Please decide whetehr better to move this into the docker container part of this forum.

There used to be a docker container in CA with a Muscbrainz mirror. Unfortunately, this has been deprecated for some years.

I believe many of us have started to look into self-hosting again since the issues with Lidarr metadata. Even once Lidarr is fixed, there may stay be interested using a local mirror for MBZ Picard.

There is a detailed description below how to set this up with docker-compose. Amy chance we can see a container on CA again? Or if note, anyone can help how to set up below with docker-compose in Unraid? Thanks!

https://github.com/blampe/hearring-aid/blob/main/docs/self-hosted-mirror-setup.md

Yea I'd love this too. The guide is just a bit confusing lol

You can follow the docker-compose instructions as long as you install the "Docker Compose Manager" plugin (which will let you use docker compose). Be sure to change the paths in volume-settings.yml and lmd-settings.yml. There's a few containers so this would be non-trivial to set up in Unraid without using Docker Compose.

I've gone down a long rabbit hole with Lidarr to finally getting it to a point where it's no longer useless, but somewhat useful. It starts with blampe's guide linked above. And once it's all set up and working, it continues with some custom scripts I've made to easily manage adding artists to Lidarr without any manual lookups or scans - including batch processing for an unlimited number of artists at one time. I'll publish that code soonish.

The easiest way to deploy this on Unraid, IMO, is to use an LXC container. So first install the LXC plugin.

Then create a Debian LXC and start it up. Open a web terminal into that LXC and follow exactly the directions in blames guide linked above. Don't change anything, including changing paths - no need to make any edits.

Just ONE thing. Before following his instructions, install nano:

apt install nano

Then just blindly follow the directions.* Every time the directions tell you to "create a file" just copy the path and filename from the instructions and at the command prompt:

$> nano <<PASTE>>

That will create the file, then copy and paste the contents from the instructions into the file and CRTL-X, ENTER, Y

• * Some of the instruction blocks contain multiple lines that may not be interpreted correctly if you paste them all at once - if you want to be safe, paste them one line at a time into the terminal to make sure the commands are doing what they're supposed to do.

That's about it. Make sure your machine doesn't go to sleep while this is being done and don't close the terminal window, unless you're using SCREEN to run the process - some of the steps will take a while, such as when pulling compressed data from musicbrainz.

After it's all done and you can confirm that it's all working, go to the Tubifarry GitHub and optionally add some more data sources, like Deezer, Discogs, etc.

It'd be nice to have this all packaged into a docker in the App Store to make it all much easier for the majority of people, but for those who are willing to take some time, the above works solidly. IMO, this is the only way to get a working Lidarr. I don't think the developers are ever going to get it working reliably with Musicbrainz on their own. They've spent most of the year on it (supposedly) and it's still horribly broken.

Edited January 2Jan 2 by Espressomatic

In case anyone was curious, once you beat this horse into submission..... You can start getting close to completion for most of what you're tracking - slskd can get you most of the way with Tubifarry. I needed to create a number of custom scripts to help prevent Lidarr tossing a lot of downloads into the bin. Even still, expect to download up to many times the MB/TB that you want to keep.

And in case anyone is worried about collection size, I'm at about 150k tracks - though you really should manually convert Lidarr to PostgreSQL if you're hitting anywhere close to 100k).

This is for FLAC only. If you're collecting without regard to format, completion is much easier and MUCH MUCH faster.

Edited January 19Jan 19 by Espressomatic