Btrfs In 2014: RAID5/RAID6, Data Deduplication

Samsung's F2FS isn't the only Linux file-system receiving new features but the Btrfs developers are currently baking some new treats of their own.

Btrfs developers were asked about the state of RAID5/6 in Btrfs, since the work has largely been silent in 2013 since the initial support was added to the kernel as still an experimental option. Chris Mason, the lead Btrfs developer and now employed by Facebook, responded.

Chris Mason says that the main item missing from Btrfs' RAID5/6 handling is the logging of parity to ensure consistent parity in situations like power loss. Chris says he's back to working on this code and is making good progress. He's also planning to fix scrub support for RAID5/6 Btrfs support after completing the parity work.

On a related note is also the new Linux kernel RAID library with Btrfs support for offering up to six levels of parity in RAID configurations.

As another exciting forthcoming item for Btrfs is online data de-duplication support. Data deduplication allows for eliminating duplicate blocks of repeating data. There's been other file-systems to offer inline data deduplication like SDFS and LessFS. There's also UKSM data deduplication as out-of-tree support for the Linux kernel.

There's now a set of fourteen patches that are up to their 8th revision on the Btrfs mailing list for offering data dedup support at the Btrfs file-system level. Separately is also a btrfs-progs patch for implementing de-duplication controls from the Btrfs program in user-space.

Another recent Btrfs news item worth reiterating is that Btrfs-progs is going to be updated in tandem with the Linux kernel.

With many changes lately and more forthcoming to Btrfs, plus now having the backing of Facebook, will 2014 finally be the year of the Btrfs file-system on Linux? OpenSUSE is likely switching to Btrfs in their next release but we haven't yet seen any signalling of other tier-one distributions making the switch in the near-term.