6 mkfs.btrfs - create a btrfs filesystem
10 *mkfs.btrfs* [options] <device> [<device>...]
14 *mkfs.btrfs* is used to create the btrfs filesystem on a single or multiple
15 devices. <device> is typically a block device but can be a file-backed image
16 as well. Multiple devices are grouped by UUID of the filesystem.
18 Before mounting such filesystem, the kernel module must know all the devices
19 either via preceding execution of *btrfs device scan* or using the *device*
20 mount option. See section *MULTIPLE DEVICES* for more details.
24 *-A|--alloc-start <offset>*::
25 (An option to help debugging chunk allocator.)
26 Specify the (physical) offset from the start of the device at which allocations
27 start. The default value is zero.
29 *-b|--byte-count <size>*::
30 Specify the size of the filesystem. If this option is not used,
31 mkfs.btrfs uses the entire device space for the filesystem.
33 *-d|--data <profile>*::
34 Specify the profile for the data block groups. Valid values are 'raid0',
35 'raid1', 'raid5', 'raid6', 'raid10' or 'single' or dup (case does not matter).
37 See 'DUP PROFILES ON A SINGLE DEVICE' for more.
39 *-m|--metadata <profile>*::
40 Specify the profile for the metadata block groups.
41 Valid values are 'raid0', 'raid1', 'raid5', 'raid6', 'raid10', 'single' or
42 'dup', (case does not matter).
44 A single device filesystem will default to 'DUP', unless a SSD is detected. Then
45 it will default to 'single'. The detection is based on the value of
46 `/sys/block/DEV/queue/rotational`, where 'DEV' is the short name of the device.
48 Note that the rotational status can be arbitrarily set by the underlying block
49 device driver and may not reflect the true status (network block device, memory-backed
50 SCSI devices etc). Use the options '--data/--metadata' to avoid confusion.
52 See 'DUP PROFILES ON A SINGLE DEVICE' for more details.
55 Normally the data and metadata block groups are isolated. The 'mixed' mode
56 will remove the isolation and store both types in the same block group type.
57 This helps to utilize the free space regardless of the purpose and is suitable
58 for small devices. The separate allocation of block groups leads to a situation
59 where the space is reserved for the other block group type, is not available for
60 allocation and can lead to ENOSPC state.
62 The recommended size for the mixed mode is for filesystems less than 1GiB. The
63 soft recommendation is to use it for filesystems smaller than 5GiB. The mixed
64 mode may lead to degraded performance on larger filesystems, but is otherwise
65 usable, even on multiple devices.
67 The 'nodesize' and 'sectorsize' must be equal, and the block group types must
70 NOTE: versions up to 4.2.x forced the mixed mode for devices smaller than 1GiB.
71 This has been removed in 4.3+ as it caused some usability issues.
73 *-l|--leafsize <size>*::
74 Alias for --nodesize. Deprecated.
76 *-n|--nodesize <size>*::
77 Specify the nodesize, the tree block size in which btrfs stores metadata. The
78 default value is 16KiB (16384) or the page size, whichever is bigger. Must be a
79 multiple of the sectorsize, but not larger than 64KiB (65536). Leafsize always
80 equals nodesize and the options are aliases.
82 Smaller node size increases fragmentation but lead to higher b-trees which in
83 turn leads to lower locking contention. Higher node sizes give better packing
84 and less fragmentation at the cost of more expensive memory operations while
85 updating the metadata blocks.
87 NOTE: versions up to 3.11 set the nodesize to 4k.
89 *-s|--sectorsize <size>*::
90 Specify the sectorsize, the minimum data block allocation unit.
92 The default value is the page size and is autodetected. If the sectorsize
93 differs from the page size, the created filesystem may not be mountable by the
94 kernel. Therefore it is not recommended to use this option unless you are going
95 to mount it on a system with the appropriate page size.
97 *-L|--label <string>*::
98 Specify a label for the filesystem. The 'string' should be less than 256
99 bytes and must not contain newline characters.
102 Do not perform whole device TRIM operation on devices that are capable of that.
103 This does not affect discard/trim operation when the filesystem is mounted.
104 Please see the mount option 'discard' for that in `btrfs`(5).
106 *-r|--rootdir <rootdir>*::
107 Populate the toplevel subvolume with files from 'rootdir'. This does not
108 require root permissions and does not mount the filesystem.
110 *-O|--features <feature1>[,<feature2>...]*::
111 A list of filesystem features turned on at mkfs time. Not all features are
112 supported by old kernels. To disable a feature, prefix it with '^'.
114 See section *FILESYSTEM FEATURES* for more details. To see all available
115 features that mkfs.btrfs supports run:
117 +mkfs.btrfs -O list-all+
120 Forcibly overwrite the block devices when an existing filesystem is detected.
121 By default, mkfs.btrfs will utilize 'libblkid' to check for any known
122 filesystem on the devices. Alternatively you can use the `wipefs` utility
123 to clear the devices.
126 Print only error or warning messages. Options --features or --help are unaffected.
129 Create the filesystem with the given 'UUID'. The UUID must not exist on any
130 filesystem currently present.
133 Print the *mkfs.btrfs* version and exit.
140 The default unit is 'byte'. All size parameters accept suffixes in the 1024
141 base. The recognized suffixes are: 'k', 'm', 'g', 't', 'p', 'e', both uppercase
147 Before mounting a multiple device filesystem, the kernel module must know the
148 association of the block devices that are attached to the filesystem UUID.
150 There is typically no action needed from the user. On a system that utilizes a
151 udev-like daemon, any new block device is automatically registered. The rules
152 call *btrfs device scan*.
154 The same command can be used to trigger the device scanning if the btrfs kernel
155 module is reloaded (naturally all previous information about the device
156 registration is lost).
158 Another possibility is to use the mount options *device* to specify the list of
159 devices to scan at the time of mount.
161 # mount -o device=/dev/sdb,device=/dev/sdc /dev/sda /mnt
163 NOTE: that this means only scanning, if the devices do not exist in the system,
164 mount will fail anyway. This can happen on systems without initramfs/initrd and
165 root partition created with RAID1/10/5/6 profiles. The mount action can happen
166 before all block devices are discovered. The waiting is usually done on the
167 initramfs/initrd systems.
172 Features that can be enabled during creation time. See also `btrfs`(5) section
173 'FILESYSTEM FEATURES'.
176 (kernel support since 2.6.37)
178 mixed data and metadata block groups, also set by option '--mixed'
181 (default since btrfs-progs 3.12, kernel support since 3.7)
183 increased hardlink limit per file in a directory to 65536, older kernels
184 supported a varying number of hardlinks depending on the sum of all file name
185 sizes that can be stored into one metadata block
188 (kernel support since 3.9)
190 extended format for RAID5/6, also enabled if raid5 or raid6 block groups
194 (default since btrfs-progs 3.18, kernel support since 3.10)
196 reduced-size metadata for extent references, saves a few percent of metadata
199 (kernel support since 3.14)
201 improved representation of file extents where holes are not explicitly
202 stored as an extent, saves a few percent of metadata if sparse files are used
204 BLOCK GROUPS, CHUNKS, RAID
205 --------------------------
207 The highlevel organizational units of a filesystem are block groups of three types:
208 data, metadata and system.
211 store data blocks and nothing else
214 store internal metadata in b-trees, can store file data if they fit into the
218 store structures that describe the mapping between the physical devices and the
219 linear logical space representing the filesystem
221 Other terms commonly used:
225 a logical range of space of a given profile, stores data, metadata or both;
226 sometimes the terms are used interchangeably
228 A typical size of metadata block group is 256MiB (filesystem smaller than
229 50GiB) and 1GiB (larger than 50GiB), for data it's 1GiB. The system block group
230 size is a few megabytes.
233 a block group profile type that utilizes RAID-like features on multiple
234 devices: striping, mirroring, parity
237 when used in connection with block groups refers to the allocation strategy
238 and constraints, see the section 'PROFILES' for more details
243 There are the following block group types available:
245 [ cols="^,^,^,^,^",width="60%" ]
246 |=============================================================
247 .2+^.<h| Profile 3+^.^h| Redundancy .2+^.<h| Min/max devices
248 ^.^h| Copies ^.^h| Parity ^.<h| Striping
249 | single | 1 | | | 1/any
250 | DUP | 2 / 1 device | | | 1/any ^(see note 1)^
251 | RAID0 | | | 1 to N | 2/any
252 | RAID1 | 2 | | | 2/any
253 | RAID10 | 2 | | 1 to N | 4/any
254 | RAID5 | 1 | 1 | 2 to N - 1 | 2/any ^(see note 2)^
255 | RAID6 | 1 | 2 | 3 to N - 2 | 3/any ^(see note 3)^
256 |=============================================================
258 WARNING: It's not recommended to build btrfs with RAID0/1/10/5/6 prfiles on
259 partitions from the same device. Neither redundancy nor performance will be
262 'Note 1:' DUP may exist on more than 1 device if it starts on a single device and
263 another one is added. Since version 4.5.1, *mkfs.btrfs* will let you create DUP
266 'Note 2:' It's not recommended to use 2 devices with RAID5. In that case,
267 parity stripe will contain the same data as the data stripe, making RAID5
268 degraded to RAID1 with more overhead.
270 'Note 3:' It's also not recommended to use 3 devices with RAID6, unless you
271 want to get effectively 3 copies in a RAID1-like manner (but not exactly that).
272 N-copies RAID1 is not implemented.
274 DUP PROFILES ON A SINGLE DEVICE
275 -------------------------------
277 The mkfs utility will let the user create a filesystem with profiles that write
278 the logical blocks to 2 physical locations. Whether there are really 2
279 physical copies highly depends on the underlying device type.
281 For example, a SSD drive can remap the blocks internally to a single copy thus
282 deduplicating them. This negates the purpose of increased redundancy and just
283 wastes filesystem space without the expected level of redundancy.
285 The duplicated data/metadata may still be useful to statistically improve the
286 chances on a device that might perform some internal optimizations. The actual
287 details are not usually disclosed by vendors. For example we could expect that
288 not all blocks get deduplicated. This will provide a non-zero probability of
289 recovery compared to a zero chance if the single profile is used. The user
290 should make the tradeoff decision. The deduplication in SSDs is thought to be
291 widely available so the reason behind the mkfs default is to not give a false
294 As another example, the widely used USB flash or SD cards use a translation
295 layer between the logical and physical view of the device. The data lifetime
296 may be affected by frequent plugging. The memory cells could get damaged,
297 hopefully not destroying both copies of particular data in case of DUP.
299 The wear levelling techniques can also lead to reduced redundancy, even if the
300 device does not do any deduplication. The controllers may put data written in
301 a short timespan into the same physical storage unit (cell, block etc). In case
302 this unit dies, both copies are lost. BTRFS does not add any artificial delay
303 between metadata writes.
305 The traditional rotational hard drives usually fail at the sector level.
307 In any case, a device that starts to misbehave and repairs from the DUP copy
308 should be replaced! *DUP is not backup*.
313 **SMALL FILESYSTEMS AND LARGE NODESIZE**
315 The combination of small filesystem size and large nodesize is not recommended
316 in general and can lead to various ENOSPC-related issues during mount time or runtime.
318 Since mixed block group creation is optional, we allow small
319 filesystem instances with differing values for 'sectorsize' and 'nodesize'
320 to be created and could end up in the following situation:
322 # mkfs.btrfs -f -n 65536 /dev/loop0
323 btrfs-progs v3.19-rc2-405-g976307c
324 See http://btrfs.wiki.kernel.org for more information.
326 Performing full device TRIM (512.00MiB) ...
328 UUID: 49fab72e-0c8b-466b-a3ca-d1bfe56475f0
331 Filesystem size: 512.00MiB
332 Block group profiles:
334 Metadata: DUP 40.00MiB
337 Incompat features: extref, skinny-metadata
341 1 512.00MiB /dev/loop0
343 # mount /dev/loop0 /mnt/
344 mount: mount /dev/loop0 on /mnt failed: No space left on device
346 The ENOSPC occurs during the creation of the UUID tree. This is caused
347 by large metadata blocks and space reservation strategy that allocates more
348 than can fit into the filesystem.
353 *mkfs.btrfs* is part of btrfs-progs.
354 Please refer to the btrfs wiki http://btrfs.wiki.kernel.org for