--- name: stack-power description: Safely stop the media stack and unmount the media drive before it is powered off or unplugged, or bring the stack back up; check the UPS and run its power-cut tests. Use only when the user explicitly asks to turn off, unplug or restart the drive or the whole stack, or asks about the UPS, a power cut or a UPS test. --- # Stopping and starting the stack This interrupts everything: playback, downloads and requests. **Get the user's explicit yes first**, and check that nothing is playing (see the `stack-update` skill for the Jellyfin sessions check). ```bash STACK=${MEDIA_STACK_DIR:-$(git rev-parse --show-toplevel 2>/dev/null)}; [ -x "$STACK/scripts/stack-health" ] || STACK=~/homelab-media-stack ``` ## Power the drive off safely ```bash "$STACK/scripts/drive-off" ``` This stops the stack, flushes writes and unmounts the media mount (`STORAGE_MOUNT` in `.env`, default `/mnt/storage`). Only then is it safe to cut the enclosure's power. A backup drive (`BACKUP_UUID` set, mounted at `BACKUP_MOUNT`, default `/mnt/backup`) is read-only between mirrors and needs no unmount when only the media drive is powered off. If both drives share a dock that is being switched off, also stop the mirror and unmount it. The backup drive spins down after 10 idle minutes, so it may be asleep; that changes nothing, it still needs both steps: ```bash sudo systemctl stop arr-mirror.service sudo umount /mnt/backup # BACKUP_MOUNT ``` ## Bring it back Power the drive on and it mounts by its UUID; `arr-stack.service` then starts the stack on its own. To start it by hand: ```bash sudo systemctl start arr-stack.service # or: "$STACK/scripts/stack-up" ``` `stack-up` refuses to start without the drive, re-attaches qBittorrent and slskd to the VPN, and syncs the forwarded port. Afterwards, run `stack-health`. ## UPS (only where `/etc/apcupsd/apcupsd.conf` exists) Status and history are safe to read: ```bash /usr/sbin/apcaccess status # STATUS ONLINE, ONBATT or COMMLOST; BCHARGE, TIMELEFT, MBATTCHG, MINTIMEL tail -20 /var/log/apcupsd.events # power events with their times systemctl is-active apcupsd; systemctl list-timers arr-ups-metrics.timer --no-pager ``` At 10% charge or 5 minutes left, apcupsd shuts the server down, the UPS cuts its outlets, and the server boots by itself when mains returns. `COMMLOST` means a power cut would not shut it down: check the USB cable and that `DEVICE` is empty in `/etc/apcupsd/apcupsd.conf`. `docs/flows/power.md` has the whole sequence. The tests cut power. **Ask first, every time**, and check that nothing is playing: - **Short test:** the user unplugs the UPS from the wall for 60 seconds. Expect "on battery" and "power restored" pushes and everything staying up. - **Full test (shuts the server down):** set `BATTERYLEVEL 95` in `/etc/apcupsd/apcupsd.conf`, `sudo systemctl restart apcupsd`, then the user unplugs the UPS. Expect a clean shutdown and the UPS output going off; plugged back in, the server boots. Afterwards check `journalctl -b -1`, run `"$STACK/scripts/host-install"` to put `BATTERYLEVEL 10` back, and `sudo systemctl restart apcupsd`. If you run on the server, your session ends with the shutdown: tell the user these follow-up steps first. ## Rebooting the server Everything comes back automatically: the drive mounts, Docker and Tailscale start, and `arr-stack.service` brings the stack up. If you are an agent running on the server, your session ends with the reboot, so tell the user to run `stack-health` once it's back up.