I was working over the weekend to move my rhasspy server instance. But something is off and I just can’t seem to locate it. I even copied the profile from my raspberry pi i had been using and still no go.
So here is what I have.
pi0w running satellite. Here is the config for that one.
and now for the server.
So here is what I am seeing
I have the satellite using the UDP audio to keep the audio local until the wake word is detected.
So when I say the wake word and im monitoring the mqtt broker. I see no audio frames being sent to the broker until the wake word has been detected. Ok, this is as expected. The next step seems to be the problem. I see the audio frames hitting the mqtt broker and I say my request. The frames keep hitting the server, but nothing happens, until it times out, then the audio frames stop and I get the little error sound.
It seems that the server is not picking up the audio frames and processing. When I do a restart on the server, I don’t see it subscribing to any topics that include the site_id = sat1.
The only error I am seeing in the server log is
[ERROR:2022-02-21 15:07:05,413] rhasspyserver_hermes: on_message
Traceback (most recent call last):
File "/usr/lib/rhasspy/rhasspy-server-hermes/rhasspyserver_hermes/__init__.py", line 1260, in on_message
for queue in self.message_queues:
RuntimeError: Set changed size during iteration
[DEBUG:2022-02-21 15:07:05,408] rhasspyserver_hermes: Sent 307 char(s) to websocket
But I am unsure if that’s related to the audio frame or not. I am thinking not, because if I switch my speech to text to local, I am getting basically the same result, so I believe the issue to be on the server side in that it doesn’t seem to be subscribing to satellite topics.
Any thoughts or suggestions?
Thanks!