Using in production

Our Debian, Ubuntu, Fedora and Alpine packages create a liquidsoap system user and group, a log directory /var/log/liquidsoap and a cache directory /var/cache/liquidsoap. The package also fills that cache for the standard library, which speeds up startup.

In production, run liquidsoap in the foreground under a service manager, such as systemd on Linux or launchd on macOS. The service manager starts your script at boot, restarts it if it crashes, runs it as the liquidsoap user and captures its output. Here is a minimal systemd unit, for instance in /etc/systemd/system/radio.service:

[Unit]
Description=My radio
After=network-online.target
Wants=network-online.target

[Service]
User=liquidsoap
Group=liquidsoap
ExecStart=/usr/bin/liquidsoap /etc/liquidsoap/radio.liq
Restart=always

[Install]
WantedBy=multi-user.target

Enable and start it with systemctl enable --now radio. Your scripts do not need the #! line when started this way.

By default, liquidsoap logs to its standard output, which systemd sends to the journal. To log to a file, set settings.log.file := true. The file goes to settings.log.file.path, which defaults to <syslogdir>/<script>.log, that is /var/log/liquidsoap/<script>.log with our packages. Liquidsoap reopens its log file when it receives SIGUSR1, so a logrotate configuration can rotate it with:

postrotate
  systemctl kill --signal=USR1 radio
endscript

Errors in a script are easier to catch before a restart. Check your script after each modification with liquidsoap --check /etc/liquidsoap/radio.liq, then restart the service.