Introduce fileCluster Table Function for Shared Filesystems

Why it matters

This feature solves the challenge of accessing and querying data files that are shared across multiple ClickHouse servers using a shared filesystem such as NFS. It allows users to treat these distributed files as a single logical table cluster, facilitating parallel data processing and simplifying data management in environments with mounted shared storage.

How to use it

To use the fileCluster function, ensure that a shared filesystem (e.g., NFS) is mounted into the user_files directory on all relevant ClickHouse nodes. Then, invoke the function in your queries to access the files as a distributed cluster table.