What is an inode? What is the difference between a hard link and a symbolic link?
Short answer: An inode is the on-disk structure that describes a file: its type, permissions, owner, size, timestamps, link count, and where its data blocks are. It does not contain the file name; names live in directory entries, which map a name to an inode number. A hard link is an additional directory entry pointing to the same inode. A symbolic (soft) link is a separate small file that contains a path to another file.
echo hello > a.txt
ln a.txt b.txt # hard link: same inode
ln -s a.txt c.txt # symbolic link: new inode containing the text "a.txt"
ls -li
# 1311 -rw-r--r-- 2 ... a.txt
# 1311 -rw-r--r-- 2 ... b.txt <- same inode number, link count 2
# 1312 lrwxrwxrwx 1 ... c.txt -> a.txt
| Hard link | Symbolic link | |
|---|---|---|
| What it is | another name for the same inode | a file containing a path |
| Inode | same as the target | its own |
| Survives deleting the original name | yes, data stays until the link count is 0 | no, becomes a dangling link |
| Across file systems | no | yes |
| To directories | no (would allow loops) | yes |
ls -l |
looks like a normal file | shows -> target |
How deleting works: rm calls unlink(), which removes one directory entry and decrements the inode's link count. The data is freed only when the link count is 0 and no process has the file open. That is why deleting a large log file that a program still has open does not free disk space until the program closes it (find these with lsof +L1).
"Everything is a file": in Linux, many objects are accessed through the file interface: regular files, directories, devices (/dev), pipes, sockets, and kernel information (/proc, /sys). They all have inodes in the VFS, and programs use open/read/write/close on them. This uniformity is why shell tools can work with devices and kernel data directly.
Inode limits: a file system has a fixed number of inodes (ext4 sets it at mkfs time). You can run out of inodes with free space still available if you create millions of tiny files: df -i shows usage. Dividing free space by free inodes gives the average space available per future file.