NILFS

Design

"NILFS is a log-structured filesystem, in that the storage medium is treated like a circular buffer and new blocks are always written to the end.[…]Log-structured filesystems are often used for flash media since they will naturally perform wear-leveling;[…]NILFS emphasizes snapshots. The log-structured approach is a specific form of copy-on-write behavior, so it naturally lends itself to the creation of filesystem snapshots. The NILFS developers talk about the creation of "continuous snapshots" which can be used to recover from user-initiated filesystem problems[…]."

Using a copy-on-write technique known as "nothing in life is free", NILFS records all data in a continuous log-like format that is only appended to, never overwritten, an approach that is designed to reduce seek times, as well as minimize the kind of data loss that occurs after a crash with conventional file systems. For example, data loss occurs on ext3 file systems when the system crashes during a write operation. When the system reboots, the journal notes that the write did not complete, and any partial data writes are lost.

Some file systems, like UFS-derived file systems used by the Solaris operating system and BSDs, provide a snapshot feature that prevents such data loss, but the snapshot configuration can be lengthy on large file systems. NILFS, in contrast, can "continuously and automatically [save] instantaneous states of the file system without interrupting service", according to NTT Labs.

The "instantaneous states" that NILFS continuously saves can actually be mounted, read-only, at the same time that the actual file system is mounted read-write — a capability useful for data recovery after hardware failures and other system crashes. The "lscp" (list checkpoint) command of an interactive NILFS "inspect" utility is first used to find the checkpoint's address, in this case "2048":

# inspect /dev/sda2...nilfs> listcp 1 6 Tue Jul 12 14:55:57 2005 MajorCP|LogiBegin|LogiEnd2048 2352 Tue Jul 12 14:55:58 2005 MajorCP|LogiEnd...nilfs> quit

The checkpoint address is then used to mount the checkpoint:

# mount -t nilfs -r -o cp=2048 /dev/sda2 /nilfs-cp# dfFilesystem 1K-blocks Used Available Use% Mounted on/dev/sda2 70332412 8044540 62283776 12% /nilfs/dev/sda2 70332412 8044540 62283776 12% /nilfs-cp

Features

NILFS supports continuous snapshotting. In addition to versioning capability of the entire file system, users can even restore files mistakenly overwritten or destroyed just a few seconds ago. Since NILFS can keep consistency like conventional LFS, it achieves quick recovery after system crashes.

NILFS creates a number of checkpoints every few seconds or per synchronous write basis (unless there is no change). Users can select significant versions among continuously created checkpoints, and can change them into snapshots which will be preserved until they are changed back to checkpoints.

There is no limit on the number of snapshots until the volume gets full. Each snapshot is mountable as a read-only file system. It is mountable concurrently with a writable mount and other snapshots, and this feature is convenient to make consistent backups during use.

Snapshot administration is easy and quickly performable. NILFS will make snapshotting or versioning of the POSIX filesystem much familiar to you. The possible use of NILFS includes, versioning, tamper detection, SOX compliance logging, and so forth. It can serve as an alternative filesystem for Linux desktop environment, or as a basis of advanced storage appliances.

The current major version of NILFS is version 2, which is referred to as NILFS2. NILFS2 realized online garbage collection that reclaims disk space with keeping multiple snapshots.

Other NILFS features include:

B-tree based file and inode management.

Immediate recovery after system crash.

64-bit data structures; support many files, large files and disks.

64-bit on-disk timestamps which are free of the year 2038 problem.

Loadable kernel module; no recompilation of the kernel is required.

Supported Features

Basic POSIX file system features

Snapshots

Automatically and continuously taken, you can retrieve past state of NILFS file system.

No limit on the number of snapshots until the volume gets full

Mountable as read-only file systems

Mountable concurrently with the writable mount (convenient to make consistent backups during use)

Quick listing

Easy administration

Background Garbage Collection

Can maintain multiple snapshots

Selectable GC Policy, which is given by a userland daemon.

Quick crash recovery on-mount

Read-ahead for meta data files as well as data files

Block sizes smaller than page size (e.g. 1KB or 2KB)

Redundant super block

Online resizing (since Linux-3.x and nilfs-utils 2.1)

Related utilities (by contribution of Jiro SEKIBA)

grub2

util-linux (blkid, libblkid, uuid mount)

udisks, palimpsest

Filesystem label (nilfs-tune)

Additional features

Fast write and recovery times

Minimal damage to file data and system consistency on hardware failure

32-bit checksums (CRC32) on data and metadata for integrity assurance (per block group, in segment summary)

Correctly ordered data and meta-data writes

Redundant superblock

Internal data is processed in 64-bit wide word size

Can create and store huge files (8 EiB)

OS Compatibility

NILFS was merged into the Linux kernel 2.6.30.

On distributions where NILFS is available out-of-the-box, the user needs to download the nilfs-utils package, following the instructions from [2].

A separate, BSD licensed implementation, currently with read-only support, is included in NetBSD.

Relative Performance

In the January 2015 presentation SD cards and filesystems for embedded systems at Linux.conf.au, it was stated:

License

The NILFS2 filesystem utilities are made available under the GNU Public License version 2, with the exception of the lib/nilfs libraries and their header files, which are made available under the GNU Lesser General Public License Version 2.1.

Developers

The Japanese primary authors and major contributors to the nilfs-utils who worked or are working at labs of NTT Corporation are:

Ryusuke Konishi (Primary maintainer, 02/2008-Present)

Koji Sato

Naruhiko Kamimura

Seiji Kihara

Yoshiji Amagai

Hisashi Hifumi and

Satoshi Moriai.

Other major contributors are:

Andreas Rohner

Dan McGee

David Arendt

David Smid

dexen deVries

Dmitry Smirnov

Eric Sandeen

Jiro SEKIBA

Matteo Frigo

Hitoshi Mitake

Takashi Iwai

Vyacheslav Dubeyko