6 btrfs-device - manage devices of btrfs filesystems
10 *btrfs device* <subcommand> <args>
14 The *btrfs device* command group is used to manage devices of the btrfs filesystems.
18 Btrfs filesystem can be created on top of single or multiple block devices.
19 Data and metadata are organized in allocation profiles with various redundancy
20 policies. There's some similarity with traditional RAID levels, but this could
21 be confusing to users familiar with the traditional meaning. Due to the
22 similarity, the RAID terminology is widely used in the documentation. See
23 `mkfs.btrfs`(9) for more details and the exact profile capabilities and
26 The device management works on a mounted filesystem. Devices can be added,
27 removed or replaced, by commands profided by *btrfs device* and *btrfs replace*.
29 The profiles can be also changed, provided there's enough workspace to do the
30 conversion, using the *btrfs balance* command and namely the filter 'convert'.
33 A profile describes an allocation policy based on the redundancy/replication
34 constrants in connection with the number of devices. The profile applies to
35 data and metadata block groups separately.
38 Where applicable, the level refers to a profile that matches constraints of the
39 standard RAID levels. At the moment the supported ones are: RAID0, RAID1,
40 RAID10, RAID5 and RAID6.
42 See the section *TYPICAL USECASES* for some examples.
46 *add* [-Kf] <dev> [<dev>...] <path>::
47 Add device(s) to the filesystem identified by <path>.
49 If applicable, a whole device discard (TRIM) operation is performed prior to
50 adding the device. A device with existing filesystem detected by `blkid`(8)
51 will prevent device addition and has to be forced. Alternatively the filesystem
52 can be wiped from the device using eg. the `wipefs`(8) tool.
54 The operation is instant and does not affect existing data. The operation merely
55 adds the device to the filesystem structures and creates some block groups
61 do not perform discard (TRIM) by default
63 force overwrite of existing filesystem on the given disk(s)
65 *remove* <dev>|<devid> [<dev>|<devid>...] <path>::
66 Remove device(s) from a filesystem identified by <path>
68 Device removal must satisfy the profile constraints, otherwise the command
69 fails. The filesystem must be converted to profile(s) that would allow the
70 removal. This can typically happen when going down from 2 devices to 1 and
71 using the RAID1 profile. See the example section below.
73 The operation can take long as it needs to move all data from the device.
75 NOTE: It is not possible to delete the device that was used to mount the
76 filesystem. This is a limitation given by the VFS.
78 *delete* <dev>|<devid> [<dev>|<devid>...] <path>::
79 Alias of remove kept for backward compatibility
82 Wait until all devices of a multiple-device filesystem are scanned and registered
83 within the kernel module.
85 *scan* [(--all-devices|-d)|<device> [<device>...]]::
86 Scan devices for a btrfs filesystem and register them with the kernel module.
87 This allows mounting multiple-device filesystem by specifying just one from the
90 If no devices are passed, all block devices that blkid reports to contain btrfs
93 The options '--all-devices' or '-d' are deprecated and kept for backward compatibility.
94 If used, behavior is the same as if no devices are passed.
96 The command can be run repeatedly. Devices that have been already registered
97 remain as such. Reloading the kernel module will drop this information. There's
98 an alternative way of mounting multiple-device filesystem without the need for
99 prior scanning. See the mount option 'device'.
101 *stats* [options] <path>|<device>::
102 Read and print the device IO error statistics for all devices of the given
103 filesystem identified by <path> or for a single <device>. The filesystem must
104 be mounted. See section *DEVICE STATS* for more information about the reported
105 statistics and the meaning.
110 Print the stats and reset the values to zero afterwards.
113 Check if the stats are all zeros and return 0 it it is so. Set bit 6 of the
114 return code if any of the statistics is no-zero. The error values is 65 if
115 reading stats from at least one device failed, otherwise it's 64.
117 *usage* [options] <path> [<path>...]::
118 Show detailed information about internal allocations in devices.
123 raw numbers in bytes, without the 'B' suffix
124 -h|--human-readable::::
125 print human friendly numbers, base 1024, this is the default
127 print human friendly numbers, base 1000
129 select the 1024 base for the following options, according to the IEC standard
131 select the 1000 base for the following options, according to the SI standard
133 show sizes in KiB, or kB with --si
135 show sizes in MiB, or MB with --si
137 show sizes in GiB, or GB with --si
139 show sizes in TiB, or TB with --si
141 If conflicting options are passed, the last one takes precedence.
146 STARTING WITH A SINGLE-DEVICE FILESYSTEM
147 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
149 Assume we've created a filesystem on a block device '/dev/sda' with profile
150 'single/single' (data/metadata), the device size is 50GiB and we've used the
151 whole device for the filesystem. The mount point is '/mnt'.
153 The amount of data stored is 16GiB, metadata have allocated 2GiB.
155 ==== ADD NEW DEVICE ====
157 We want to increase the total size of the filesystem and keep the profiles. The
158 size of the new device '/dev/sdb' is 100GiB.
160 $ btrfs device add /dev/sdb /mnt
162 The amount of free data space increases by less than 100GiB, some space is
163 allocated for metadata.
165 ==== CONVERT TO RAID1 ====
167 Now we want to increase the redundancy level of both data and metadata, but
168 we'll do that in steps. Note, that the device sizes are not equal and we'll use
169 that to show the capabilities of split data/metadata and independent profiles.
171 The constraint for RAID1 gives us at most 50GiB of usable space and exactly 2
172 copies will be stored on the devices.
174 First we'll convert the metadata. As the metadata occupy less than 50GiB and
175 there's enough workspace for the conversion process, we can do:
177 $ btrfs balance start -mconvert=raid1 /mnt
179 This operation can take a while as the metadata have to be moved and all block
180 pointers updated. Depending on the physical locations of the old and new
181 blocks, the disk seeking is the key factor affecting performance.
183 You'll note that the system block group has been also converted to RAID1, this
184 normally happens as the system block group also holds metadata (the physical to
189 * available data space decreased by 3GiB, usable roughly (50 - 3) + (100 - 3) = 144 GiB
190 * metadata redundancy increased
192 IOW, the unequal device sizes allow for combined space for data yet improved
193 redundancy for metadata. If we decide to increase redundancy of data as well,
194 we're going to lose 50GiB of the second device for obvious reasons.
196 $ btrfs balance start -dconvert=raid1 /mnt
198 The balance process needs some workspace (ie. a free device space without any
199 data or metadata block groups) so the command could fail if there's too much
200 data or the block groups occupy the whole first device.
202 The device size of '/dev/sdb' as seen by the filesystem remains unchanged, but
203 the logical space from 50-100GiB will be unused.
208 The device stats keep persistent record of several error classes related to
209 doing IO. The current values are printed at mount time and updated during
210 filesystem lifetime or from a scrub run.
212 $ btrfs device stats /dev/sda3
213 [/dev/sda3].write_io_errs 0
214 [/dev/sda3].read_io_errs 0
215 [/dev/sda3].flush_io_errs 0
216 [/dev/sda3].corruption_errs 0
217 [/dev/sda3].generation_errs 0
220 Failed writes to the block devices, means that the layers beneath the
221 filesystem were not able to satisfy the write request.
223 Read request analogy to write_io_errs.
225 Number of failed writes with the 'FLUSH' flag set. The flushing is a method of
226 forcing a particular order between write requests and is crucial for
227 implementing crash consistency. In case of btrfs, all the metadata blocks must
228 be permanently stored on the block device before the superblock is written.
230 A block checksum mismatched or a corrupted metadata header was found.
232 The block generation does not match the expected value (eg. stored in the
237 *btrfs device* returns a zero exit status if it succeeds. Non zero is
238 returned in case of failure.
240 If the '-s' option is used, *btrfs device stats* will add 64 to the
241 exit status if any of the error counters is non-zero.
245 *btrfs* is part of btrfs-progs.
246 Please refer to the btrfs wiki http://btrfs.wiki.kernel.org for