Computer Science › Operating Systems
Inode
The file system record that holds a file's metadata.
Also known as: inode, index node, inode number
An inode (index node) is the file system’s record for a file. It holds the file’s metadata — its type, size, permissions, owner, timestamps, link count, and pointers to where its data lives on disk. Crucially, it does not hold the file’s name. Names live in directories, which map a name to an inode number.
directory entry: "report.txt" → inode 4021
inode 4021: type=file, size=2KB, perms=644, links=1, blocks→[...]This separation is why hard links work: two names can point to the same inode, so “the same file” has two paths. It’s also why renaming a file within a file system is instant — you change a directory entry, not the data. And it’s why moving a file between file systems copies it — the inode is local to one file system.
The classic mistakes:
- Running out of inodes, not space. A file system reserves a fixed number of inodes at creation. Millions of tiny files can exhaust inodes while gigabytes of space remain — a surprising “disk full” that
df -i(notdf -h) reveals. - Confusing inodes with filenames. The name is a pointer. Deleting a name decrements the link count; the data is freed only when the count reaches zero.
- Expecting metadata changes to touch data.
chmod,chownandtouchchange the inode, not the file contents. (This is why they’re fast, and why their timestamps differ.) - Assuming inode numbers are meaningful across systems. They’re unique only within one file system.
- Forgetting symlinks have their own inode. A symbolic link is a small separate file whose content is a path; its target’s inode is elsewhere.
Inodes are a core piece of how Unix file systems organize storage: metadata separate from names, data referenced by block pointers. They explain disk usage inode limits, hard links, and why some file operations are instant while others copy gigabytes.