How can I disable Pulseaudio system mode and use the normal systemd per user services on a headless server?

Question

I have tried a lot of different suggestions from both askubuntu and other sources, but I'm really stuck at the moment.

I'm trying to configure a headless server with the following setup:

1. Line out which goes to the speakers.
2. Line in which should be "forwarded" to the speakers regardless of anything playing on the server.
3. A MPD instance which outputs to snapcast, which outputs to the "line out" speakers.

I have come to understand that I need a combination of Pulseaudio and ALSA for this to work, because ALSA doesn't really support multiple audio sources on the same sound card.

For some reason, when I install pulseaudio, it is started in system mode and I can't figure out why. Htop shows the following processes (all as root user):

s6-supervise pulseaudio
pulseaudio --system -vvv --disallow-exit --exit-idle-time=-1 --disable-shm
pulseaudio --system -vvv --disallow-exit --exit-idle-time=-1 --disable-shm
pulseaudio --system -vvv --disallow-exit --exit-idle-time=-1 --disable-shm

But there is no systemd service that starts this as far as I'm aware:

$ sudo systemctl status pulseaudio.service
Unit pulseaudio.service could not be found.
$ sudo systemctl status pulseaudio.socket
Unit pulseaudio.socket could not be found.

I'm running Ubuntu server:

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 20.04.1 LTS
Release:        20.04
Codename:       focal

Alsamixer shows all playback sources as unmuted (green 00). But I can't connect to Pulseaudio:

$ pulsemixer
Failed to connect to pulseaudio: Connection refused

Even though all users have access to pulse-access:

$ grep pulse-acc /etc/group
pulse-access:x:125:ruben,_snapclient,openhab,snapserver,pulse,root

And I have configured the following in /etc/pulse/{default|system}.pa:

load-module module-native-protocol-unix auth-group-enable=true auth-group=pulse-access socket=/tmp/pulse-socket

And in /etc/pulse/client.conf:

default-server = unix:/tmp/pulse-socket

The /tmp/pulse-socket is never created either. My question is as follows:

How can I sort out this mess, disable system mode and use Pulseaudio per user on a headless server in combination with snapcast?

If any information is missing, please let me know.

Answer 1

It turns out that it had nothing to do with a "normal" version of Pulseaudio running in system mode, but that my instance of Homeassistant Supervised was taking over the audio of the host system, see https://github.com/home-assistant/plugin-audio/issues/12 for information.

So now I'll revert back to Homeassistant core and hopefully the issue is solved.