Permissions issue over SSH and ESP home wont connect

[soulgun16]

Hello, I am unsure if this is the intended way that permissions work, but the picompose image builds the containers and most files with root and the default admin user pi does not have permission over SSH to modify anything which makes file modification a bit tedious as I either have to use sudo nano to edit the file on the pi connected to a monitor or I have to create a password for the root user and then modify the SSH file to allow root over SSH. I am unsure if there is something i am missing or if this is intended? Also when trying to connect to home assistant via ESP home it tells me "unable to connect, make sure your configuration includes an API section" so just to be safe i modified the .env for LVA to remove the comment and include port 6053 manually even though I know that should be used by default, but no change. I have read over both the picompose instructions and LVA instructions extensively and I dont see either of them mention any additional dependencies that may be needed, so i am unsure if there is another service i need to be running for this to work or if it is just some sort of configuration issue on the pi side. Any help with this issue would be most appreciated, I was previously running wyoming satellite but I know this is the new deployment method and I am excited to have an easy snapcast addition and the power of expandable dockers for features, I just seem to be running into some bumps following along and getting it all set up the first time. Thank you!

[florian-asche]

Hey, can you specify what files ?

[soulgun16]

Well it seems that when picompose sets everything up it assigns root as the owner of all of the files in the compose folder, and also var/lib/docker/volumes where the wake word files would be located. Really the entire directory is almost exclusively root owned by default and my work around for now has been enabling root user over ssh to manage the files. That being said, I still haven't gotten LVA to run yet, I've been trying to do more of my own troubleshooting in the meantime and doing a docker ps shows the LVA docker as boot looping so clearly something is going wrong there before I even get a chance to worry about the port. I previously tried using the respeaker release so I could have LVA and snapcast pre configured in hopes of a smooth install despite my speaker not being the respeaker, I just have a generic Anker one but I manually grabbed the device ID and added it to the .env file. That had the same permission issues over ssh and HA said the port was not open, so I tried manually adding the same default port in the .env by uncommenting the port but had the same effect. So I took a copy of those docker folders and reformated with the current stable base image, enabled the root ssh again and pasted in the containers to reach my current state. I can elaborate or paste screenshots/configs as needed and am much appreciative of your help :)

[soulgun16]

Hello, I wanted to follow up on this, so to start on the permissions issue, simply put, i think the permissions are clearly intended as is since almost all files are owned and created by root user, and essentially I suppose I would like to suggest you consider adding Read/Write permissions to the GUID 1000 that the Pi user is in, that would allow modification of files by the pi user over SSH without requiring you to change the proper ownership from being root since you could keep UID 0 while its creating the files. That would be very helpful for management of the docker env's, YAMLs, and whole container additions, etc as right now you cannot add, delete, or modify over SSH. If you really wanted to you could even limit this modification to just the compose folder so that way users can manage containers/apps in the compose folder over SSH but not be able to modify the other essential files that they probably shouldnt be in anyway.

A follow up on the LVA deployment, and i apologize that i have lumped 2 issues in one here but I once again tried a fresh image and this time i used the development builds instead of stable (which i think was 4/22) and chose the 5/22 build + LVA (i think thats the date) and upon deployment it seemed to not use the user I had pre-defined in the pi imager (I am using Pi 3B) and instead actually had the pi user set with the default raspberry password as mentioned in the guide, this time the LVA actually started the container properly and after some time HA was able to connect successfully with ESP! However the part where you configure the voice assistant says that the device cant connect to HA which seems to contradict itself and I know its on the same subnet since they are both in my local range and are clearly talking, and logs on the LVA container look like they are saying micro wake word didnt start properly? I will attach those logs to this post so you can look, however I think its also trying to use micro wake word and I am currently trying to look into all of the documentation to see how I can swap it to open wake word since I would like to swap to that. Currently I dont see any mention of micro vs open wake word in the .env or YAML so I am still trying to figure that out but I am done for the night. I enabled the sound when activated feature and can can confirm "ok Nabu" sets off the sound, however after that it never responds again and the logs arent mentioning a failure or anything. If you have any ideas on why its not triggering HA to perform actions (it never changes assist mode from idle) any help would be greatly appreciated. Thank you!

LVA log.txt