Redis + MySQl or mongoDB?

Question

MySQL

I plan on what to make a site with music, Projects will resemble the audio list of VK, all on ajax without reloading.
A total of about 5,000 - 10,000 audio files, categories, sorted by popularity, name, artist, search. I count on 3000 users per month, given that they will constantly sort and search.
Naturally, there will be no file caching.
I have never worked with NoSQL, after reading a lot of articles I have not decided what to do better.
1. MySQL + Redis
2. MongoDB
Other options may be more logical, but NoSQL is unambiguous.

Answer 1

MySQL + Sphinx + Memcached
With a similar amount of data, you can keep a million visitors a month on this bundle.
Sphinx is needed for more or less decent full text search, usually a key point when searching for music.
MongoDB makes sense if you have at least 20 million songs, plus there will be a bunch of activity like plays, comments, likes and other nonsense.
MySQL doesn't scale well with a mixed load (when writes/reads are almost equally divided), but if you have only reads and the data is rarely updated, then you can have millions of compositions with a million daily traffic. There are solutions like Galera where scaling problems are solved quite well.
MongoDB should be used where schema flexibility is critical and significant scalability is important. This is a good solution, for example, for storing patient data and medical history. Each patient has a very complex and unique story. You can build a system of state-wide medical records on Monga. On MySQL it is also possible, but it will be a much more complicated solution.