Error handler on failed worker

#48 (closed) It's not possible to restart process in not cluster more, so if user want to have server that never crashes(or in another words server that reloads on failure) then clusterMode in config.js must be true