What Is a File System?
What a File System Keeps Track Of
A file system is the layer of software that organizes data on a storage device into files and folders. It records where each file's data sits on the disk, who owns it, when it changed and which blocks, the fixed-size units a disk is divided into, are free for new data.
Each file has metadata, which is information about the file rather than its contents. That includes the name, size, permissions and timestamps. On Linux file systems such as ext4, this information is stored in a record called an inode. The file system keeps a table of these records together with a map of free space.
Journaling and Crash Recovery
Saving a file often requires several separate updates, such as the data itself, the file size and the free-space map. If the power fails halfway through, the disk can be left inconsistent. Most modern file systems reduce this risk with journaling.
A journal is a log in which the file system records planned changes before applying them. After a crash, the system replays or checks the journal rather than scanning the whole disk. This makes recovery faster and lowers the risk of damaged files. Journaling does not replace backups, but it greatly reduces corruption after an unclean shutdown.
Common File Systems on Linux Servers
ext4 is a general-purpose journaling file system and one of the most widely used on Linux. It is the default on many distributions. Its traits are compared with another common option in ext4 vs XFS.
XFS was designed for large files and parallel workloads. It is the default on many enterprise distributions and handles large volumes well.
ZFS combines a file system with volume management. It adds checksums, snapshots and pooled storage, and it is often used on dedicated storage servers, as described in what ZFS is.
Btrfs is a copy-on-write file system with snapshots and built-in multi-disk features. It is included in the Linux kernel and used on desktops and some servers.
These file systems share the same basic concepts, but they differ in speed, features and maturity, so the choice affects how a server behaves under load.
When a Full Disk Is Not Really Full
Consider a web server that stores a session file for every visitor. After a busy month, one directory holds millions of tiny files. The disk still shows gigabytes of free space, yet the application reports that it cannot create new files.
The cause is the inode table, which was sized when the file system was created and is now full. Each file needs an inode regardless of its size. Removing old session files frees inodes again, and the lesson for future setups is to plan for the number of files, not only the amount of data.
Frequently asked questions
What is the most common file system on Linux?
ext4 is one of the most widely used file systems on Linux. It is the default on many distributions and has a long record of stable operation. XFS is also common on enterprise systems, while ZFS and Btrfs are chosen when snapshots, checksums or built-in volume management matter.
Do Windows servers use the same file systems?
No. Windows uses NTFS and, in some cases, ReFS. Linux and BSD systems typically use ext4, XFS, ZFS or Btrfs. The underlying ideas, such as metadata and journaling, are shared, but the on-disk formats are not compatible, so a disk formatted for one family usually cannot be mounted by the other.