No text-to-speech resources are available when a VoiceXML document attempts to play a text-to-speech string. This can occur if not enough TTS engines are defined in the procman.cfg file on the server to service the number of active calls.

Ensure that the system frees up each text-to-speech engine immediately a text-to-speech process ends, rather than when the call ends. To do this, stop the node with dtjstop, add the attribute Engine_Free=True to InitSessionString= in the default.cff configuration file, run dtjconf and start the node again with dtjstart.