Hi,
what is the price to pay, if I run always with no-lifeguard-reset?
When some crusial system service goes down (e.g. because it was killed
by kernel in out-of-memory situation, or because it was buggy and
crashed) and you not being able to use the device.
Another alternative would be to change the SW watchdog policies in
the init scripts under /etc from reseting the device to restarting
the failing process. For more information, see dsmetool -h
and grep the files under /etc/ on the device.
I've tested a bit what happens when different processes die
and I'm not sure doing this is much of use:
- If X server dies, all it's clients exit (all UI processes)
- If either D-BUS daemon dies, all it's clients exit (all UI
processes and most of others)
- If maemo-launcher dies, about all UI processes (including
desktop) exit/die too
Restarting Desktop and Matchbox window manager would be feasible,
but you might have some funny artefacts (like fullscreen wide
statusbar that pushes all app windows its height downwards).
There might be also some other always running processes which
trigger device reset, please feel free to test the effects of
kill -KILL PID. :-)
- Eero
___
maemo-developers mailing list
maemo-developers@maemo.org
https://maemo.org/mailman/listinfo/maemo-developers
