tell me the most ass over backward shit you do to keep your system chugging?
here’s mine:
sway struggles with my dual monitors, when my screen powers off and back on it causes sway to crash.
system service ‘switch-to-tty1.service’

[Unit]
Description=Switch to tty1 on resume
After=suspend.target

[Service]
Type=simple
ExecStart=/usr/local/bin/switch-to-tty1.sh

[Install]
WantedBy=suspend.target

‘switch-to-tty1.service’ executes ‘/usr/local/bin/switch-to-tty1.sh’ and send user to tty1

#!/bin/bash
# Switch to tty1
chvt 1

.bashrc login from tty1 then kicks user to tty2 and logs out tty1.

if [[ "$(tty)" == "/dev/tty1" ]]; then
    chvt 2
    logout
fi

also tty2 is blocked from keyboard inputs (Alt+Ctrl+F2) so its a somewhat secure lock-screen which on sway lock-screen aren’t great.

  • TimeSquirrel@kbin.melroy.org
    0·
    3 months ago

    I’ve set Raspberry Pis to auto-reboot themselves at night if they are being used for headless network services that need to be available 24/7, just to clear out memory leaks or other things that may have gotten locked up. Not sure if that’s duct tape or just a standard practice. They aren’t the most stable things sometimes. They’re known for power supply and SD card issues.

    • leisesprecher@feddit.org
      0·
      3 months ago

      I did this with my sensors running in Pi picos.

      There was some wonkyness with some of the electrical stuff and since I have no idea how to debug that, I just restarted them every 24 hours and at start “drained” all pins by repeatedly reading from them.

      I’m reasonably sure, this setup is cursed enough to kill an electrical engineer on sight, but it kind of works good-ish enough.