Remove duplicate mp3 with different name, size, and hash

Use MusicBrainz's Picard. It scans the music to generate a 'fingerprint' and then compares it against an online database to identify it (pulling all the info like Artist, Album, etc).

After it's identified the songs, it can organize them into a custom directory & name structure, and you can use that to find the duplicates.

Fully compatible with linux.

There's an open source acoustic fingerprinting system called Echoprint, you might want to check I want to deduplicate a big collection. I haven't used it before (just stumbled across it), but it sounds awesome to be able to do this locally.

I'd recommend using beets with its duplicates plugin; beets is similar to picard in terms of functionality (it can lookup ID3 info from the musicbrainz database and organize your files based on that info), but goes beyond picard's abilities with some of its plugins.

Personally I use them together--picard for cleaning up ID3 tags and beets for organizing the files and cleaning up duplicates.