blob: a8b430627473aa243995ab6f6e173b9cf1ff819e [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001Tools that manage md devices can be found at
2 http://www.<country>.kernel.org/pub/linux/utils/raid/....
3
4
5Boot time assembly of RAID arrays
6---------------------------------
7
8You can boot with your md device with the following kernel command
9lines:
10
11for old raid arrays without persistent superblocks:
12 md=<md device no.>,<raid level>,<chunk size factor>,<fault level>,dev0,dev1,...,devn
13
14for raid arrays with persistent superblocks
15 md=<md device no.>,dev0,dev1,...,devn
16or, to assemble a partitionable array:
17 md=d<md device no.>,dev0,dev1,...,devn
18
19md device no. = the number of the md device ...
20 0 means md0,
21 1 md1,
22 2 md2,
23 3 md3,
24 4 md4
25
26raid level = -1 linear mode
27 0 striped mode
28 other modes are only supported with persistent super blocks
29
30chunk size factor = (raid-0 and raid-1 only)
31 Set the chunk size as 4k << n.
32
33fault level = totally ignored
34
35dev0-devn: e.g. /dev/hda1,/dev/hdc1,/dev/sda1,/dev/sdb1
36
37A possible loadlin line (Harald Hoyer <HarryH@Royal.Net>) looks like this:
38
39e:\loadlin\loadlin e:\zimage root=/dev/md0 md=0,0,4,0,/dev/hdb2,/dev/hdc3 ro
40
41
42Boot time autodetection of RAID arrays
43--------------------------------------
44
45When md is compiled into the kernel (not as module), partitions of
46type 0xfd are scanned and automatically assembled into RAID arrays.
47This autodetection may be suppressed with the kernel parameter
48"raid=noautodetect". As of kernel 2.6.9, only drives with a type 0
49superblock can be autodetected and run at boot time.
50
51The kernel parameter "raid=partitionable" (or "raid=part") means
52that all auto-detected arrays are assembled as partitionable.
53
NeilBrown6ff8d8e2006-01-06 00:20:15 -080054Boot time assembly of degraded/dirty arrays
55-------------------------------------------
56
57If a raid5 or raid6 array is both dirty and degraded, it could have
58undetectable data corruption. This is because the fact that it is
59'dirty' means that the parity cannot be trusted, and the fact that it
60is degraded means that some datablocks are missing and cannot reliably
61be reconstructed (due to no parity).
62
63For this reason, md will normally refuse to start such an array. This
64requires the sysadmin to take action to explicitly start the array
Matt LaPlantefff92892006-10-03 22:47:42 +020065despite possible corruption. This is normally done with
NeilBrown6ff8d8e2006-01-06 00:20:15 -080066 mdadm --assemble --force ....
67
68This option is not really available if the array has the root
69filesystem on it. In order to support this booting from such an
70array, md supports a module parameter "start_dirty_degraded" which,
71when set to 1, bypassed the checks and will allows dirty degraded
72arrays to be started.
73
74So, to boot with a root filesystem of a dirty degraded raid[56], use
75
76 md-mod.start_dirty_degraded=1
77
Linus Torvalds1da177e2005-04-16 15:20:36 -070078
79Superblock formats
80------------------
81
82The md driver can support a variety of different superblock formats.
83Currently, it supports superblock formats "0.90.0" and the "md-1" format
84introduced in the 2.5 development series.
85
86The kernel will autodetect which format superblock is being used.
87
88Superblock format '0' is treated differently to others for legacy
89reasons - it is the original superblock format.
90
91
92General Rules - apply for all superblock formats
93------------------------------------------------
94
95An array is 'created' by writing appropriate superblocks to all
96devices.
97
98It is 'assembled' by associating each of these devices with an
99particular md virtual device. Once it is completely assembled, it can
100be accessed.
101
102An array should be created by a user-space tool. This will write
103superblocks to all devices. It will usually mark the array as
104'unclean', or with some devices missing so that the kernel md driver
105can create appropriate redundancy (copying in raid1, parity
106calculation in raid4/5).
107
108When an array is assembled, it is first initialized with the
109SET_ARRAY_INFO ioctl. This contains, in particular, a major and minor
110version number. The major version number selects which superblock
111format is to be used. The minor number might be used to tune handling
112of the format, such as suggesting where on each device to look for the
113superblock.
114
115Then each device is added using the ADD_NEW_DISK ioctl. This
116provides, in particular, a major and minor number identifying the
117device to add.
118
119The array is started with the RUN_ARRAY ioctl.
120
121Once started, new devices can be added. They should have an
122appropriate superblock written to them, and then passed be in with
123ADD_NEW_DISK.
124
125Devices that have failed or are not yet active can be detached from an
126array using HOT_REMOVE_DISK.
127
128
129Specific Rules that apply to format-0 super block arrays, and
130 arrays with no superblock (non-persistent).
131-------------------------------------------------------------
132
133An array can be 'created' by describing the array (level, chunksize
134etc) in a SET_ARRAY_INFO ioctl. This must has major_version==0 and
135raid_disks != 0.
136
137Then uninitialized devices can be added with ADD_NEW_DISK. The
138structure passed to ADD_NEW_DISK must specify the state of the device
139and it's role in the array.
140
141Once started with RUN_ARRAY, uninitialized spares can be added with
142HOT_ADD_DISK.
NeilBrownbb636542005-11-08 21:39:45 -0800143
144
145
146MD devices in sysfs
147-------------------
148md devices appear in sysfs (/sys) as regular block devices,
149e.g.
150 /sys/block/md0
151
152Each 'md' device will contain a subdirectory called 'md' which
153contains further md-specific information about the device.
154
155All md devices contain:
156 level
NeilBrownd33a56d2006-10-03 01:15:58 -0700157 a text file indicating the 'raid level'. e.g. raid0, raid1,
158 raid5, linear, multipath, faulty.
NeilBrownbb636542005-11-08 21:39:45 -0800159 If no raid level has been set yet (array is still being
NeilBrownd33a56d2006-10-03 01:15:58 -0700160 assembled), the value will reflect whatever has been written
161 to it, which may be a name like the above, or may be a number
162 such as '0', '5', etc.
NeilBrownbb636542005-11-08 21:39:45 -0800163
164 raid_disks
165 a text file with a simple number indicating the number of devices
166 in a fully functional array. If this is not yet known, the file
167 will be empty. If an array is being resized (not currently
168 possible) this will contain the larger of the old and new sizes.
NeilBrownda943b92006-01-06 00:20:54 -0800169 Some raid level (RAID1) allow this value to be set while the
170 array is active. This will reconfigure the array. Otherwise
171 it can only be set while assembling an array.
NeilBrownbb636542005-11-08 21:39:45 -0800172
NeilBrown3b343802006-01-06 00:20:47 -0800173 chunk_size
174 This is the size if bytes for 'chunks' and is only relevant to
175 raid levels that involve striping (1,4,5,6,10). The address space
176 of the array is conceptually divided into chunks and consecutive
177 chunks are striped onto neighbouring devices.
Matt LaPlante3f6dee92006-10-03 22:45:33 +0200178 The size should be at least PAGE_SIZE (4k) and should be a power
NeilBrown3b343802006-01-06 00:20:47 -0800179 of 2. This can only be set while assembling an array
180
NeilBrown08a02ec2007-05-09 02:35:38 -0700181 layout
182 The "layout" for the array for the particular level. This is
183 simply a number that is interpretted differently by different
184 levels. It can be written while assembling an array.
185
186 reshape_position
187 This is either "none" or a sector number within the devices of
188 the array where "reshape" is up to. If this is set, the three
189 attributes mentioned above (raid_disks, chunk_size, layout) can
190 potentially have 2 values, an old and a new value. If these
191 values differ, reading the attribute returns
192 new (old)
193 and writing will effect the 'new' value, leaving the 'old'
194 unchanged.
195
NeilBrowna35b0d62006-01-06 00:20:49 -0800196 component_size
197 For arrays with data redundancy (i.e. not raid0, linear, faulty,
198 multipath), all components must be the same size - or at least
199 there must a size that they all provide space for. This is a key
200 part or the geometry of the array. It is measured in sectors
201 and can be read from here. Writing to this value may resize
202 the array if the personality supports it (raid1, raid5, raid6),
203 and if the component drives are large enough.
204
NeilBrown8bb93aa2006-01-06 00:20:50 -0800205 metadata_version
206 This indicates the format that is being used to record metadata
207 about the array. It can be 0.90 (traditional format), 1.0, 1.1,
208 1.2 (newer format in varying locations) or "none" indicating that
209 the kernel isn't managing metadata at all.
210
NeilBrowna94213b2006-06-26 00:28:00 -0700211 resync_start
212 The point at which resync should start. If no resync is needed,
213 this will be a very large number. At array creation it will
214 default to 0, though starting the array as 'clean' will
215 set it much larger.
216
NeilBrown6d7ff732006-01-06 00:21:16 -0800217 new_dev
218 This file can be written but not read. The value written should
219 be a block device number as major:minor. e.g. 8:0
220 This will cause that device to be attached to the array, if it is
221 available. It will then appear at md/dev-XXX (depending on the
222 name of the device) and further configuration is then possible.
223
NeilBrown16f17b32006-06-26 00:27:37 -0700224 safe_mode_delay
225 When an md array has seen no write requests for a certain period
226 of time, it will be marked as 'clean'. When another write
Matt LaPlantefff92892006-10-03 22:47:42 +0200227 request arrives, the array is marked as 'dirty' before the write
228 commences. This is known as 'safe_mode'.
NeilBrown16f17b32006-06-26 00:27:37 -0700229 The 'certain period' is controlled by this file which stores the
230 period as a number of seconds. The default is 200msec (0.200).
231 Writing a value of 0 disables safemode.
232
NeilBrown9e653b62006-06-26 00:27:58 -0700233 array_state
234 This file contains a single word which describes the current
235 state of the array. In many cases, the state can be set by
236 writing the word for the desired state, however some states
237 cannot be explicitly set, and some transitions are not allowed.
238
239 clear
240 No devices, no size, no level
241 Writing is equivalent to STOP_ARRAY ioctl
242 inactive
243 May have some settings, but array is not active
244 all IO results in error
245 When written, doesn't tear down array, but just stops it
246 suspended (not supported yet)
247 All IO requests will block. The array can be reconfigured.
248 Writing this, if accepted, will block until array is quiessent
249 readonly
250 no resync can happen. no superblocks get written.
251 write requests fail
252 read-auto
253 like readonly, but behaves like 'clean' on a write request.
254
255 clean - no pending writes, but otherwise active.
256 When written to inactive array, starts without resync
257 If a write request arrives then
258 if metadata is known, mark 'dirty' and switch to 'active'.
259 if not known, block and switch to write-pending
260 If written to an active array that has pending writes, then fails.
261 active
262 fully active: IO and resync can be happening.
263 When written to inactive array, starts with resync
264
265 write-pending
266 clean, but writes are blocked waiting for 'active' to be written.
267
268 active-idle
269 like active, but no writes have been seen for a while (safe_mode_delay).
270
271
NeilBrownbb636542005-11-08 21:39:45 -0800272As component devices are added to an md array, they appear in the 'md'
273directory as new directories named
274 dev-XXX
275where XXX is a name that the kernel knows for the device, e.g. hdb1.
276Each directory contains:
277
278 block
279 a symlink to the block device in /sys/block, e.g.
280 /sys/block/md0/md/dev-hdb1/block -> ../../../../block/hdb/hdb1
281
282 super
283 A file containing an image of the superblock read from, or
284 written to, that device.
285
286 state
287 A file recording the current state of the device in the array
288 which can be a comma separated list of
289 faulty - device has been kicked from active use due to
290 a detected fault
291 in_sync - device is a fully in-sync member of the array
NeilBrownf6556752006-06-26 00:28:01 -0700292 writemostly - device will only be subject to read
293 requests if there are no other options.
294 This applies only to raid1 arrays.
NeilBrownbb636542005-11-08 21:39:45 -0800295 spare - device is working, but not a full member.
296 This includes spares that are in the process
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +0200297 of being recovered to
298 This list may grow in future.
NeilBrown45dc2de2006-06-26 00:27:58 -0700299 This can be written to.
300 Writing "faulty" simulates a failure on the device.
301 Writing "remove" removes the device from the array.
NeilBrownf6556752006-06-26 00:28:01 -0700302 Writing "writemostly" sets the writemostly flag.
303 Writing "-writemostly" clears the writemostly flag.
NeilBrownbb636542005-11-08 21:39:45 -0800304
NeilBrown4dbcdc72006-01-06 00:20:52 -0800305 errors
306 An approximate count of read errors that have been detected on
307 this device but have not caused the device to be evicted from
308 the array (either because they were corrected or because they
309 happened while the array was read-only). When using version-1
310 metadata, this value persists across restarts of the array.
311
312 This value can be written while assembling an array thus
313 providing an ongoing count for arrays with metadata managed by
314 userspace.
315
NeilBrown014236d2006-01-06 00:20:55 -0800316 slot
317 This gives the role that the device has in the array. It will
318 either be 'none' if the device is not active in the array
319 (i.e. is a spare or has failed) or an integer less than the
Matt LaPlante992caac2006-10-03 22:52:05 +0200320 'raid_disks' number for the array indicating which position
NeilBrown014236d2006-01-06 00:20:55 -0800321 it currently fills. This can only be set while assembling an
322 array. A device for which this is set is assumed to be working.
323
NeilBrown93c8cad2006-01-06 00:20:56 -0800324 offset
325 This gives the location in the device (in sectors from the
326 start) where data from the array will be stored. Any part of
327 the device before this offset us not touched, unless it is
328 used for storing metadata (Formats 1.1 and 1.2).
329
NeilBrown83303b62006-01-06 00:21:06 -0800330 size
331 The amount of the device, after the offset, that can be used
332 for storage of data. This will normally be the same as the
333 component_size. This can be written while assembling an
334 array. If a value less than the current component_size is
335 written, component_size will be reduced to this value.
336
NeilBrownbb636542005-11-08 21:39:45 -0800337
338An active md device will also contain and entry for each active device
339in the array. These are named
340
341 rdNN
342
Matt LaPlante992caac2006-10-03 22:52:05 +0200343where 'NN' is the position in the array, starting from 0.
NeilBrownbb636542005-11-08 21:39:45 -0800344So for a 3 drive array there will be rd0, rd1, rd2.
345These are symbolic links to the appropriate 'dev-XXX' entry.
346Thus, for example,
347 cat /sys/block/md*/md/rd*/state
348will show 'in_sync' on every line.
349
350
351
352Active md devices for levels that support data redundancy (1,4,5,6)
353also have
354
355 sync_action
356 a text file that can be used to monitor and control the rebuild
357 process. It contains one word which can be one of:
358 resync - redundancy is being recalculated after unclean
359 shutdown or creation
360 recover - a hot spare is being built to replace a
361 failed/missing device
362 idle - nothing is happening
363 check - A full check of redundancy was requested and is
364 happening. This reads all block and checks
365 them. A repair may also happen for some raid
366 levels.
367 repair - A full check and repair is happening. This is
368 similar to 'resync', but was requested by the
369 user, and the write-intent bitmap is NOT used to
370 optimise the process.
371
372 This file is writable, and each of the strings that could be
373 read are meaningful for writing.
374
375 'idle' will stop an active resync/recovery etc. There is no
376 guarantee that another resync/recovery may not be automatically
377 started again, though some event will be needed to trigger
378 this.
379 'resync' or 'recovery' can be used to restart the
380 corresponding operation if it was stopped with 'idle'.
381 'check' and 'repair' will start the appropriate process
382 providing the current state is 'idle'.
383
384 mismatch_count
385 When performing 'check' and 'repair', and possibly when
386 performing 'resync', md will count the number of errors that are
387 found. The count in 'mismatch_cnt' is the number of sectors
388 that were re-written, or (for 'check') would have been
389 re-written. As most raid levels work in units of pages rather
390 than sectors, this my be larger than the number of actual errors
391 by a factor of the number of sectors in a page.
392
Paul Clements9b1d1da2006-10-03 01:15:49 -0700393 bitmap_set_bits
394 If the array has a write-intent bitmap, then writing to this
395 attribute can set bits in the bitmap, indicating that a resync
396 would need to check the corresponding blocks. Either individual
397 numbers or start-end pairs can be written. Multiple numbers
398 can be separated by a space.
399 Note that the numbers are 'bit' numbers, not 'block' numbers.
400 They should be scaled by the bitmap_chunksize.
401
NeilBrown08a02ec2007-05-09 02:35:38 -0700402 sync_speed_min
403 sync_speed_max
404 This are similar to /proc/sys/dev/raid/speed_limit_{min,max}
405 however they only apply to the particular array.
406 If no value has been written to these, of if the word 'system'
407 is written, then the system-wide value is used. If a value,
408 in kibibytes-per-second is written, then it is used.
409 When the files are read, they show the currently active value
410 followed by "(local)" or "(system)" depending on whether it is
411 a locally set or system-wide value.
412
413 sync_completed
414 This shows the number of sectors that have been completed of
415 whatever the current sync_action is, followed by the number of
416 sectors in total that could need to be processed. The two
417 numbers are separated by a '/' thus effectively showing one
418 value, a fraction of the process that is complete.
NeilBrownc6207272008-02-06 01:39:52 -0800419 A 'select' on this attribute will return when resync completes,
420 when it reaches the current sync_max (below) and possibly at
421 other times.
422
423 sync_max
424 This is a number of sectors at which point a resync/recovery
425 process will pause. When a resync is active, the value can
426 only ever be increased, never decreased. The value of 'max'
427 effectively disables the limit.
428
NeilBrown08a02ec2007-05-09 02:35:38 -0700429
430 sync_speed
431 This shows the current actual speed, in K/sec, of the current
432 sync_action. It is averaged over the last 30 seconds.
433
434 suspend_lo
435 suspend_hi
436 The two values, given as numbers of sectors, indicate a range
437 within the array where IO will be blocked. This is currently
438 only supported for raid4/5/6.
439
440
NeilBrownbb636542005-11-08 21:39:45 -0800441Each active md device may also have attributes specific to the
442personality module that manages it.
443These are specific to the implementation of the module and could
444change substantially if the implementation changes.
445
446These currently include
447
448 stripe_cache_size (currently raid5 only)
449 number of entries in the stripe cache. This is writable, but
450 there are upper and lower limits (32768, 16). Default is 128.
451 strip_cache_active (currently raid5 only)
452 number of active entries in the stripe cache
Dan Williams8b3e6cd2008-04-28 02:15:53 -0700453 preread_bypass_threshold (currently raid5 only)
454 number of times a stripe requiring preread will be bypassed by
455 a stripe that does not require preread. For fairness defaults
456 to 1. Setting this to 0 disables bypass accounting and
457 requires preread stripes to wait until all full-width stripe-
458 writes are complete. Valid values are 0 to stripe_cache_size.