Linux Filesystem Types: ext4 vs xfs vs btrfs

2026-06-06 Reading time: About 13 min Difficulty: Intermediate

What Is a Filesystem, and Which One Should You Choose?

Conclusion: When in doubt, pick ext4 (mature and stable). Choose xfs for large, highly parallel I/O, and btrfs when you want snapshots and checksums.

A filesystem maps the raw blocks on a disk into a structure of files and directories, and manages metadata such as permissions, timestamps, and sizes. On Linux, the same physical disk behaves differently depending on which filesystem you format it with — performance, operational features, and resize flexibility all change.

This guide compares the three filesystems most common on modern Linux — ext4, xfs, and btrfs — using practical decision criteria.

What Is ext4, and When Should You Use It?

Conclusion: ext4 is the most battle-tested general-purpose filesystem. Journaling and extents give it crash resilience and good throughput, making it a safe default.

ext4 (fourth extended filesystem) continues the ext2/ext3 lineage and has long been the default on Debian, Ubuntu, and others. Its biggest strength is maturity: it has by far the most production track record, and its failure modes and fixes are widely documented.

Key features:

• Journaling: Writes are recorded to a journal first, so after a power loss fsck can restore consistency quickly.
• Extents: Contiguous blocks are tracked as ranges (start + length), reducing fragmentation and handling large files efficiently.
• Online grow: You can grow it while mounted with resize2fs; shrinking requires unmounting.

# Format as ext4 (always double-check the target device)
$ sudo mkfs.ext4 /dev/sdX1

# Grow while mounted (after extending the LV or partition)
$ sudo resize2fs /dev/sdX1

What Is xfs, and When Should You Use It?

Conclusion: xfs is a 64-bit filesystem optimized for large volumes and high parallel I/O. The key operational caveat: it can grow online but cannot shrink.

xfs was originally developed by SGI for IRIX and is now the default filesystem on RHEL, CentOS Stream, Rocky, and AlmaLinux. It excels at large files, heavy metadata operations, and parallel I/O.

Key features:

• Allocation groups: The volume is split into independent regions whose metadata can be updated in parallel, boosting throughput.
• Built for scale: A 64-bit design that scales to exabyte-class volumes, well suited to file servers and large storage.
• Online grow only: xfs_growfs grows it while mounted, but there is no shrink feature.

# Format as xfs
$ sudo mkfs.xfs /dev/sdX1

# Grow online by mount point
$ sudo xfs_growfs /mnt/data

What Is btrfs, and When Should You Use It?

Conclusion: btrfs is feature-rich, with CoW, snapshots, checksums, and subvolumes. It is compelling for data protection and rollback, but operating it is more complex than ext4 or xfs.

btrfs (B-tree filesystem) is a next-generation filesystem built on Copy-on-Write (CoW), used as the default on openSUSE and the Fedora desktop. It pulls volume-management features into the filesystem itself rather than being a plain store.

Key features:

• Copy-on-Write (CoW): Instead of overwriting existing data, it writes to a new location. This pairs well with snapshots and makes old data less likely to be corrupted by a crash mid-write.
• Snapshots: Capture a point-in-time state instantly and roll back a failed change — handy as a safety net before updates.
• Checksums: Both data and metadata are checksummed, so silent bit rot can be detected.
• Subvolumes / built-in RAID: Partition one filesystem logically and span multiple devices.
• Online grow and shrink: btrfs filesystem resize changes size in either direction while mounted.

# Format as btrfs
$ sudo mkfs.btrfs /dev/sdX1

# Example snapshot (snapshot of / into /snap)
$ sudo btrfs subvolume snapshot / /snap/root-backup

How Do You Compare and Choose Among ext4 / xfs / btrfs?

Conclusion: The base axes are "stability → ext4", "large/parallel → xfs", and "snapshots/data protection → btrfs". Shrink support and feature count are the main trade-offs.

How Do You Check Your System's Filesystem?

Conclusion: df -T shows usage with the type, lsblk -f shows the FS per device, and blkid shows UUID and TYPE. Use them to confirm the device before formatting too.

# List mounted filesystems including the Type column
$ df -T

Filesystem     Type     1K-blocks      Used Available Use% Mounted on
/dev/sda2      ext4      48000000  20000000  25600000  44% /
/dev/sda1      vfat        523248      6132    517116   2% /boot/efi

# Show FS type, UUID, and mount point per block device
$ lsblk -f

# Show TYPE / UUID for a single device (avoid picking the wrong one before mkfs)
$ blkid /dev/sdX1

$ findmnt -no FSTYPE /
$ stat -f -c %T /