1BTRFS-DEVICE(8) Btrfs Manual BTRFS-DEVICE(8)
2
6 btrfs-device - manage devices of btrfs filesystems
7
9 btrfs device <subcommand> <args>
10
12 The btrfs device command group is used to manage devices of the btrfs
13 filesystems.
14
16 Btrfs filesystem can be created on top of single or multiple block
17 devices. Data and metadata are organized in allocation profiles with
18 various redundancy policies. There’s some similarity with traditional
19 RAID levels, but this could be confusing to users familiar with the
20 traditional meaning. Due to the similarity, the RAID terminology is
21 widely used in the documentation. See mkfs.btrfs(8) for more details
22 and the exact profile capabilities and constraints.
23 The device management works on a mounted filesystem. Devices can be
25 added, removed or replaced, by commands provided by btrfs device and
26 btrfs replace.
27 The profiles can be also changed, provided there’s enough workspace to
29 do the conversion, using the btrfs balance command and namely the
30 filter convert.
31 Profile
33 A profile describes an allocation policy based on the
34 redundancy/replication constraints in connection with the number of
35 devices. The profile applies to data and metadata block groups
36 separately.
37 RAID level
39 Where applicable, the level refers to a profile that matches
40 constraints of the standard RAID levels. At the moment the
41 supported ones are: RAID0, RAID1, RAID10, RAID5 and RAID6.
42 See the section TYPICAL USECASES for some examples.
44
46 add [-Kf] <device> [<device>...] <path>
47 Add device(s) to the filesystem identified by <path>.
48 If applicable, a whole device discard (TRIM) operation is performed
50 prior to adding the device. A device with existing filesystem
51 detected by blkid(8) will prevent device addition and has to be
52 forced. Alternatively the filesystem can be wiped from the device
53 using eg. the wipefs(8) tool.
54 The operation is instant and does not affect existing data. The
56 operation merely adds the device to the filesystem structures and
57 creates some block groups headers.
58 Options
60 -K|--nodiscard
62 do not perform discard (TRIM) by default
63 -f|--force
65 force overwrite of existing filesystem on the given disk(s)
66 --enqueue
68 wait if there’s another exclusive operation running, otherwise
69 continue
70 remove [options] <device>|<devid> [<device>|<devid>...] <path>
72 Remove device(s) from a filesystem identified by <path>
73 Device removal must satisfy the profile constraints, otherwise the
75 command fails. The filesystem must be converted to profile(s) that
76 would allow the removal. This can typically happen when going down
77 from 2 devices to 1 and using the RAID1 profile. See the TYPICAL
78 USECASES section below.
79 The operation can take long as it needs to move all data from the
81 device.
82 It is possible to delete the device that was used to mount the
84 filesystem. The device entry in the mount table will be replaced by
85 another device name with the lowest device id.
86 If the filesystem is mounted in degraded mode (-o degraded),
88 special term missing can be used for device. In that case, the
89 first device that is described by the filesystem metadata, but not
90 present at the mount time will be removed.
91 Note
93 In most cases, there is only one missing device in degraded
94 mode, otherwise mount fails. If there are two or more devices
95 missing (e.g. possible in RAID6), you need specify missing as
96 many times as the number of missing devices to remove all of
97 them.
98 Options
99 --enqueue
101 wait if there’s another exclusive operation running, otherwise
102 continue
103 delete <device>|<devid> [<device>|<devid>...] <path>
105 Alias of remove kept for backward compatibility
106 ready <device>
108 Wait until all devices of a multiple-device filesystem are scanned
109 and registered within the kernel module. This is to provide a way
110 for automatic filesystem mounting tools to wait before the mount
111 can start. The device scan is only one of the preconditions and the
112 mount can fail for other reasons. Normal users usually do not need
113 this command and may safely ignore it.
114 scan [options] [<device> [<device>...]]
116 Scan devices for a btrfs filesystem and register them with the
117 kernel module. This allows mounting multiple-device filesystem by
118 specifying just one from the whole group.
119 If no devices are passed, all block devices that blkid reports to
121 contain btrfs are scanned.
122 The options --all-devices or -d can be used as a fallback in case
124 blkid is not available. If used, behavior is the same as if no
125 devices are passed.
126 The command can be run repeatedly. Devices that have been already
128 registered remain as such. Reloading the kernel module will drop
129 this information. There’s an alternative way of mounting
130 multiple-device filesystem without the need for prior scanning. See
131 the mount option device.
132 Options
134 -d|--all-devices
136 Enumerate and register all devices, use as a fallback in case
137 blkid is not available.
138 -u|--forget
140 Unregister a given device or all stale devices if no path is
141 given, the device must be unmounted otherwise it’s an error.
142 stats [options] <path>|<device>
144 Read and print the device IO error statistics for all devices of
145 the given filesystem identified by <path> or for a single <device>.
146 The filesystem must be mounted. See section DEVICE STATS for more
147 information about the reported statistics and the meaning.
148 Options
150 -z|--reset
152 Print the stats and reset the values to zero afterwards.
153 -c|--check
155 Check if the stats are all zeros and return 0 if it is so. Set
156 bit 6 of the return code if any of the statistics is no-zero.
157 The error values is 65 if reading stats from at least one
158 device failed, otherwise it’s 64.
159 usage [options] <path> [<path>...]
161 Show detailed information about internal allocations in devices.
162 Options
164 -b|--raw
166 raw numbers in bytes, without the B suffix
167 -h|--human-readable
169 print human friendly numbers, base 1024, this is the default
170 -H
172 print human friendly numbers, base 1000
173 --iec
175 select the 1024 base for the following options, according to
176 the IEC standard
177 --si
179 select the 1000 base for the following options, according to
180 the SI standard
181 -k|--kbytes
183 show sizes in KiB, or kB with --si
184 -m|--mbytes
186 show sizes in MiB, or MB with --si
187 -g|--gbytes
189 show sizes in GiB, or GB with --si
190 -t|--tbytes
192 show sizes in TiB, or TB with --si
193 If conflicting options are passed, the last one takes precedence.
195
197 STARTING WITH A SINGLE-DEVICE FILESYSTEM
198 Assume we’ve created a filesystem on a block device /dev/sda with
199 profile single/single (data/metadata), the device size is 50GiB and
200 we’ve used the whole device for the filesystem. The mount point is
201 /mnt.
202 The amount of data stored is 16GiB, metadata have allocated 2GiB.
204 ADD NEW DEVICE
206 We want to increase the total size of the filesystem and keep the
207 profiles. The size of the new device /dev/sdb is 100GiB.
208 $ btrfs device add /dev/sdb /mnt
210 The amount of free data space increases by less than 100GiB, some
212 space is allocated for metadata.
213 CONVERT TO RAID1
215 Now we want to increase the redundancy level of both data and
216 metadata, but we’ll do that in steps. Note, that the device sizes
217 are not equal and we’ll use that to show the capabilities of split
218 data/metadata and independent profiles.
219 The constraint for RAID1 gives us at most 50GiB of usable space and
221 exactly 2 copies will be stored on the devices.
222 First we’ll convert the metadata. As the metadata occupy less than
224 50GiB and there’s enough workspace for the conversion process, we
225 can do:
226 $ btrfs balance start -mconvert=raid1 /mnt
228 This operation can take a while, because all metadata have to be
230 moved and all block pointers updated. Depending on the physical
231 locations of the old and new blocks, the disk seeking is the key
232 factor affecting performance.
233 You’ll note that the system block group has been also converted to
235 RAID1, this normally happens as the system block group also holds
236 metadata (the physical to logical mappings).
237 What changed:
239 · available data space decreased by 3GiB, usable roughly (50 - 3)
241 + (100 - 3) = 144 GiB
242 · metadata redundancy increased
244 IOW, the unequal device sizes allow for combined space for data yet
246 improved redundancy for metadata. If we decide to increase
247 redundancy of data as well, we’re going to lose 50GiB of the second
248 device for obvious reasons.
249 $ btrfs balance start -dconvert=raid1 /mnt
251 The balance process needs some workspace (ie. a free device space
253 without any data or metadata block groups) so the command could
254 fail if there’s too much data or the block groups occupy the whole
255 first device.
256 The device size of /dev/sdb as seen by the filesystem remains
258 unchanged, but the logical space from 50-100GiB will be unused.
259 REMOVE DEVICE
261 Device removal must satisfy the profile constraints, otherwise the
262 command fails. For example:
263 $ btrfs device remove /dev/sda /mnt
265 ERROR: error removing device '/dev/sda': unable to go below two devices on raid1
266 In order to remove a device, you need to convert the profile in
268 this case:
269 $ btrfs balance start -mconvert=dup -dconvert=single /mnt
271 $ btrfs device remove /dev/sda /mnt
272
274 The device stats keep persistent record of several error classes
275 related to doing IO. The current values are printed at mount time and
276 updated during filesystem lifetime or from a scrub run.
277 $ btrfs device stats /dev/sda3
279 [/dev/sda3].write_io_errs 0
280 [/dev/sda3].read_io_errs 0
281 [/dev/sda3].flush_io_errs 0
282 [/dev/sda3].corruption_errs 0
283 [/dev/sda3].generation_errs 0
284 write_io_errs
286 Failed writes to the block devices, means that the layers beneath
287 the filesystem were not able to satisfy the write request.
288 read_io_errors
290 Read request analogy to write_io_errs.
291 flush_io_errs
293 Number of failed writes with the FLUSH flag set. The flushing is a
294 method of forcing a particular order between write requests and is
295 crucial for implementing crash consistency. In case of btrfs, all
296 the metadata blocks must be permanently stored on the block device
297 before the superblock is written.
298 corruption_errs
300 A block checksum mismatched or a corrupted metadata header was
301 found.
302 generation_errs
304 The block generation does not match the expected value (eg. stored
305 in the parent node).
306
308 btrfs device returns a zero exit status if it succeeds. Non zero is
309 returned in case of failure.
310 If the -s option is used, btrfs device stats will add 64 to the exit
312 status if any of the error counters is non-zero.
313
315 btrfs is part of btrfs-progs. Please refer to the btrfs wiki
316 http://btrfs.wiki.kernel.org for further details.
317
319 mkfs.btrfs(8), btrfs-replace(8), btrfs-balance(8)
320Btrfs v5.10 01/18/2021 BTRFS-DEVICE(8)