search

pimoroni / clean-shutdown

Python daemon to watch a GPIO pin and trigger a clean shutdown.
MIT License
125 stars 40 forks source link

Proposal: Minimum uptime #6

Closed dbbnrl closed 7 years ago

dbbnrl commented 7 years ago

I've been experimenting with the Zero LIPO and clean-shutdown and run into the following scenario:

Sometimes I want to power my Pi Zero through USB power, not the LIPO connector. The Zero LIPO supports this just fine, but in conjunction with clean-shutdown yields a problem, since without the battery connected the low-battery GPIO fires, and clean-shutdown almost immediately turns the system back off.

My proposed solution is to add a system uptime check to clean-shutdown that will cause it not to act until the system has been up for less than min_uptime seconds (a new config parameter). Once min_uptime seconds elapse, the daemon may shut the system down, but this gives the user a grace period to log in and disable the service if we're actually running from USB.

Another option would be to permanently disable clean-shutdown (until restarted) if the very first poll of the GPIO indicates low battery (on the assumption we're actually not using a battery). This would be really convenient when running from USB and avoid a new config parameter, but will behave badly if we're actually booting the system off a low battery.

I'd be happy to submit a pull request if this feature would be welcome.

Gadgetoid commented 7 years ago

These are some good suggestions! I'll see what we can do. I'll summon @RogueM since this script is very much his baby.

RogueM commented 7 years ago

from the README at the top of this repo:

you may add the following to your /boot/config.txt file from another computer: disable_cleanshutd=1

RogueM commented 7 years ago

incidentally, the shutdown_delay should already be 5 minutes? it's not exactly what you are asking, but as default for the Zero LIPO, at least if installed by passing the product parameter, I think it does end up having the same effect?

dbbnrl commented 7 years ago

Huh. Good point about shutdown_delay -- I agree it should be functionally equivalent to my suggestion. Silly of me not to have considered it. I need to go back and figure out why my device was shutting down almost instantly, since I didn't intentionally change that parameter.

Closing this issue since equivalent functionality should already exist.