Documentation/filesystems/nilfs2.txt

   1 NILFS2
   2 ------
   3
   4 NILFS2 is a log-structured file system (LFS) supporting continuous
   5 snapshotting.  In addition to versioning capability of the entire file
   6 system, users can even restore files mistakenly overwritten or
   7 destroyed just a few seconds ago.  Since NILFS2 can keep consistency
   8 like conventional LFS, it achieves quick recovery after system
   9 crashes.
  10
  11 NILFS2 creates a number of checkpoints every few seconds or per
  12 synchronous write basis (unless there is no change).  Users can select
  13 significant versions among continuously created checkpoints, and can
  14 change them into snapshots which will be preserved until they are
  15 changed back to checkpoints.
  16
  17 There is no limit on the number of snapshots until the volume gets
  18 full.  Each snapshot is mountable as a read-only file system
  19 concurrently with its writable mount, and this feature is convenient
  20 for online backup.
  21
  22 The userland tools are included in nilfs-utils package, which is
  23 available from the following download page.  At least "mkfs.nilfs2",
  24 "mount.nilfs2", "umount.nilfs2", and "nilfs_cleanerd" (so called
  25 cleaner or garbage collector) are required.  Details on the tools are
  26 described in the man pages included in the package.
  27
  28 Project web page:    http://www.nilfs.org/en/
  29 Download page:       http://www.nilfs.org/en/download.html
  30 Git tree web page:   http://www.nilfs.org/git/
  31 List info:           http://vger.kernel.org/vger-lists.html#linux-nilfs
  32
  33 Caveats
  34 =======
  35
  36 Features which NILFS2 does not support yet:
  37
  38         - atime
  39         - extended attributes
  40         - POSIX ACLs
  41         - quotas
  42         - fsck
  43         - defragmentation
  44
  45 Mount options
  46 =============
  47
  48 NILFS2 supports the following mount options:
  49 (*) == default
  50
  51 barrier(*)              This enables/disables the use of write barriers.  This
  52 nobarrier               requires an IO stack which can support barriers, and
  53                         if nilfs gets an error on a barrier write, it will
  54                         disable again with a warning.
  55 errors=continue         Keep going on a filesystem error.
  56 errors=remount-ro(*)    Remount the filesystem read-only on an error.
  57 errors=panic            Panic and halt the machine if an error occurs.
  58 cp=n                    Specify the checkpoint-number of the snapshot to be
  59                         mounted.  Checkpoints and snapshots are listed by lscp
  60                         user command.  Only the checkpoints marked as snapshot
  61                         are mountable with this option.  Snapshot is read-only,
  62                         so a read-only mount option must be specified together.
  63 order=relaxed(*)        Apply relaxed order semantics that allows modified data
  64                         blocks to be written to disk without making a
  65                         checkpoint if no metadata update is going.  This mode
  66                         is equivalent to the ordered data mode of the ext3
  67                         filesystem except for the updates on data blocks still
  68                         conserve atomicity.  This will improve synchronous
  69                         write performance for overwriting.
  70 order=strict            Apply strict in-order semantics that preserves sequence
  71                         of all file operations including overwriting of data
  72                         blocks.  That means, it is guaranteed that no
  73                         overtaking of events occurs in the recovered file
  74                         system after a crash.
  75 norecovery              Disable recovery of the filesystem on mount.
  76                         This disables every write access on the device for
  77                         read-only mounts or snapshots.  This option will fail
  78                         for r/w mounts on an unclean volume.
  79 discard                 This enables/disables the use of discard/TRIM commands.
  80 nodiscard(*)            The discard/TRIM commands are sent to the underlying
  81                         block device when blocks are freed.  This is useful
  82                         for SSD devices and sparse/thinly-provisioned LUNs.
  83
  84 NILFS2 usage
  85 ============
  86
  87 To use nilfs2 as a local file system, simply:
  88
  89  # mkfs -t nilfs2 /dev/block_device
  90  # mount -t nilfs2 /dev/block_device /dir
  91
  92 This will also invoke the cleaner through the mount helper program
  93 (mount.nilfs2).
  94
  95 Checkpoints and snapshots are managed by the following commands.
  96 Their manpages are included in the nilfs-utils package above.
  97
  98   lscp     list checkpoints or snapshots.
  99   mkcp     make a checkpoint or a snapshot.
 100   chcp     change an existing checkpoint to a snapshot or vice versa.
 101   rmcp     invalidate specified checkpoint(s).
 102
 103 To mount a snapshot,
 104
 105  # mount -t nilfs2 -r -o cp=<cno> /dev/block_device /snap_dir
 106
 107 where <cno> is the checkpoint number of the snapshot.
 108
 109 To unmount the NILFS2 mount point or snapshot, simply:
 110
 111  # umount /dir
 112
 113 Then, the cleaner daemon is automatically shut down by the umount
 114 helper program (umount.nilfs2).
 115
 116 Disk format
 117 ===========
 118
 119 A nilfs2 volume is equally divided into a number of segments except
 120 for the super block (SB) and segment #0.  A segment is the container
 121 of logs.  Each log is composed of summary information blocks, payload
 122 blocks, and an optional super root block (SR):
 123
 124    ______________________________________________________
 125   | |SB| | Segment | Segment | Segment | ... | Segment | |
 126   |_|__|_|____0____|____1____|____2____|_____|____N____|_|
 127   0 +1K +4K       +8M       +16M      +24M  +(8MB x N)
 128        .             .            (Typical offsets for 4KB-block)
 129     .                  .
 130   .______________________.
 131   | log | log |... | log |
 132   |__1__|__2__|____|__m__|
 133         .       .
 134       .               .
 135     .                       .
 136   .______________________________.
 137   | Summary | Payload blocks  |SR|
 138   |_blocks__|_________________|__|
 139
 140 The payload blocks are organized per file, and each file consists of
 141 data blocks and B-tree node blocks:
 142
 143     |<---       File-A        --->|<---       File-B        --->|
 144    _______________________________________________________________
 145     | Data blocks | B-tree blocks | Data blocks | B-tree blocks | ...
 146    _|_____________|_______________|_____________|_______________|_
 147
 148
 149 Since only the modified blocks are written in the log, it may have
 150 files without data blocks or B-tree node blocks.
 151
 152 The organization of the blocks is recorded in the summary information
 153 blocks, which contains a header structure (nilfs_segment_summary), per
 154 file structures (nilfs_finfo), and per block structures (nilfs_binfo):
 155
 156   _________________________________________________________________________
 157  | Summary | finfo | binfo | ... | binfo | finfo | binfo | ... | binfo |...
 158  |_blocks__|___A___|_(A,1)_|_____|(A,Na)_|___B___|_(B,1)_|_____|(B,Nb)_|___
 159
 160
 161 The logs include regular files, directory files, symbolic link files
 162 and several meta data files.  The mata data files are the files used
 163 to maintain file system meta data.  The current version of NILFS2 uses
 164 the following meta data files:
 165
 166  1) Inode file (ifile)             -- Stores on-disk inodes
 167  2) Checkpoint file (cpfile)       -- Stores checkpoints
 168  3) Segment usage file (sufile)    -- Stores allocation state of segments
 169  4) Data address translation file  -- Maps virtual block numbers to usual
 170     (DAT)                             block numbers.  This file serves to
 171                                       make on-disk blocks relocatable.
 172
 173 The following figure shows a typical organization of the logs:
 174
 175   _________________________________________________________________________
 176  | Summary | regular file | file  | ... | ifile | cpfile | sufile | DAT |SR|
 177  |_blocks__|_or_directory_|_______|_____|_______|________|________|_____|__|
 178
 179
 180 To stride over segment boundaries, this sequence of files may be split
 181 into multiple logs.  The sequence of logs that should be treated as
 182 logically one log, is delimited with flags marked in the segment
 183 summary.  The recovery code of nilfs2 looks this boundary information
 184 to ensure atomicity of updates.
 185
 186 The super root block is inserted for every checkpoints.  It includes
 187 three special inodes, inodes for the DAT, cpfile, and sufile.  Inodes
 188 of regular files, directories, symlinks and other special files, are
 189 included in the ifile.  The inode of ifile itself is included in the
 190 corresponding checkpoint entry in the cpfile.  Thus, the hierarchy
 191 among NILFS2 files can be depicted as follows:
 192
 193   Super block (SB)
 194        |
 195        v
 196   Super root block (the latest cno=xx)
 197        |-- DAT
 198        |-- sufile
 199        `-- cpfile
 200               |-- ifile (cno=c1)
 201               |-- ifile (cno=c2) ---- file (ino=i1)
 202               :        :          |-- file (ino=i2)
 203               `-- ifile (cno=xx)  |-- file (ino=i3)
 204                                   :        :
 205                                   `-- file (ino=yy)
 206                                     ( regular file, directory, or symlink )
 207
 208 For detail on the format of each file, please see include/linux/nilfs2_fs.h.