What Are Inodes?
Inode is short for index node, the filesystem metadata that stores information about each file and where it is. Every file and directory has an inode that contains ownership details, permissions, the file’s size, the filetype, and pointers to its data. The contents and structure differ by filesystem, but inodes play a similar role in all Unix filesystems.
Because every file has an inode, enforcing an inode quota is the same as enforcing a limit on the number of files. Inode quotas don’t control the amount of data a user can store in their home directory; they put an upper limit on the number of files, not the size.
In the past, inode quotas were essential because each filesystem had a fixed quantity determined when the disk volume was formatted. If users generated too many files, they could use all the inodes, making it impossible to create new files. This was a common occurrence on email servers with inboxes that contain many small files. However, beginning with CentOS 7’s release in 2014, the default filesystem is XFS, which allocates inodes dynamically—you can’t run out of inodes.
However, large numbers of files create other problems. They can degrade hard drive performance and increase memory consumption. Additionally, each inode consumes a small amount of disk space, by default 512 bytes on XFS. In rare circumstances, a user could generate so many files they fill the volume with inodes, leaving no space for data. This usually only happens when buggy scripts autogenerate billions of tiny or empty files, but it’s a risk that should be guarded against.
https://blog.cpanel.com/how-to-manage-i ... ode-usage/