Here are some issues you may encounter down the road.

• Connect to MySQL after you start() the daemon. Otherwise only the parent process will have a MySQL connection, and since that dies.. It's lost and you will get a 'MySQL has gone away' error.

• Error handling Good error handling is imperative. Daemons are often mission critical applications and you don't want an uncatched error to bring it to it's knees.

  • Reconnect to MySQL A connection may be interrupted. Think about network downtime or lock-ups when your database server makes backups. Whatever the cause: You don't want your daemon to die for this, let it try again later.

  • Write your own php error handler It should forward all the PHP errors to the log() method, so they end up in your logfile. Otherwise it's very hard to spot errors, since the daemon has no way of writing to your console anymore! See set_error_handler.

• Monit Monit is a standalone program that can kickstart any daemon, based on your parameters. Should your daemon fail, monit will mail you and try to restart it.

I know I'm saying MySQL a lot, but you can obviously replace that with Oracle, MSSQL, PostgreSQL, etc.