Hi,
If I want to use Mycroft on a Raspberry Pi 4, which I want to use as a desktop computer at the same time, what is the best way to do this?
Install Mycroft on Raspberry? Or add the desktop to Picroft?
(I would use the Google AIY Voice HAT)
I’m not familiar with “Docker” maybe that’s the way?
Git installation mycroft-core → build mycroft-gui on top → desktop-link(s) performing mycroft-start all , mycroft-stop , mycroft-gui-app and mycroft-cli-client (in terminal) and you’re good.
Using docker is perfectly fine, yet i would preferably use it in performance critical applications. (I run a pi4 server with serveral pods managing the household data) And if you plan on orchastrating several docker -with cross communication-, it’s not a straight forward task at first.
Although i’m pretty curious about mycroft driven by rootless podman, but i would guess this needs additional dev time on podman to not end up in too much of a hassle.
Thanks for your feedback! And thanks for bringing that ansible-thing to notice, I didn’t know it. I’m not sure which RPi 4 I have though, will look it up later.
Thanks for your feedback too. Now I have 1 vote for and 1 vote against :-).
I don’t plan to use several docker-applications or docker on different RPi’s.
And you seem a lot more savvy then me, I don’t know what a podman is, dev time and all that kind of things :-).
Podman is on my to do, last time I tried the arm64 support was not great. I tried docker-rootless and felt a performance issue mostly related to the filesystem.
@Taxicletter the best way to know will be to try different things! Enjoy
It describes installing Ubuntu server, but Raspberry Pi OS shouldn’t be much different.
It’s still very “alpha”, but if you use it, any feedback would be appreciated. Maybe I could add a section on installing Raspberry Pi OS as an alternative.
I think everything went OK, but now I’m stuck. I’m using these instructions: Docker - Mycroft AI,
and I’m stuck on this part: “Running Mycroft for Docker”.
I don’t know where I should do this: " Just replace the directory_on_local_machine with the path you want the container mapped to on your local machine" and they don’t explain how to do this “Set PULSE_SERVER env variable” and “Share pulseaudio’s cookie”.
I tried copying the code underneath, but somehow it doesn’t copy and copying line per line gives an error in the first line, I need to set an argument…
I’ll try further, but meanwhile, if you have hints or tips
-v directory_on_local_machine:/root/.mycroft
what reads create a volume (-v) mapping directory_on_local_machine to the one inside the container (/root/.mycroft)
the ones attributed with -e are environmental variable. ie -e PULSE_SERVER=unix:${XDG_RUNTIME_DIR}/pulse/native is the one talked aout
this goes along with -v ${XDG_RUNTIME_DIR}/pulse/native:${XDG_RUNTIME_DIR}/pulse/native
you have to check if there is something at ${XDG_RUNTIME_DIR}/pulse/native, resp. where it is to map it in.
The same would apply to the cookie.
That looks more complicated than I understand… Thought it would be an easier process. I really don’t understand “what reads create a volume (-v) mapping directory_on_local_machine to the one inside the container” Creating a volume?? Mapping??
I don’t know what smartgics or microservice are either, I didn’t think I was dockering anything up there.
I understand basic terminal commando’s and I can follow a manual, but since this manual seems to stop halfway through, I’m stuck as well :-).
The info on Network Setup – PulseAudio is way to complicated for me. They offer choices and I have to decide on info I don’t understand.
Is the way to install Mycroft WITHOUT docker as complicated as this?
(Thanks for your feedback, by the way, I forgot to mention that )
forget about docker and install it using the usual method
That’s why i suggested that in the first place. You would have to dig through docker documentary otherwise to understand the syntax to be able to apply the specifics.