, but if needed ...  Huh - I suppose
I could also switch to Fluxbox for a while, and see if it has the same
issue.

Any ideas appreciated, any thoughts on where to look or how to debug ...

--
Giles
https://www.gilesorr.com/
gilesorr at gmail.com


More information about the talk mailing list