Linux kernel 2.6.29 gets two new file systems
Summary
In January 2009 the file systems Btrfs and Squashfs were accepted into the development branch for Linux 2.6.29. Squashfs compresses files, directories and inodes and is used mainly for live CDs and embedded systems, but offers no write access. Oracle’s Btrfs was regarded as a possible future standard but was not yet production-ready, which is why Ext4 was to follow Ext3 for the time being.
Ideas
- A read-only, compressed file system saves space on live and embedded media.
- Ext3 was reaching its limits; Btrfs was to follow in the long term.
- Ext4 served as an interim solution until Btrfs matured.
- Btrfs brought snapshots, checksums, compression, online checking and subvolumes.
Insights
- New file systems need years from entering the kernel to production readiness.
- A proven interim solution gives new approaches time to mature.
Facts
- Btrfs was developed at Oracle.
References
Critique
- The report names no concrete limits of Ext3 that justify the change.
Remarks
- Years later Btrfs became the default at openSUSE and Fedora; Squashfs is now in almost every live system and in Snap packages.
Recommendations
- Use Btrfs snapshots before system updates so you can roll back quickly.
- Use Squashfs for immutable images and embedded systems.
Links to the original source and the Web Archive open in a new tab.