btrfs
The btrfs (B -tree FS, alternatively " Butter FS" or "Better FS" ) is a copy-on - write file system (GPL ) for the Oracle Corporation since 2007 as free software under the GNU General Public License the Linux operating system is developed. Tests since the year 2011 has resulted a steady state (as of April 2013).
Maybe it is to replace the hitherto prevailing in the Linux environment ext4 file system in the future because this has lifted only part of the limitations of ext3 as file size and total file system size. Why is Andrew Morton, one of the most prominent Linux kernel developers, in the longer term on btrfs.
Properties
Btrfs has many similarities with ZFS. It will mainly provide other functions that make it stand out from the current standard Linux ext3/ext4, but also of other file systems such as XFS or JFS, this includes:
- Extended storage area (264 bytes)
- Efficient storage of small files and folders
- Dynamic inodes
- Snapshots
- Several root directories ( subvolume )
- Integrated RAID, mirroring and concatenation of objects (RAID 0, RAID 1, RAID 10; experimental support for RAID 5 and 6)
- Checksums
- Data compression
- File system check and defrag during operation
- Efficient internal incremental backup
- Conversion of existing ext3-/ext4-Dateisysteme
- Copy- on-write
The built- in file system RAID subsystem provides over traditional hardware or software RAID implementations the advantage that one can distinguish between used and unused blocks of data and thus only used disk space must be mirrored in the reconstruction of a mirrored RAID volume. This results in case of damage, especially for poorly filled file systems, a tremendous time saver.
History
Btrfs can be described as a Linux analog to ZFS. ZFS was indeed already years earlier from now the same manufacturer (Sun Microsystems, merged into Oracle) designed as the definitive file system, because of his license status was, however, unsuitable for use with Linux. Furthermore btrfs builds with the B -tree structure on a central concept of XFS. It is thus also related to Reiser4, to which it is seen as an adequate guard.
The core structure of Btrfs - the copy-on -write B -tree structure - was originally proposed by the IBM researcher Ohad Rodeh during a presentation at the USENIX 2007. Rodeh also suggested adding Referenzierungszählern for memory blocks and certain relaxations of the balancing algorithms of normal B - trees that make the B- trees for high-performance storage solutions with copy- on-write snapshots fit, while maintaining a good concurrency.
Chris Mason, then a ReiserFS developers at SUSE, was set in the same year by Oracle and began his work on a new file system that is used almost exclusively those B- trees - not only for meta- and Nutzdateien, but also recursively to persecution of the memory allocation of the trees themselves. This means that all operations can be handled by the same routines.
January 9, 2009 btrfs was first included in the Linux kernel 2.6.29.
The support of BTRFS is already included in the main version of the Linux kernel (since 2.6.29 -rc1 ). In some Linux distributions, the file system is already officially in the installation to choose from.
Support in the Linux distributions
For the experimental operation, the file system is first supported in OpenSUSE 11.3 as well. Using Oracle Linux Release 2 and there are packages for the distributions Gentoo, Arch Linux, Debian and Ubuntu
Btrfs is the standard file system of MeeGo and should be used in Fedora as the default file system for many releases. However, this was again discarded, and finally moved to version 20. [ Deprecated] Thus, Btrfs is so far only in the developer version of the distribution.
Criticism of the design
Red Hat commissioned in the second quarter of 2010, Edward Shishkin, one of the original Reiser4 developers with a code review. Shishkin's conclusion was that the design is flawed because the original algorithm in key points are not followed. The design error cause in special cases, the disk space can run out, even though there is enough space.