1 RAID arrays 1 RAID arrays
2 =========== 2 ===========
3 3
4 Boot time assembly of RAID arrays 4 Boot time assembly of RAID arrays
5 --------------------------------- 5 ---------------------------------
6 6
7 Tools that manage md devices can be found at 7 Tools that manage md devices can be found at
8 https://www.kernel.org/pub/linux/utils/raid 8 https://www.kernel.org/pub/linux/utils/raid/
9 9
10 10
11 You can boot with your md device with the foll 11 You can boot with your md device with the following kernel command
12 lines: 12 lines:
13 13
14 for old raid arrays without persistent superbl 14 for old raid arrays without persistent superblocks::
15 15
16 md=<md device no.>,<raid level>,<chunk size 16 md=<md device no.>,<raid level>,<chunk size factor>,<fault level>,dev0,dev1,...,devn
17 17
18 for raid arrays with persistent superblocks:: 18 for raid arrays with persistent superblocks::
19 19
20 md=<md device no.>,dev0,dev1,...,devn 20 md=<md device no.>,dev0,dev1,...,devn
21 21
22 or, to assemble a partitionable array:: 22 or, to assemble a partitionable array::
23 23
24 md=d<md device no.>,dev0,dev1,...,devn 24 md=d<md device no.>,dev0,dev1,...,devn
25 25
26 ``md device no.`` 26 ``md device no.``
27 +++++++++++++++++ 27 +++++++++++++++++
28 28
29 The number of the md device 29 The number of the md device
30 30
31 ================= ========= 31 ================= =========
32 ``md device no.`` device 32 ``md device no.`` device
33 ================= ========= 33 ================= =========
34 0 md0 34 0 md0
35 1 md1 35 1 md1
36 2 md2 36 2 md2
37 3 md3 37 3 md3
38 4 md4 38 4 md4
39 ================= ========= 39 ================= =========
40 40
41 ``raid level`` 41 ``raid level``
42 ++++++++++++++ 42 ++++++++++++++
43 43
44 level of the RAID array 44 level of the RAID array
45 45
46 =============== ============= 46 =============== =============
47 ``raid level`` level 47 ``raid level`` level
48 =============== ============= 48 =============== =============
49 -1 linear mode 49 -1 linear mode
50 0 striped mode 50 0 striped mode
51 =============== ============= 51 =============== =============
52 52
53 other modes are only supported with persistent 53 other modes are only supported with persistent super blocks
54 54
55 ``chunk size factor`` 55 ``chunk size factor``
56 +++++++++++++++++++++ 56 +++++++++++++++++++++
57 57
58 (raid-0 and raid-1 only) 58 (raid-0 and raid-1 only)
59 59
60 Set the chunk size as 4k << n. 60 Set the chunk size as 4k << n.
61 61
62 ``fault level`` 62 ``fault level``
63 +++++++++++++++ 63 +++++++++++++++
64 64
65 Totally ignored 65 Totally ignored
66 66
67 ``dev0`` to ``devn`` 67 ``dev0`` to ``devn``
68 ++++++++++++++++++++ 68 ++++++++++++++++++++
69 69
70 e.g. ``/dev/hda1``, ``/dev/hdc1``, ``/dev/sda1 70 e.g. ``/dev/hda1``, ``/dev/hdc1``, ``/dev/sda1``, ``/dev/sdb1``
71 71
72 A possible loadlin line (Harald Hoyer <HarryH@R 72 A possible loadlin line (Harald Hoyer <HarryH@Royal.Net>) looks like this::
73 73
74 e:\loadlin\loadlin e:\zimage root=/dev 74 e:\loadlin\loadlin e:\zimage root=/dev/md0 md=0,0,4,0,/dev/hdb2,/dev/hdc3 ro
75 75
76 76
77 Boot time autodetection of RAID arrays 77 Boot time autodetection of RAID arrays
78 -------------------------------------- 78 --------------------------------------
79 79
80 When md is compiled into the kernel (not as mo 80 When md is compiled into the kernel (not as module), partitions of
81 type 0xfd are scanned and automatically assemb 81 type 0xfd are scanned and automatically assembled into RAID arrays.
82 This autodetection may be suppressed with the 82 This autodetection may be suppressed with the kernel parameter
83 ``raid=noautodetect``. As of kernel 2.6.9, on 83 ``raid=noautodetect``. As of kernel 2.6.9, only drives with a type 0
84 superblock can be autodetected and run at boot 84 superblock can be autodetected and run at boot time.
85 85
86 The kernel parameter ``raid=partitionable`` (o 86 The kernel parameter ``raid=partitionable`` (or ``raid=part``) means
87 that all auto-detected arrays are assembled as 87 that all auto-detected arrays are assembled as partitionable.
88 88
89 Boot time assembly of degraded/dirty arrays 89 Boot time assembly of degraded/dirty arrays
90 ------------------------------------------- 90 -------------------------------------------
91 91
92 If a raid5 or raid6 array is both dirty and de 92 If a raid5 or raid6 array is both dirty and degraded, it could have
93 undetectable data corruption. This is because 93 undetectable data corruption. This is because the fact that it is
94 ``dirty`` means that the parity cannot be trus 94 ``dirty`` means that the parity cannot be trusted, and the fact that it
95 is degraded means that some datablocks are mis 95 is degraded means that some datablocks are missing and cannot reliably
96 be reconstructed (due to no parity). 96 be reconstructed (due to no parity).
97 97
98 For this reason, md will normally refuse to st 98 For this reason, md will normally refuse to start such an array. This
99 requires the sysadmin to take action to explic 99 requires the sysadmin to take action to explicitly start the array
100 despite possible corruption. This is normally 100 despite possible corruption. This is normally done with::
101 101
102 mdadm --assemble --force .... 102 mdadm --assemble --force ....
103 103
104 This option is not really available if the arr 104 This option is not really available if the array has the root
105 filesystem on it. In order to support this bo 105 filesystem on it. In order to support this booting from such an
106 array, md supports a module parameter ``start_ 106 array, md supports a module parameter ``start_dirty_degraded`` which,
107 when set to 1, bypassed the checks and will al 107 when set to 1, bypassed the checks and will allows dirty degraded
108 arrays to be started. 108 arrays to be started.
109 109
110 So, to boot with a root filesystem of a dirty 110 So, to boot with a root filesystem of a dirty degraded raid 5 or 6, use::
111 111
112 md-mod.start_dirty_degraded=1 112 md-mod.start_dirty_degraded=1
113 113
114 114
115 Superblock formats 115 Superblock formats
116 ------------------ 116 ------------------
117 117
118 The md driver can support a variety of differe 118 The md driver can support a variety of different superblock formats.
119 Currently, it supports superblock formats ``0. 119 Currently, it supports superblock formats ``0.90.0`` and the ``md-1`` format
120 introduced in the 2.5 development series. 120 introduced in the 2.5 development series.
121 121
122 The kernel will autodetect which format superb 122 The kernel will autodetect which format superblock is being used.
123 123
124 Superblock format ``0`` is treated differently 124 Superblock format ``0`` is treated differently to others for legacy
125 reasons - it is the original superblock format 125 reasons - it is the original superblock format.
126 126
127 127
128 General Rules - apply for all superblock forma 128 General Rules - apply for all superblock formats
129 ---------------------------------------------- 129 ------------------------------------------------
130 130
131 An array is ``created`` by writing appropriate 131 An array is ``created`` by writing appropriate superblocks to all
132 devices. 132 devices.
133 133
134 It is ``assembled`` by associating each of the 134 It is ``assembled`` by associating each of these devices with an
135 particular md virtual device. Once it is comp 135 particular md virtual device. Once it is completely assembled, it can
136 be accessed. 136 be accessed.
137 137
138 An array should be created by a user-space too 138 An array should be created by a user-space tool. This will write
139 superblocks to all devices. It will usually m 139 superblocks to all devices. It will usually mark the array as
140 ``unclean``, or with some devices missing so t 140 ``unclean``, or with some devices missing so that the kernel md driver
141 can create appropriate redundancy (copying in 141 can create appropriate redundancy (copying in raid 1, parity
142 calculation in raid 4/5). 142 calculation in raid 4/5).
143 143
144 When an array is assembled, it is first initia 144 When an array is assembled, it is first initialized with the
145 SET_ARRAY_INFO ioctl. This contains, in parti 145 SET_ARRAY_INFO ioctl. This contains, in particular, a major and minor
146 version number. The major version number sele 146 version number. The major version number selects which superblock
147 format is to be used. The minor number might 147 format is to be used. The minor number might be used to tune handling
148 of the format, such as suggesting where on eac 148 of the format, such as suggesting where on each device to look for the
149 superblock. 149 superblock.
150 150
151 Then each device is added using the ADD_NEW_DI 151 Then each device is added using the ADD_NEW_DISK ioctl. This
152 provides, in particular, a major and minor num 152 provides, in particular, a major and minor number identifying the
153 device to add. 153 device to add.
154 154
155 The array is started with the RUN_ARRAY ioctl. 155 The array is started with the RUN_ARRAY ioctl.
156 156
157 Once started, new devices can be added. They 157 Once started, new devices can be added. They should have an
158 appropriate superblock written to them, and th 158 appropriate superblock written to them, and then be passed in with
159 ADD_NEW_DISK. 159 ADD_NEW_DISK.
160 160
161 Devices that have failed or are not yet active 161 Devices that have failed or are not yet active can be detached from an
162 array using HOT_REMOVE_DISK. 162 array using HOT_REMOVE_DISK.
163 163
164 164
165 Specific Rules that apply to format-0 super bl 165 Specific Rules that apply to format-0 super block arrays, and arrays with no superblock (non-persistent)
166 ---------------------------------------------- 166 --------------------------------------------------------------------------------------------------------
167 167
168 An array can be ``created`` by describing the 168 An array can be ``created`` by describing the array (level, chunksize
169 etc) in a SET_ARRAY_INFO ioctl. This must hav 169 etc) in a SET_ARRAY_INFO ioctl. This must have ``major_version==0`` and
170 ``raid_disks != 0``. 170 ``raid_disks != 0``.
171 171
172 Then uninitialized devices can be added with A 172 Then uninitialized devices can be added with ADD_NEW_DISK. The
173 structure passed to ADD_NEW_DISK must specify 173 structure passed to ADD_NEW_DISK must specify the state of the device
174 and its role in the array. 174 and its role in the array.
175 175
176 Once started with RUN_ARRAY, uninitialized spa 176 Once started with RUN_ARRAY, uninitialized spares can be added with
177 HOT_ADD_DISK. 177 HOT_ADD_DISK.
178 178
179 179
180 MD devices in sysfs 180 MD devices in sysfs
181 ------------------- 181 -------------------
182 182
183 md devices appear in sysfs (``/sys``) as regul 183 md devices appear in sysfs (``/sys``) as regular block devices,
184 e.g.:: 184 e.g.::
185 185
186 /sys/block/md0 186 /sys/block/md0
187 187
188 Each ``md`` device will contain a subdirectory 188 Each ``md`` device will contain a subdirectory called ``md`` which
189 contains further md-specific information about 189 contains further md-specific information about the device.
190 190
191 All md devices contain: 191 All md devices contain:
192 192
193 level 193 level
194 a text file indicating the ``raid level`` 194 a text file indicating the ``raid level``. e.g. raid0, raid1,
195 raid5, linear, multipath, faulty. 195 raid5, linear, multipath, faulty.
196 If no raid level has been set yet (array 196 If no raid level has been set yet (array is still being
197 assembled), the value will reflect whatev 197 assembled), the value will reflect whatever has been written
198 to it, which may be a name like the above 198 to it, which may be a name like the above, or may be a number
199 such as ``0``, ``5``, etc. 199 such as ``0``, ``5``, etc.
200 200
201 raid_disks 201 raid_disks
202 a text file with a simple number indicati 202 a text file with a simple number indicating the number of devices
203 in a fully functional array. If this is 203 in a fully functional array. If this is not yet known, the file
204 will be empty. If an array is being resi 204 will be empty. If an array is being resized this will contain
205 the new number of devices. 205 the new number of devices.
206 Some raid levels allow this value to be s 206 Some raid levels allow this value to be set while the array is
207 active. This will reconfigure the array. 207 active. This will reconfigure the array. Otherwise it can only
208 be set while assembling an array. 208 be set while assembling an array.
209 A change to this attribute will not be pe 209 A change to this attribute will not be permitted if it would
210 reduce the size of the array. To reduce 210 reduce the size of the array. To reduce the number of drives
211 in an e.g. raid5, the array size must fir 211 in an e.g. raid5, the array size must first be reduced by
212 setting the ``array_size`` attribute. 212 setting the ``array_size`` attribute.
213 213
214 chunk_size 214 chunk_size
215 This is the size in bytes for ``chunks`` 215 This is the size in bytes for ``chunks`` and is only relevant to
216 raid levels that involve striping (0,4,5, 216 raid levels that involve striping (0,4,5,6,10). The address space
217 of the array is conceptually divided into 217 of the array is conceptually divided into chunks and consecutive
218 chunks are striped onto neighbouring devi 218 chunks are striped onto neighbouring devices.
219 The size should be at least PAGE_SIZE (4k 219 The size should be at least PAGE_SIZE (4k) and should be a power
220 of 2. This can only be set while assembl 220 of 2. This can only be set while assembling an array
221 221
222 layout 222 layout
223 The ``layout`` for the array for the part 223 The ``layout`` for the array for the particular level. This is
224 simply a number that is interpreted diffe 224 simply a number that is interpreted differently by different
225 levels. It can be written while assembli 225 levels. It can be written while assembling an array.
226 226
227 array_size 227 array_size
228 This can be used to artificially constrai 228 This can be used to artificially constrain the available space in
229 the array to be less than is actually ava 229 the array to be less than is actually available on the combined
230 devices. Writing a number (in Kilobytes) 230 devices. Writing a number (in Kilobytes) which is less than
231 the available size will set the size. An 231 the available size will set the size. Any reconfiguration of the
232 array (e.g. adding devices) will not caus 232 array (e.g. adding devices) will not cause the size to change.
233 Writing the word ``default`` will cause t 233 Writing the word ``default`` will cause the effective size of the
234 array to be whatever size is actually ava 234 array to be whatever size is actually available based on
235 ``level``, ``chunk_size`` and ``component 235 ``level``, ``chunk_size`` and ``component_size``.
236 236
237 This can be used to reduce the size of th 237 This can be used to reduce the size of the array before reducing
238 the number of devices in a raid4/5/6, or 238 the number of devices in a raid4/5/6, or to support external
239 metadata formats which mandate such clipp 239 metadata formats which mandate such clipping.
240 240
241 reshape_position 241 reshape_position
242 This is either ``none`` or a sector numbe 242 This is either ``none`` or a sector number within the devices of
243 the array where ``reshape`` is up to. If 243 the array where ``reshape`` is up to. If this is set, the three
244 attributes mentioned above (raid_disks, c 244 attributes mentioned above (raid_disks, chunk_size, layout) can
245 potentially have 2 values, an old and a n 245 potentially have 2 values, an old and a new value. If these
246 values differ, reading the attribute retu 246 values differ, reading the attribute returns::
247 247
248 new (old) 248 new (old)
249 249
250 and writing will effect the ``new`` value 250 and writing will effect the ``new`` value, leaving the ``old``
251 unchanged. 251 unchanged.
252 252
253 component_size 253 component_size
254 For arrays with data redundancy (i.e. not 254 For arrays with data redundancy (i.e. not raid0, linear, faulty,
255 multipath), all components must be the sa 255 multipath), all components must be the same size - or at least
256 there must a size that they all provide s 256 there must a size that they all provide space for. This is a key
257 part or the geometry of the array. It is 257 part or the geometry of the array. It is measured in sectors
258 and can be read from here. Writing to th 258 and can be read from here. Writing to this value may resize
259 the array if the personality supports it 259 the array if the personality supports it (raid1, raid5, raid6),
260 and if the component drives are large eno 260 and if the component drives are large enough.
261 261
262 metadata_version 262 metadata_version
263 This indicates the format that is being u 263 This indicates the format that is being used to record metadata
264 about the array. It can be 0.90 (traditi 264 about the array. It can be 0.90 (traditional format), 1.0, 1.1,
265 1.2 (newer format in varying locations) o 265 1.2 (newer format in varying locations) or ``none`` indicating that
266 the kernel isn't managing metadata at all 266 the kernel isn't managing metadata at all.
267 Alternately it can be ``external:`` follo 267 Alternately it can be ``external:`` followed by a string which
268 is set by user-space. This indicates tha 268 is set by user-space. This indicates that metadata is managed
269 by a user-space program. Any device fail 269 by a user-space program. Any device failure or other event that
270 requires a metadata update will cause arr 270 requires a metadata update will cause array activity to be
271 suspended until the event is acknowledged 271 suspended until the event is acknowledged.
272 272
273 resync_start 273 resync_start
274 The point at which resync should start. 274 The point at which resync should start. If no resync is needed,
275 this will be a very large number (or ``no 275 this will be a very large number (or ``none`` since 2.6.30-rc1). At
276 array creation it will default to 0, thou 276 array creation it will default to 0, though starting the array as
277 ``clean`` will set it much larger. 277 ``clean`` will set it much larger.
278 278
279 new_dev 279 new_dev
280 This file can be written but not read. T 280 This file can be written but not read. The value written should
281 be a block device number as major:minor. 281 be a block device number as major:minor. e.g. 8:0
282 This will cause that device to be attache 282 This will cause that device to be attached to the array, if it is
283 available. It will then appear at md/dev 283 available. It will then appear at md/dev-XXX (depending on the
284 name of the device) and further configura 284 name of the device) and further configuration is then possible.
285 285
286 safe_mode_delay 286 safe_mode_delay
287 When an md array has seen no write reques 287 When an md array has seen no write requests for a certain period
288 of time, it will be marked as ``clean``. 288 of time, it will be marked as ``clean``. When another write
289 request arrives, the array is marked as ` 289 request arrives, the array is marked as ``dirty`` before the write
290 commences. This is known as ``safe_mode` 290 commences. This is known as ``safe_mode``.
291 The ``certain period`` is controlled by t 291 The ``certain period`` is controlled by this file which stores the
292 period as a number of seconds. The defau 292 period as a number of seconds. The default is 200msec (0.200).
293 Writing a value of 0 disables safemode. 293 Writing a value of 0 disables safemode.
294 294
295 array_state 295 array_state
296 This file contains a single word which de 296 This file contains a single word which describes the current
297 state of the array. In many cases, the s 297 state of the array. In many cases, the state can be set by
298 writing the word for the desired state, h 298 writing the word for the desired state, however some states
299 cannot be explicitly set, and some transi 299 cannot be explicitly set, and some transitions are not allowed.
300 300
301 Select/poll works on this file. All chan 301 Select/poll works on this file. All changes except between
302 Active_idle and active (which can be freq 302 Active_idle and active (which can be frequent and are not
303 very interesting) are notified. active-> 303 very interesting) are notified. active->active_idle is
304 reported if the metadata is externally ma 304 reported if the metadata is externally managed.
305 305
306 clear 306 clear
307 No devices, no size, no level 307 No devices, no size, no level
308 308
309 Writing is equivalent to STOP_ARRAY i 309 Writing is equivalent to STOP_ARRAY ioctl
310 310
311 inactive 311 inactive
312 May have some settings, but array is 312 May have some settings, but array is not active
313 all IO results in error 313 all IO results in error
314 314
315 When written, doesn't tear down array 315 When written, doesn't tear down array, but just stops it
316 316
317 suspended (not supported yet) 317 suspended (not supported yet)
318 All IO requests will block. The array 318 All IO requests will block. The array can be reconfigured.
319 319
320 Writing this, if accepted, will block 320 Writing this, if accepted, will block until array is quiescent
321 321
322 readonly 322 readonly
323 no resync can happen. no superblocks 323 no resync can happen. no superblocks get written.
324 324
325 Write requests fail 325 Write requests fail
326 326
327 read-auto 327 read-auto
328 like readonly, but behaves like ``cle 328 like readonly, but behaves like ``clean`` on a write request.
329 329
330 clean 330 clean
331 no pending writes, but otherwise acti 331 no pending writes, but otherwise active.
332 332
333 When written to inactive array, start 333 When written to inactive array, starts without resync
334 334
335 If a write request arrives then 335 If a write request arrives then
336 if metadata is known, mark ``dirty`` 336 if metadata is known, mark ``dirty`` and switch to ``active``.
337 if not known, block and switch to wri 337 if not known, block and switch to write-pending
338 338
339 If written to an active array that ha 339 If written to an active array that has pending writes, then fails.
340 active 340 active
341 fully active: IO and resync can be ha 341 fully active: IO and resync can be happening.
342 When written to inactive array, start 342 When written to inactive array, starts with resync
343 343
344 write-pending 344 write-pending
345 clean, but writes are blocked waiting 345 clean, but writes are blocked waiting for ``active`` to be written.
346 346
347 active-idle 347 active-idle
348 like active, but no writes have been 348 like active, but no writes have been seen for a while (safe_mode_delay).
349 349
350 bitmap/location 350 bitmap/location
351 This indicates where the write-intent bit 351 This indicates where the write-intent bitmap for the array is
352 stored. 352 stored.
353 353
354 It can be one of ``none``, ``file`` or `` 354 It can be one of ``none``, ``file`` or ``[+-]N``.
355 ``file`` may later be extended to ``file: 355 ``file`` may later be extended to ``file:/file/name``
356 ``[+-]N`` means that many sectors from th 356 ``[+-]N`` means that many sectors from the start of the metadata.
357 357
358 This is replicated on all devices. For a 358 This is replicated on all devices. For arrays with externally
359 managed metadata, the offset is from the 359 managed metadata, the offset is from the beginning of the
360 device. 360 device.
361 361
362 bitmap/chunksize 362 bitmap/chunksize
363 The size, in bytes, of the chunk which wi 363 The size, in bytes, of the chunk which will be represented by a
364 single bit. For RAID456, it is a portion 364 single bit. For RAID456, it is a portion of an individual
365 device. For RAID10, it is a portion of th 365 device. For RAID10, it is a portion of the array. For RAID1, it
366 is both (they come to the same thing). 366 is both (they come to the same thing).
367 367
368 bitmap/time_base 368 bitmap/time_base
369 The time, in seconds, between looking for 369 The time, in seconds, between looking for bits in the bitmap to
370 be cleared. In the current implementation 370 be cleared. In the current implementation, a bit will be cleared
371 between 2 and 3 times ``time_base`` after 371 between 2 and 3 times ``time_base`` after all the covered blocks
372 are known to be in-sync. 372 are known to be in-sync.
373 373
374 bitmap/backlog 374 bitmap/backlog
375 When write-mostly devices are active in a 375 When write-mostly devices are active in a RAID1, write requests
376 to those devices proceed in the backgroun 376 to those devices proceed in the background - the filesystem (or
377 other user of the device) does not have t 377 other user of the device) does not have to wait for them.
378 ``backlog`` sets a limit on the number of 378 ``backlog`` sets a limit on the number of concurrent background
379 writes. If there are more than this, new 379 writes. If there are more than this, new writes will by
380 synchronous. 380 synchronous.
381 381
382 bitmap/metadata 382 bitmap/metadata
383 This can be either ``internal`` or ``exte 383 This can be either ``internal`` or ``external``.
384 384
385 ``internal`` 385 ``internal``
386 is the default and means the metadata f 386 is the default and means the metadata for the bitmap
387 is stored in the first 256 bytes of the 387 is stored in the first 256 bytes of the allocated space and is
388 managed by the md module. 388 managed by the md module.
389 389
390 ``external`` 390 ``external``
391 means that bitmap metadata is managed e 391 means that bitmap metadata is managed externally to
392 the kernel (i.e. by some userspace prog 392 the kernel (i.e. by some userspace program)
393 393
394 bitmap/can_clear 394 bitmap/can_clear
395 This is either ``true`` or ``false``. If 395 This is either ``true`` or ``false``. If ``true``, then bits in the
396 bitmap will be cleared when the correspon 396 bitmap will be cleared when the corresponding blocks are thought
397 to be in-sync. If ``false``, bits will n 397 to be in-sync. If ``false``, bits will never be cleared.
398 This is automatically set to ``false`` if 398 This is automatically set to ``false`` if a write happens on a
399 degraded array, or if the array becomes d 399 degraded array, or if the array becomes degraded during a write.
400 When metadata is managed externally, it s 400 When metadata is managed externally, it should be set to true
401 once the array becomes non-degraded, and 401 once the array becomes non-degraded, and this fact has been
402 recorded in the metadata. 402 recorded in the metadata.
403 403
404 consistency_policy 404 consistency_policy
405 This indicates how the array maintains co 405 This indicates how the array maintains consistency in case of unexpected
406 shutdown. It can be: 406 shutdown. It can be:
407 407
408 none 408 none
409 Array has no redundancy information, e. 409 Array has no redundancy information, e.g. raid0, linear.
410 410
411 resync 411 resync
412 Full resync is performed and all redund 412 Full resync is performed and all redundancy is regenerated when the
413 array is started after unclean shutdown 413 array is started after unclean shutdown.
414 414
415 bitmap 415 bitmap
416 Resync assisted by a write-intent bitma 416 Resync assisted by a write-intent bitmap.
417 417
418 journal 418 journal
419 For raid4/5/6, journal device is used t 419 For raid4/5/6, journal device is used to log transactions and replay
420 after unclean shutdown. 420 after unclean shutdown.
421 421
422 ppl 422 ppl
423 For raid5 only, Partial Parity Log is u 423 For raid5 only, Partial Parity Log is used to close the write hole and
424 eliminate resync. 424 eliminate resync.
425 425
426 The accepted values when writing to this 426 The accepted values when writing to this file are ``ppl`` and ``resync``,
427 used to enable and disable PPL. 427 used to enable and disable PPL.
428 428
429 uuid 429 uuid
430 This indicates the UUID of the array in t 430 This indicates the UUID of the array in the following format:
431 xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx 431 xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
432 432
433 433
434 As component devices are added to an md array, 434 As component devices are added to an md array, they appear in the ``md``
435 directory as new directories named:: 435 directory as new directories named::
436 436
437 dev-XXX 437 dev-XXX
438 438
439 where ``XXX`` is a name that the kernel knows 439 where ``XXX`` is a name that the kernel knows for the device, e.g. hdb1.
440 Each directory contains: 440 Each directory contains:
441 441
442 block 442 block
443 a symlink to the block device in /sys/ 443 a symlink to the block device in /sys/block, e.g.::
444 444
445 /sys/block/md0/md/dev-hdb1/block 445 /sys/block/md0/md/dev-hdb1/block -> ../../../../block/hdb/hdb1
446 446
447 super 447 super
448 A file containing an image of the supe 448 A file containing an image of the superblock read from, or
449 written to, that device. 449 written to, that device.
450 450
451 state 451 state
452 A file recording the current state of 452 A file recording the current state of the device in the array
453 which can be a comma separated list of 453 which can be a comma separated list of:
454 454
455 faulty 455 faulty
456 device has been kicked 456 device has been kicked from active use due to
457 a detected fault, or i 457 a detected fault, or it has unacknowledged bad
458 blocks 458 blocks
459 459
460 in_sync 460 in_sync
461 device is a fully in-s 461 device is a fully in-sync member of the array
462 462
463 writemostly 463 writemostly
464 device will only be su 464 device will only be subject to read
465 requests if there are 465 requests if there are no other options.
466 466
467 This applies only to r 467 This applies only to raid1 arrays.
468 468
469 blocked 469 blocked
470 device has failed, and 470 device has failed, and the failure hasn't been
471 acknowledged yet by th 471 acknowledged yet by the metadata handler.
472 472
473 Writes that would writ 473 Writes that would write to this device if
474 it were not faulty are 474 it were not faulty are blocked.
475 475
476 spare 476 spare
477 device is working, but 477 device is working, but not a full member.
478 478
479 This includes spares t 479 This includes spares that are in the process
480 of being recovered to 480 of being recovered to
481 481
482 write_error 482 write_error
483 device has ever seen a 483 device has ever seen a write error.
484 484
485 want_replacement 485 want_replacement
486 device is (mostly) wor 486 device is (mostly) working but probably
487 should be replaced, ei 487 should be replaced, either due to errors or
488 due to user request. 488 due to user request.
489 489
490 replacement 490 replacement
491 device is a replacemen 491 device is a replacement for another active
492 device with same raid_ 492 device with same raid_disk.
493 493
494 494
495 This list may grow in future. 495 This list may grow in future.
496 496
497 This can be written to. 497 This can be written to.
498 498
499 Writing ``faulty`` simulates a failur 499 Writing ``faulty`` simulates a failure on the device.
500 500
501 Writing ``remove`` removes the device 501 Writing ``remove`` removes the device from the array.
502 502
503 Writing ``writemostly`` sets the write 503 Writing ``writemostly`` sets the writemostly flag.
504 504
505 Writing ``-writemostly`` clears the wr 505 Writing ``-writemostly`` clears the writemostly flag.
506 506
507 Writing ``blocked`` sets the ``blocked 507 Writing ``blocked`` sets the ``blocked`` flag.
508 508
509 Writing ``-blocked`` clears the ``bloc 509 Writing ``-blocked`` clears the ``blocked`` flags and allows writes
510 to complete and possibly simulates an 510 to complete and possibly simulates an error.
511 511
512 Writing ``in_sync`` sets the in_sync f 512 Writing ``in_sync`` sets the in_sync flag.
513 513
514 Writing ``write_error`` sets writeerro 514 Writing ``write_error`` sets writeerrorseen flag.
515 515
516 Writing ``-write_error`` clears writee 516 Writing ``-write_error`` clears writeerrorseen flag.
517 517
518 Writing ``want_replacement`` is allowe 518 Writing ``want_replacement`` is allowed at any time except to a
519 replacement device or a spare. It set 519 replacement device or a spare. It sets the flag.
520 520
521 Writing ``-want_replacement`` is allow 521 Writing ``-want_replacement`` is allowed at any time. It clears
522 the flag. 522 the flag.
523 523
524 Writing ``replacement`` or ``-replacem 524 Writing ``replacement`` or ``-replacement`` is only allowed before
525 starting the array. It sets or clears 525 starting the array. It sets or clears the flag.
526 526
527 527
528 This file responds to select/poll. Any 528 This file responds to select/poll. Any change to ``faulty``
529 or ``blocked`` causes an event. 529 or ``blocked`` causes an event.
530 530
531 errors 531 errors
532 An approximate count of read errors th 532 An approximate count of read errors that have been detected on
533 this device but have not caused the de 533 this device but have not caused the device to be evicted from
534 the array (either because they were co 534 the array (either because they were corrected or because they
535 happened while the array was read-only 535 happened while the array was read-only). When using version-1
536 metadata, this value persists across r 536 metadata, this value persists across restarts of the array.
537 537
538 This value can be written while assemb 538 This value can be written while assembling an array thus
539 providing an ongoing count for arrays 539 providing an ongoing count for arrays with metadata managed by
540 userspace. 540 userspace.
541 541
542 slot 542 slot
543 This gives the role that the device ha 543 This gives the role that the device has in the array. It will
544 either be ``none`` if the device is no 544 either be ``none`` if the device is not active in the array
545 (i.e. is a spare or has failed) or an 545 (i.e. is a spare or has failed) or an integer less than the
546 ``raid_disks`` number for the array in 546 ``raid_disks`` number for the array indicating which position
547 it currently fills. This can only be 547 it currently fills. This can only be set while assembling an
548 array. A device for which this is set 548 array. A device for which this is set is assumed to be working.
549 549
550 offset 550 offset
551 This gives the location in the device 551 This gives the location in the device (in sectors from the
552 start) where data from the array will 552 start) where data from the array will be stored. Any part of
553 the device before this offset is not t 553 the device before this offset is not touched, unless it is
554 used for storing metadata (Formats 1.1 554 used for storing metadata (Formats 1.1 and 1.2).
555 555
556 size 556 size
557 The amount of the device, after the of 557 The amount of the device, after the offset, that can be used
558 for storage of data. This will normal 558 for storage of data. This will normally be the same as the
559 component_size. This can be written w 559 component_size. This can be written while assembling an
560 array. If a value less than the curre 560 array. If a value less than the current component_size is
561 written, it will be rejected. 561 written, it will be rejected.
562 562
563 recovery_start 563 recovery_start
564 When the device is not ``in_sync``, th 564 When the device is not ``in_sync``, this records the number of
565 sectors from the start of the device w 565 sectors from the start of the device which are known to be
566 correct. This is normally zero, but d 566 correct. This is normally zero, but during a recovery
567 operation it will steadily increase, a 567 operation it will steadily increase, and if the recovery is
568 interrupted, restoring this value can 568 interrupted, restoring this value can cause recovery to
569 avoid repeating the earlier blocks. W 569 avoid repeating the earlier blocks. With v1.x metadata, this
570 value is saved and restored automatica 570 value is saved and restored automatically.
571 571
572 This can be set whenever the device is 572 This can be set whenever the device is not an active member of
573 the array, either before the array is 573 the array, either before the array is activated, or before
574 the ``slot`` is set. 574 the ``slot`` is set.
575 575
576 Setting this to ``none`` is equivalent 576 Setting this to ``none`` is equivalent to setting ``in_sync``.
577 Setting to any other value also clears 577 Setting to any other value also clears the ``in_sync`` flag.
578 578
579 bad_blocks 579 bad_blocks
580 This gives the list of all known bad b 580 This gives the list of all known bad blocks in the form of
581 start address and length (in sectors r 581 start address and length (in sectors respectively). If output
582 is too big to fit in a page, it will b 582 is too big to fit in a page, it will be truncated. Writing
583 ``sector length`` to this file adds ne 583 ``sector length`` to this file adds new acknowledged (i.e.
584 recorded to disk safely) bad blocks. 584 recorded to disk safely) bad blocks.
585 585
586 unacknowledged_bad_blocks 586 unacknowledged_bad_blocks
587 This gives the list of known-but-not-y 587 This gives the list of known-but-not-yet-saved-to-disk bad
588 blocks in the same form of ``bad_block 588 blocks in the same form of ``bad_blocks``. If output is too big
589 to fit in a page, it will be truncated 589 to fit in a page, it will be truncated. Writing to this file
590 adds bad blocks without acknowledging 590 adds bad blocks without acknowledging them. This is largely
591 for testing. 591 for testing.
592 592
593 ppl_sector, ppl_size 593 ppl_sector, ppl_size
594 Location and size (in sectors) of the 594 Location and size (in sectors) of the space used for Partial Parity Log
595 on this device. 595 on this device.
596 596
597 597
598 An active md device will also contain an entry 598 An active md device will also contain an entry for each active device
599 in the array. These are named:: 599 in the array. These are named::
600 600
601 rdNN 601 rdNN
602 602
603 where ``NN`` is the position in the array, sta 603 where ``NN`` is the position in the array, starting from 0.
604 So for a 3 drive array there will be rd0, rd1, 604 So for a 3 drive array there will be rd0, rd1, rd2.
605 These are symbolic links to the appropriate `` 605 These are symbolic links to the appropriate ``dev-XXX`` entry.
606 Thus, for example:: 606 Thus, for example::
607 607
608 cat /sys/block/md*/md/rd*/state 608 cat /sys/block/md*/md/rd*/state
609 609
610 will show ``in_sync`` on every line. 610 will show ``in_sync`` on every line.
611 611
612 612
613 613
614 Active md devices for levels that support data 614 Active md devices for levels that support data redundancy (1,4,5,6,10)
615 also have 615 also have
616 616
617 sync_action 617 sync_action
618 a text file that can be used to monitor a 618 a text file that can be used to monitor and control the rebuild
619 process. It contains one word which can 619 process. It contains one word which can be one of:
620 620
621 resync 621 resync
622 redundancy is being recalculat 622 redundancy is being recalculated after unclean
623 shutdown or creation 623 shutdown or creation
624 624
625 recover 625 recover
626 a hot spare is being built to 626 a hot spare is being built to replace a
627 failed/missing device 627 failed/missing device
628 628
629 idle 629 idle
630 nothing is happening 630 nothing is happening
631 check 631 check
632 A full check of redundancy was 632 A full check of redundancy was requested and is
633 happening. This reads all blo 633 happening. This reads all blocks and checks
634 them. A repair may also happen 634 them. A repair may also happen for some raid
635 levels. 635 levels.
636 636
637 repair 637 repair
638 A full check and repair is hap 638 A full check and repair is happening. This is
639 similar to ``resync``, but was 639 similar to ``resync``, but was requested by the
640 user, and the write-intent bit 640 user, and the write-intent bitmap is NOT used to
641 optimise the process. 641 optimise the process.
642 642
643 This file is writable, and each of the s 643 This file is writable, and each of the strings that could be
644 read are meaningful for writing. 644 read are meaningful for writing.
645 645
646 ``idle`` will stop an active resync/re 646 ``idle`` will stop an active resync/recovery etc. There is no
647 guarantee that another resync/recovery 647 guarantee that another resync/recovery may not be automatically
648 started again, though some event will 648 started again, though some event will be needed to trigger
649 this. 649 this.
650 650
651 ``resync`` or ``recovery`` can be used 651 ``resync`` or ``recovery`` can be used to restart the
652 corresponding operation if it was stop 652 corresponding operation if it was stopped with ``idle``.
653 653
654 ``check`` and ``repair`` will start th 654 ``check`` and ``repair`` will start the appropriate process
655 providing the current state is ``idle` 655 providing the current state is ``idle``.
656 656
657 This file responds to select/poll. Any 657 This file responds to select/poll. Any important change in the value
658 triggers a poll event. Sometimes the va 658 triggers a poll event. Sometimes the value will briefly be
659 ``recover`` if a recovery seems to be ne 659 ``recover`` if a recovery seems to be needed, but cannot be
660 achieved. In that case, the transition t 660 achieved. In that case, the transition to ``recover`` isn't
661 notified, but the transition away is. 661 notified, but the transition away is.
662 662
663 degraded 663 degraded
664 This contains a count of the number of d 664 This contains a count of the number of devices by which the
665 arrays is degraded. So an optimal array 665 arrays is degraded. So an optimal array will show ``0``. A
666 single failed/missing drive will show `` 666 single failed/missing drive will show ``1``, etc.
667 667
668 This file responds to select/poll, any i 668 This file responds to select/poll, any increase or decrease
669 in the count of missing devices will tri 669 in the count of missing devices will trigger an event.
670 670
671 mismatch_count 671 mismatch_count
672 When performing ``check`` and ``repair`` 672 When performing ``check`` and ``repair``, and possibly when
673 performing ``resync``, md will count the 673 performing ``resync``, md will count the number of errors that are
674 found. The count in ``mismatch_cnt`` is 674 found. The count in ``mismatch_cnt`` is the number of sectors
675 that were re-written, or (for ``check``) 675 that were re-written, or (for ``check``) would have been
676 re-written. As most raid levels work in 676 re-written. As most raid levels work in units of pages rather
677 than sectors, this may be larger than th 677 than sectors, this may be larger than the number of actual errors
678 by a factor of the number of sectors in 678 by a factor of the number of sectors in a page.
679 679
680 bitmap_set_bits 680 bitmap_set_bits
681 If the array has a write-intent bitmap, 681 If the array has a write-intent bitmap, then writing to this
682 attribute can set bits in the bitmap, in 682 attribute can set bits in the bitmap, indicating that a resync
683 would need to check the corresponding bl 683 would need to check the corresponding blocks. Either individual
684 numbers or start-end pairs can be writte 684 numbers or start-end pairs can be written. Multiple numbers
685 can be separated by a space. 685 can be separated by a space.
686 686
687 Note that the numbers are ``bit`` number 687 Note that the numbers are ``bit`` numbers, not ``block`` numbers.
688 They should be scaled by the bitmap_chun 688 They should be scaled by the bitmap_chunksize.
689 689
690 sync_speed_min, sync_speed_max 690 sync_speed_min, sync_speed_max
691 This are similar to ``/proc/sys/dev/raid/ 691 This are similar to ``/proc/sys/dev/raid/speed_limit_{min,max}``
692 however they only apply to the particular 692 however they only apply to the particular array.
693 693
694 If no value has been written to these, or 694 If no value has been written to these, or if the word ``system``
695 is written, then the system-wide value is 695 is written, then the system-wide value is used. If a value,
696 in kibibytes-per-second is written, then 696 in kibibytes-per-second is written, then it is used.
697 697
698 When the files are read, they show the cu 698 When the files are read, they show the currently active value
699 followed by ``(local)`` or ``(system)`` d 699 followed by ``(local)`` or ``(system)`` depending on whether it is
700 a locally set or system-wide value. 700 a locally set or system-wide value.
701 701
702 sync_completed 702 sync_completed
703 This shows the number of sectors that hav 703 This shows the number of sectors that have been completed of
704 whatever the current sync_action is, foll 704 whatever the current sync_action is, followed by the number of
705 sectors in total that could need to be pr 705 sectors in total that could need to be processed. The two
706 numbers are separated by a ``/`` thus ef 706 numbers are separated by a ``/`` thus effectively showing one
707 value, a fraction of the process that is 707 value, a fraction of the process that is complete.
708 708
709 A ``select`` on this attribute will retur 709 A ``select`` on this attribute will return when resync completes,
710 when it reaches the current sync_max (bel 710 when it reaches the current sync_max (below) and possibly at
711 other times. 711 other times.
712 712
713 sync_speed 713 sync_speed
714 This shows the current actual speed, in K 714 This shows the current actual speed, in K/sec, of the current
715 sync_action. It is averaged over the las 715 sync_action. It is averaged over the last 30 seconds.
716 716
717 suspend_lo, suspend_hi 717 suspend_lo, suspend_hi
718 The two values, given as numbers of secto 718 The two values, given as numbers of sectors, indicate a range
719 within the array where IO will be blocked 719 within the array where IO will be blocked. This is currently
720 only supported for raid4/5/6. 720 only supported for raid4/5/6.
721 721
722 sync_min, sync_max 722 sync_min, sync_max
723 The two values, given as numbers of secto 723 The two values, given as numbers of sectors, indicate a range
724 within the array where ``check``/``repair 724 within the array where ``check``/``repair`` will operate. Must be
725 a multiple of chunk_size. When it reaches 725 a multiple of chunk_size. When it reaches ``sync_max`` it will
726 pause, rather than complete. 726 pause, rather than complete.
727 You can use ``select`` or ``poll`` on ``s 727 You can use ``select`` or ``poll`` on ``sync_completed`` to wait for
728 that number to reach sync_max. Then you 728 that number to reach sync_max. Then you can either increase
729 ``sync_max``, or can write ``idle`` to `` 729 ``sync_max``, or can write ``idle`` to ``sync_action``.
730 730
731 The value of ``max`` for ``sync_max`` eff 731 The value of ``max`` for ``sync_max`` effectively disables the limit.
732 When a resync is active, the value can on 732 When a resync is active, the value can only ever be increased,
733 never decreased. 733 never decreased.
734 The value of ``0`` is the minimum for ``s 734 The value of ``0`` is the minimum for ``sync_min``.
735 735
736 736
737 737
738 Each active md device may also have attributes 738 Each active md device may also have attributes specific to the
739 personality module that manages it. 739 personality module that manages it.
740 These are specific to the implementation of th 740 These are specific to the implementation of the module and could
741 change substantially if the implementation cha 741 change substantially if the implementation changes.
742 742
743 These currently include: 743 These currently include:
744 744
745 stripe_cache_size (currently raid5 only) 745 stripe_cache_size (currently raid5 only)
746 number of entries in the stripe cache. 746 number of entries in the stripe cache. This is writable, but
747 there are upper and lower limits (32768, 747 there are upper and lower limits (32768, 17). Default is 256.
748 748
749 strip_cache_active (currently raid5 only) 749 strip_cache_active (currently raid5 only)
750 number of active entries in the stripe c 750 number of active entries in the stripe cache
751 751
752 preread_bypass_threshold (currently raid5 on 752 preread_bypass_threshold (currently raid5 only)
753 number of times a stripe requiring prere 753 number of times a stripe requiring preread will be bypassed by
754 a stripe that does not require preread. 754 a stripe that does not require preread. For fairness defaults
755 to 1. Setting this to 0 disables bypass 755 to 1. Setting this to 0 disables bypass accounting and
756 requires preread stripes to wait until a 756 requires preread stripes to wait until all full-width stripe-
757 writes are complete. Valid values are 0 757 writes are complete. Valid values are 0 to stripe_cache_size.
758 758
759 journal_mode (currently raid5 only) 759 journal_mode (currently raid5 only)
760 The cache mode for raid5. raid5 could in 760 The cache mode for raid5. raid5 could include an extra disk for
761 caching. The mode can be "write-throuth" 761 caching. The mode can be "write-throuth" and "write-back". The
762 default is "write-through". 762 default is "write-through".
763 763
764 ppl_write_hint 764 ppl_write_hint
765 NVMe stream ID to be set for each PPL wr 765 NVMe stream ID to be set for each PPL write request.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.