1MKFS.BTRFS(8) Btrfs Manual MKFS.BTRFS(8)
2
6 mkfs.btrfs - create a btrfs filesystem
7
9 mkfs.btrfs [options] <device> [<device>...]
10
12 mkfs.btrfs is used to create the btrfs filesystem on a single or
13 multiple devices. <device> is typically a block device but can be a
14 file-backed image as well. Multiple devices are grouped by UUID of the
15 filesystem.
16 Before mounting such filesystem, the kernel module must know all the
18 devices either via preceding execution of btrfs device scan or using
19 the device mount option. See section MULTIPLE DEVICES for more details.
20
22 -b|--byte-count <size>
23 Specify the size of the filesystem. If this option is not used,
24 then mkfs.btrfs uses the entire device space for the filesystem.
25 --csum <type>, --checksum <type>
27 Specify the checksum algorithm. Default is crc32c. Valid values are
28 crc32c, xxhash, sha256 or blake2. To mount such filesystem kernel
29 must support the checksums as well. See CHECKSUM ALGORITHMS in
30 btrfs(5).
31 -d|--data <profile>
33 Specify the profile for the data block groups. Valid values are
34 raid0, raid1, raid5, raid6, raid10 or single or dup (case does not
35 matter).
36 See DUP PROFILES ON A SINGLE DEVICE for more details.
38 -m|--metadata <profile>
40 Specify the profile for the metadata block groups. Valid values are
41 raid0, raid1, raid5, raid6, raid10, single or dup (case does not
42 matter).
43 A single device filesystem will default to DUP, unless an SSD is
45 detected, in which case it will default to single. The detection is
46 based on the value of /sys/block/DEV/queue/rotational, where DEV is
47 the short name of the device.
48 Note that the rotational status can be arbitrarily set by the
50 underlying block device driver and may not reflect the true status
51 (network block device, memory-backed SCSI devices etc). Use the
52 options --data/--metadata to avoid confusion.
53 See DUP PROFILES ON A SINGLE DEVICE for more details.
55 -M|--mixed
57 Normally the data and metadata block groups are isolated. The mixed
58 mode will remove the isolation and store both types in the same
59 block group type. This helps to utilize the free space regardless
60 of the purpose and is suitable for small devices. The separate
61 allocation of block groups leads to a situation where the space is
62 reserved for the other block group type, is not available for
63 allocation and can lead to ENOSPC state.
64 The recommended size for the mixed mode is for filesystems less
66 than 1GiB. The soft recommendation is to use it for filesystems
67 smaller than 5GiB. The mixed mode may lead to degraded performance
68 on larger filesystems, but is otherwise usable, even on multiple
69 devices.
70 The nodesize and sectorsize must be equal, and the block group
72 types must match.
73 Note
75 versions up to 4.2.x forced the mixed mode for devices smaller
76 than 1GiB. This has been removed in 4.3+ as it caused some
77 usability issues.
78 -l|--leafsize <size>
80 Alias for --nodesize. Deprecated.
81 -n|--nodesize <size>
83 Specify the nodesize, the tree block size in which btrfs stores
84 metadata. The default value is 16KiB (16384) or the page size,
85 whichever is bigger. Must be a multiple of the sectorsize and a
86 power of 2, but not larger than 64KiB (65536). Leafsize always
87 equals nodesize and the options are aliases.
88 Smaller node size increases fragmentation but leads to taller
90 b-trees which in turn leads to lower locking contention. Higher
91 node sizes give better packing and less fragmentation at the cost
92 of more expensive memory operations while updating the metadata
93 blocks.
94 Note
96 versions up to 3.11 set the nodesize to 4k.
97 -s|--sectorsize <size>
99 Specify the sectorsize, the minimum data block allocation unit.
100 The default value is the page size and is autodetected. If the
102 sectorsize differs from the page size, the created filesystem may
103 not be mountable by the kernel. Therefore it is not recommended to
104 use this option unless you are going to mount it on a system with
105 the appropriate page size.
106 -L|--label <string>
108 Specify a label for the filesystem. The string should be less than
109 256 bytes and must not contain newline characters.
110 -K|--nodiscard
112 Do not perform whole device TRIM operation on devices that are
113 capable of that. This does not affect discard/trim operation when
114 the filesystem is mounted. Please see the mount option discard for
115 that in btrfs(5).
116 -r|--rootdir <rootdir>
118 Populate the toplevel subvolume with files from rootdir. This does
119 not require root permissions to write the new files or to mount the
120 filesystem.
121 Note
123 This option may enlarge the image or file to ensure it’s big
124 enough to contain the files from rootdir. Since version 4.14.1
125 the filesystem size is not minimized. Please see option
126 --shrink if you need that functionality.
127 --shrink
129 Shrink the filesystem to its minimal size, only works with
130 --rootdir option.
131 If the destination is a regular file, this option will also
133 truncate the file to the minimal size. Otherwise it will reduce the
134 filesystem available space. Extra space will not be usable unless
135 the filesystem is mounted and resized using btrfs filesystem
136 resize.
137 Note
139 prior to version 4.14.1, the shrinking was done automatically.
140 -O|--features <feature1>[,<feature2>...]
142 A list of filesystem features turned on at mkfs time. Not all
143 features are supported by old kernels. To disable a feature, prefix
144 it with ^.
145 See section FILESYSTEM FEATURES for more details. To see all
147 available features that mkfs.btrfs supports run:
148 mkfs.btrfs -O list-all
150 -f|--force
152 Forcibly overwrite the block devices when an existing filesystem is
153 detected. By default, mkfs.btrfs will utilize libblkid to check for
154 any known filesystem on the devices. Alternatively you can use the
155 wipefs utility to clear the devices.
156 -q|--quiet
158 Print only error or warning messages. Options --features or --help
159 are unaffected.
160 -U|--uuid <UUID>
162 Create the filesystem with the given UUID. The UUID must not exist
163 on any filesystem currently present.
164 -V|--version
166 Print the mkfs.btrfs version and exit.
167 --help
169 Print help.
170 -A|--alloc-start <offset>
172 deprecated, will be removed (An option to help debugging chunk
173 allocator.) Specify the (physical) offset from the start of the
174 device at which allocations start. The default value is zero.
175
177 The default unit is byte. All size parameters accept suffixes in the
178 1024 base. The recognized suffixes are: k, m, g, t, p, e, both
179 uppercase and lowercase.
180
182 Before mounting a multiple device filesystem, the kernel module must
183 know the association of the block devices that are attached to the
184 filesystem UUID.
185 There is typically no action needed from the user. On a system that
187 utilizes a udev-like daemon, any new block device is automatically
188 registered. The rules call btrfs device scan.
189 The same command can be used to trigger the device scanning if the
191 btrfs kernel module is reloaded (naturally all previous information
192 about the device registration is lost).
193 Another possibility is to use the mount options device to specify the
195 list of devices to scan at the time of mount.
196 # mount -o device=/dev/sdb,device=/dev/sdc /dev/sda /mnt
198 Note
201 that this means only scanning, if the devices do not exist in the
202 system, mount will fail anyway. This can happen on systems without
203 initramfs/initrd and root partition created with RAID1/10/5/6
204 profiles. The mount action can happen before all block devices are
205 discovered. The waiting is usually done on the initramfs/initrd
206 systems.
207 As of kernel 4.14, RAID5/6 is still considered experimental and
209 shouldn’t be employed for production use.
210
212 Features that can be enabled during creation time. See also btrfs(5)
213 section FILESYSTEM FEATURES.
214 mixed-bg
216 (kernel support since 2.6.37)
217 mixed data and metadata block groups, also set by option --mixed
219 extref
221 (default since btrfs-progs 3.12, kernel support since 3.7)
222 increased hardlink limit per file in a directory to 65536, older
224 kernels supported a varying number of hardlinks depending on the
225 sum of all file name sizes that can be stored into one metadata
226 block
227 raid56
229 (kernel support since 3.9)
230 extended format for RAID5/6, also enabled if raid5 or raid6 block
232 groups are selected
233 skinny-metadata
235 (default since btrfs-progs 3.18, kernel support since 3.10)
236 reduced-size metadata for extent references, saves a few percent of
238 metadata
239 no-holes
241 (kernel support since 3.14)
242 improved representation of file extents where holes are not
244 explicitly stored as an extent, saves a few percent of metadata if
245 sparse files are used
246
248 The highlevel organizational units of a filesystem are block groups of
249 three types: data, metadata and system.
250 DATA
252 store data blocks and nothing else
253 METADATA
255 store internal metadata in b-trees, can store file data if they fit
256 into the inline limit
257 SYSTEM
259 store structures that describe the mapping between the physical
260 devices and the linear logical space representing the filesystem
261 Other terms commonly used:
263 block group, chunk
265 a logical range of space of a given profile, stores data, metadata
266 or both; sometimes the terms are used interchangeably
267 A typical size of metadata block group is 256MiB (filesystem
269 smaller than 50GiB) and 1GiB (larger than 50GiB), for data it’s
270 1GiB. The system block group size is a few megabytes.
271 RAID
273 a block group profile type that utilizes RAID-like features on
274 multiple devices: striping, mirroring, parity
275 profile
277 when used in connection with block groups refers to the allocation
278 strategy and constraints, see the section PROFILES for more details
279
281 There are the following block group types available:
282 ┌────────┬────────────────────────────┬─────────────┬─────────────┐
284 │ │ │ │ │
285 │Profile │ Redundancy │ Space │ Min/max │
286 │ ├────────┬────────┬──────────┤ utilization │ devices │
287 │ │ │ │ │ │ │
288 │ │ Copies │ Parity │ Striping │ │ │
289 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
290 │ │ │ │ │ │ │
291 │single │ 1 │ │ │ 100% │ 1/any │
292 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
293 │ │ │ │ │ │ │
294 │DUP │ 2 / 1 │ │ │ 50% │ 1/any ^(see │
295 │ │ device │ │ │ │ note 1) │
296 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
297 │ │ │ │ │ │ │
298 │RAID0 │ │ │ 1 to N │ 100% │ 2/any │
299 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
300 │ │ │ │ │ │ │
301 │RAID1 │ 2 │ │ │ 50% │ 2/any │
302 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
303 │ │ │ │ │ │ │
304 │RAID1C3 │ 3 │ │ │ 33% │ 3/any │
305 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
306 │ │ │ │ │ │ │
307 │RAID1C4 │ 4 │ │ │ 25% │ 4/any │
308 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
309 │ │ │ │ │ │ │
310 │RAID10 │ 2 │ │ 1 to N │ 50% │ 4/any │
311 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
312 │ │ │ │ │ │ │
313 │RAID5 │ 1 │ 1 │ 2 to N-1 │ (N-1)/N │ 2/any ^(see │
314 │ │ │ │ │ │ note 2) │
315 ├────────┼────────┼────────┼──────────┼─────────────┼─────────────┤
316 │ │ │ │ │ │ │
317 │RAID6 │ 1 │ 2 │ 3 to N-2 │ (N-2)/N │ 3/any ^(see │
318 │ │ │ │ │ │ note 3) │
319 └────────┴────────┴────────┴──────────┴─────────────┴─────────────┘
320 Warning
322 It’s not recommended to create filesystems with RAID0/1/10/5/6
323 profiles on partitions from the same device. Neither redundancy nor
324 performance will be improved.
325 Note 1: DUP may exist on more than 1 device if it starts on a single
327 device and another one is added. Since version 4.5.1, mkfs.btrfs will
328 let you create DUP on multiple devices without restrictions.
329 Note 2: It’s not recommended to use 2 devices with RAID5. In that case,
331 parity stripe will contain the same data as the data stripe, making
332 RAID5 degraded to RAID1 with more overhead.
333 Note 3: It’s also not recommended to use 3 devices with RAID6, unless
335 you want to get effectively 3 copies in a RAID1-like manner (but not
336 exactly that).
337 Note 4: Since kernel 5.5 it’s possible to use RAID1C3 as replacement
339 for RAID6, higher space cost but reliable.
340 PROFILE LAYOUT
342 For the following examples, assume devices numbered by 1, 2, 3 and 4,
343 data or metadata blocks A, B, C, D, with possible stripes eg. A1, A2
344 that would be logically A, etc. For parity profiles PA and QA are
345 parity and syndrom, associated with the given stripe. The simple
346 layouts single or DUP are left out. Actual physical block placement on
347 devices depends on current state of the free/allocated space and may
348 appear random. All devices are assumed to be present at the time of the
349 blocks would have been written.
350 RAID1
352 ┌─────────┬──────────┬──────────┬──────────┐
354 │device 1 │ device 2 │ device 3 │ device 4 │
355 ├─────────┼──────────┼──────────┼──────────┤
356 │ │ │ │ │
357 │ A │ D │ │ │
358 ├─────────┼──────────┼──────────┼──────────┤
359 │ │ │ │ │
360 │ B │ │ │ C │
361 ├─────────┼──────────┼──────────┼──────────┤
362 │ │ │ │ │
363 │ C │ │ │ │
364 ├─────────┼──────────┼──────────┼──────────┤
365 │ │ │ │ │
366 │ D │ A │ B │ │
367 └─────────┴──────────┴──────────┴──────────┘
368 RAID1C3
370 ┌─────────┬──────────┬──────────┬──────────┐
372 │device 1 │ device 2 │ device 3 │ device 4 │
373 ├─────────┼──────────┼──────────┼──────────┤
374 │ │ │ │ │
375 │ A │ A │ D │ │
376 ├─────────┼──────────┼──────────┼──────────┤
377 │ │ │ │ │
378 │ B │ │ B │ │
379 ├─────────┼──────────┼──────────┼──────────┤
380 │ │ │ │ │
381 │ C │ │ A │ C │
382 ├─────────┼──────────┼──────────┼──────────┤
383 │ │ │ │ │
384 │ D │ D │ C │ B │
385 └─────────┴──────────┴──────────┴──────────┘
386 RAID0
388 ┌─────────┬──────────┬──────────┬──────────┐
390 │device 1 │ device 2 │ device 3 │ device 4 │
391 ├─────────┼──────────┼──────────┼──────────┤
392 │ │ │ │ │
393 │ A2 │ C3 │ A3 │ C2 │
394 ├─────────┼──────────┼──────────┼──────────┤
395 │ │ │ │ │
396 │ B1 │ A1 │ D2 │ B3 │
397 ├─────────┼──────────┼──────────┼──────────┤
398 │ │ │ │ │
399 │ C1 │ D3 │ B4 │ D1 │
400 ├─────────┼──────────┼──────────┼──────────┤
401 │ │ │ │ │
402 │ D4 │ B2 │ C4 │ A4 │
403 └─────────┴──────────┴──────────┴──────────┘
404 RAID5
406 ┌─────────┬──────────┬──────────┬──────────┐
408 │device 1 │ device 2 │ device 3 │ device 4 │
409 ├─────────┼──────────┼──────────┼──────────┤
410 │ │ │ │ │
411 │ A2 │ C3 │ A3 │ C2 │
412 ├─────────┼──────────┼──────────┼──────────┤
413 │ │ │ │ │
414 │ B1 │ A1 │ D2 │ B3 │
415 ├─────────┼──────────┼──────────┼──────────┤
416 │ │ │ │ │
417 │ C1 │ D3 │ PB │ D1 │
418 ├─────────┼──────────┼──────────┼──────────┤
419 │ │ │ │ │
420 │ PD │ B2 │ PC │ PA │
421 └─────────┴──────────┴──────────┴──────────┘
422 RAID6
424 ┌─────────┬──────────┬──────────┬──────────┐
426 │device 1 │ device 2 │ device 3 │ device 4 │
427 ├─────────┼──────────┼──────────┼──────────┤
428 │ │ │ │ │
429 │ A2 │ QC │ QA │ C2 │
430 ├─────────┼──────────┼──────────┼──────────┤
431 │ │ │ │ │
432 │ B1 │ A1 │ D2 │ QB │
433 ├─────────┼──────────┼──────────┼──────────┤
434 │ │ │ │ │
435 │ C1 │ QD │ PB │ D1 │
436 ├─────────┼──────────┼──────────┼──────────┤
437 │ │ │ │ │
438 │ PD │ B2 │ PC │ PA │
439 └─────────┴──────────┴──────────┴──────────┘
440
442 The mkfs utility will let the user create a filesystem with profiles
443 that write the logical blocks to 2 physical locations. Whether there
444 are really 2 physical copies highly depends on the underlying device
445 type.
446 For example, a SSD drive can remap the blocks internally to a single
448 copy—thus deduplicating them. This negates the purpose of increased
449 redundancy and just wastes filesystem space without providing the
450 expected level of redundancy.
451 The duplicated data/metadata may still be useful to statistically
453 improve the chances on a device that might perform some internal
454 optimizations. The actual details are not usually disclosed by vendors.
455 For example we could expect that not all blocks get deduplicated. This
456 will provide a non-zero probability of recovery compared to a zero
457 chance if the single profile is used. The user should make the tradeoff
458 decision. The deduplication in SSDs is thought to be widely available
459 so the reason behind the mkfs default is to not give a false sense of
460 redundancy.
461 As another example, the widely used USB flash or SD cards use a
463 translation layer between the logical and physical view of the device.
464 The data lifetime may be affected by frequent plugging. The memory
465 cells could get damaged, hopefully not destroying both copies of
466 particular data in case of DUP.
467 The wear levelling techniques can also lead to reduced redundancy, even
469 if the device does not do any deduplication. The controllers may put
470 data written in a short timespan into the same physical storage unit
471 (cell, block etc). In case this unit dies, both copies are lost. BTRFS
472 does not add any artificial delay between metadata writes.
473 The traditional rotational hard drives usually fail at the sector
475 level.
476 In any case, a device that starts to misbehave and repairs from the DUP
478 copy should be replaced! DUP is not backup.
479
481 SMALL FILESYSTEMS AND LARGE NODESIZE
482 The combination of small filesystem size and large nodesize is not
484 recommended in general and can lead to various ENOSPC-related issues
485 during mount time or runtime.
486 Since mixed block group creation is optional, we allow small filesystem
488 instances with differing values for sectorsize and nodesize to be
489 created and could end up in the following situation:
490 # mkfs.btrfs -f -n 65536 /dev/loop0
492 btrfs-progs v3.19-rc2-405-g976307c
493 See http://btrfs.wiki.kernel.org for more information.
494 Performing full device TRIM (512.00MiB) ...
496 Label: (null)
497 UUID: 49fab72e-0c8b-466b-a3ca-d1bfe56475f0
498 Node size: 65536
499 Sector size: 4096
500 Filesystem size: 512.00MiB
501 Block group profiles:
502 Data: single 8.00MiB
503 Metadata: DUP 40.00MiB
504 System: DUP 12.00MiB
505 SSD detected: no
506 Incompat features: extref, skinny-metadata
507 Number of devices: 1
508 Devices:
509 ID SIZE PATH
510 1 512.00MiB /dev/loop0
511 # mount /dev/loop0 /mnt/
513 mount: mount /dev/loop0 on /mnt failed: No space left on device
514 The ENOSPC occurs during the creation of the UUID tree. This is caused
516 by large metadata blocks and space reservation strategy that allocates
517 more than can fit into the filesystem.
518
520 mkfs.btrfs is part of btrfs-progs. Please refer to the btrfs wiki
521 http://btrfs.wiki.kernel.org for further details.
522
524 btrfs(5), btrfs(8), wipefs(8)
525Btrfs v5.4 12/03/2019 MKFS.BTRFS(8)