Introduce the random filesystem type

[lucapette]

What

In #96 we introduced mount and unmount commands for the file system. They rely on a /mnt/random file system type but the work is unfinished:

• you can mount and unmount random endpoints
• you can't make use of them (cat /mnt/random/int fails because the catalog knows nothing about it)
• you can't specify anything else than int as a type. It'd be nice to be able to generate other data types but it's outside the scope of this issue. We'll create a milestone with the goal of extending the random filesystem.

Motivation

The random filesystem type is less challenging than the upcoming http and jdbc ones. At the same time, to make this work we'll have to solve a variety of basic problems (how does the data catalog knows about newly mounted endpoints, how does it compile the node down to something kafka streams can make use of) that will simplify the upcoming ones.