Each inode stores the attributes and disk block locations of the object’s data. POSIX per-file system unique identifier for a file (each drive being an independant filesystem) - wikipedia / File System Continued

# inode ⮺

• inode number
• inode block group - inode are store within a group, indexed by inode number → block group + inode index within that group
• inode generation - allows to reuse deleted inode and detect that they don’t reference the same file as previously
• inode bitmap - indicates which entries are free/in use inside a block group

Filesystem
│
├── Block group 0loc
│   ├── inode bitmap
│   └── inode table
│       ├── inode #...
│       ├── inode #...
│       └── ...
│
├── Block group 1
│   ├── inode bitmap
│   └── inode table
│
└── ...

# Hardlink

Allows to share content + metadata except the name which may differ. hardlink cannot link to directory => This would cause a loop in the filesystem and that’s why it’s prohibited by the filesystem

• fastest way to find all the file with the same inode?

$ find . -inum 1234

# Symlink ⮺

In Linux, a symlink (symbolic link) itself is a separate file, and it has its own metadata, independent of the target file.

Metadata Stored on a Symlink (Independent of Target)

• File Type: symlink
• Ownership (User and Group): A symlink has its own owner (UID) and group (GID).
• Timestamps
• Size: The size of a symlink is the length of the path it stores

What Metadata is NOT Stored in a Symlink?

• File content
• target file permission
• target file size