launchd / systemd unit

The daemon is meant to run as a service: starts on login, restarts on failure, logs centrally. Pick the unit file for your OS, drop it in, load it, done.

macOS — launchd

launchd is macOS’s service manager. User agents (run as you, not root) live in ~/Library/LaunchAgents/. They start when you log in and stop when you log out — perfect for a personal backup daemon.

1. Create the plist

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
  <key>Label</key>
  <string>com.unclez.quay.daemon</string>

  <key>ProgramArguments</key>
  <array>
    <string>/usr/local/bin/quay</string>
    <string>daemon</string>
  </array>

  <key>RunAtLoad</key>     <true/>
  <key>KeepAlive</key>     <true/>
  <key>ThrottleInterval</key> <integer>10</integer>

  <key>EnvironmentVariables</key>
  <dict>
    <key>RUST_LOG</key>
    <string>quay=info</string>
  </dict>

  <key>StandardOutPath</key>  <string>/tmp/quay-daemon.log</string>
  <key>StandardErrorPath</key> <string>/tmp/quay-daemon.err</string>
</dict>
</plist>

2. Load it

launchctl load -w ~/Library/LaunchAgents/com.unclez.quay.daemon.plist

3. Verify

launchctl list | grep quay
# com.unclez.quay.daemon  -  ...

tail -f /tmp/quay-daemon.log

Restart / stop

# Restart (reads any plist changes)
launchctl kickstart -k gui/$(id -u)/com.unclez.quay.daemon

# Stop temporarily
launchctl unload ~/Library/LaunchAgents/com.unclez.quay.daemon.plist

”But I want it to run when I’m not logged in”

That’s a system-level daemon (/Library/LaunchDaemons/, runs as root). Don’t do that for backups — the daemon needs your user’s ~/Library/Application Support/com.unclez.quay/ config files, which won’t exist for root. Stay with the user-agent pattern.

If the host needs to run truly headless (a Mac mini that nobody logs into), enable Auto-login in System Settings → Users & Groups so the user agent loads on boot.

Linux — systemd user unit

Same idea: a per-user systemd service that survives reboots and restarts on failure.

1. Create the unit file

[Unit]
Description=Quay scheduled-backup daemon
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
ExecStart=/usr/local/bin/quay daemon
Environment=RUST_LOG=quay=info
Restart=on-failure
RestartSec=10

[Install]
WantedBy=default.target

2. Enable + start

systemctl --user daemon-reload
systemctl --user enable --now quay-daemon

3. Verify

systemctl --user status quay-daemon
journalctl --user -u quay-daemon -f

Survive logout

By default a user systemd unit stops when you log out. Enable lingering to keep it running in the background:

sudo loginctl enable-linger $(whoami)

Now quay-daemon runs 24/7 even if you’re not logged in.

Restart / stop

systemctl --user restart quay-daemon
systemctl --user stop quay-daemon
systemctl --user disable quay-daemon

Windows

Windows isn’t covered yet — v0.1 focuses on the Unix-style service managers. For now use Task Scheduler with quay schedule run-once <id> per schedule, or run quay daemon in a long-lived PowerShell window. v0.2 will ship a Windows Service wrapper.

Health check

The daemon logs a heartbeat every tick. If you want explicit liveness monitoring, the log file’s mtime moves on every fire and the process PID is in the launchd / systemd status.

A simple “is it alive” alert:

# macOS — alert if last log line is more than 70min old
test "$(stat -f %m /tmp/quay-daemon.log)" -gt $(($(date +%s) - 4200)) || \
  echo "Quay daemon log is stale" | mail -s "alert" you@example.com

# Linux — equivalent via journalctl
journalctl --user -u quay-daemon --since "1 hour ago" --no-pager | tail -1

For a real production setup pair the unit with your existing log/ metric pipeline (Datadog, Grafana Agent, etc.) tailing the journal.