What is the best way to organize a file server in a virtual machine?

Question

Virtualization

I mean data storage itself. Now I have 2.5 TB of data in a physical file dump.
If I transfer the server to a virtual machine, is it better for me to create a VM with a disk of several TB (that is, it will eventually be one large vhdx file on the hypervisor, if I understand correctly) based on an iron raid and overtake the data there? Or somehow connect existing storages to a virtual machine that are located on physical hard drives?
Tell me best practice, so to speak, how is it done according to the mind?

Answer 1

In the end, I stuffed everything into virtual machines and bought a used server - a real one, on server hardware, put an iron raid there, a dozen on raid-edition screws.

Answer 2

I threw a physical Raid controller into the VM with a backup on 2 other physical disks connected to the motherboard controller, united by means of the OS in Raid 1. Of the minuses of the solution - any emergency shutdown of the VM / reset through the proxmox webmord launches an array rebuild, however I have it even with 3 disks normally rebuilt.
I would not make a file exchanger and a file dump on virtual disks, because it's an extra abstraction and a loss in performance.

Answer 3

Thanks, colleagues. From your answers, I realized that there is no universal correct solution, and there is already someone who is much better.
This is where conflicting opinions come in. I read on reddit that, on the contrary, it is better to shove everything into a container.
But I'll probably also prokinu disks in the VM better. For some reason, a multi-terabyte file does not inspire confidence in me.