Computer Science › Operating Systems

File System

How files and directories are stored and organized.

Also known as: filesystem, ext4, NTFS

A file system is how an operating system stores, names and organizes data on a storage device. It turns a pile of numbered blocks on a disk into the files and folders you know.

/
├── home/
│   └── ana/
│       └── notes.txt
└── etc/
    └── hosts

What it does

• Names and paths: /home/ana/notes.txt identifies a file, in a tree of directories.
• Tracks where the data is: a file’s contents may be scattered in blocks, and the file system remembers which.
• Stores metadata: size, timestamps, owner and permissions.
• Allocates and frees space as files grow, shrink and are deleted.
• Keeps things consistent after a crash, often with a journal of pending changes.

Common file systems

Different file systems differ in maximum file size, case sensitivity, supported features and speed.

Things developers run into

• Case sensitivity: Linux treats File.txt and file.txt as different. Windows and macOS (by default) don’t, causing bugs that appear only after deployment.
• Path separators: / on Unix, \ on Windows. Use your language’s path functions.
• Running out of space or inodes: a disk can be “full” even with free bytes if it runs out of file entries (inodes).
• Atomicity: writing a file isn’t atomic. A crash mid-write leaves it half-written. Write to a temp file and rename.
• Slow with millions of files in one folder, in some file systems.
• Containers and volumes mount file systems into places (Docker volumes).
• Object storage (like S3) is not a file system, although it looks similar (object storage).

See Linux directory structure for where things live.