ZFS v28 for FreeBSD started testing

Pawel Jakub Dawidek completed the work on porting the latest version of the ZFS file system to FreeBSD and encouraged enthusiasts to take part in testing the code before integrating it into the FreeBSD HEAD code base. Separately, it is reported that despite the stagnation of the OpenSolaris project, work on ZFS in FreeBSD will continue on its own, a good team of developers, some of whom have previously participated in the development of OpenSolaris, have already been selected and several companies have expressed their interest in the project. Much attention will also be paid to collaboration with the IllumOS project, within which a completely free branch of the OpenSolaris code base developed by the independent developer community has been created.

Of the features of ZFS v28 can be noted:

• Support automatic recognition and merging of duplicate data. If several files contain similar data blocks, they will be saved to physical media only once, which will significantly reduce disk space consumption and increase performance — instead of copying blocks, the entry in the corresponding table will be changed.
  
• RAIDZ3 support, a RAIDZ variant with storage of three copies of structures responsible for ensuring the integrity, which can significantly increase the reliability of storage compared to dual-duplication RAID modes - RAID-6 and RAIDZ2. For example, RAIDZ3 ensures data integrity while simultaneously leaving three disks at a time while standing up and allows to increase storage reliability during array rebuilding after disk replacement;
  
• Support for the zfs diff utility, which allows you to view a list of changes between two ZFS snapshots or between snapshots and the current state of the file system. The utility displays the facts of changing, renaming, creating and deleting files and directories;
  
• Support for the zpool split command, designed to split a mirrored zpool partition into several separate pools. This feature allows, for example, to simplify data cloning, when several disks are added to the mirror, synchronization is performed and disks are excluded from the pool, after which the disks contain a full copy of the data and can be used to load similar servers. Using zpool split, you can easily exclude a disk from the pool and then create a new pool based on it;
  
• Supports reference counting for snapshots, for more flexible management of deleting unused snapshots. For example, by increasing the counter, the user may mark that the snapshot is in use and cannot be deleted;
  
• Support for the “zpool import -F” command, which allows you to rewind the damaged pool to a state corresponding to an earlier group of transactions;
  
• The ability to import a pool in read-only mode;
  
• Performance optimization and error handling.