Techniques for building a distributed disk system without a single entry point?

Question

linux

As there are techniques for building systems without a single entry point, with replication and fault tolerance.
I know ceph is suitable for these purposes, and what alternatives are there, I really like the possibility of duplicating data and that the storage is distributed, that is, it’s not scary that one server or a couple of disks will fail, you can safely change them.
The goal is for all hypervisors to connect to such a scalable storage of many servers and disks, and virtual machines are already created there.

Answer 1

https://habrahabr.ru/company/oleg-bunin/blog/313364/
https://habrahabr.ru/company/oleg-bunin/blog/313330/

Answer 2

We used gluster, stored about 150tb in general, it worked not bad.