blob: 23e6cce40f9c9855d1ac1621530e9e308e5a1c97 [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
54
55Superblock formats
56------------------
57
58The md driver can support a variety of different superblock formats.
59Currently, it supports superblock formats "0.90.0" and the "md-1" format
60introduced in the 2.5 development series.
61
62The kernel will autodetect which format superblock is being used.
63
64Superblock format '0' is treated differently to others for legacy
65reasons - it is the original superblock format.
66
67
68General Rules - apply for all superblock formats
69------------------------------------------------
70
71An array is 'created' by writing appropriate superblocks to all
72devices.
73
74It is 'assembled' by associating each of these devices with an
75particular md virtual device. Once it is completely assembled, it can
76be accessed.
77
78An array should be created by a user-space tool. This will write
79superblocks to all devices. It will usually mark the array as
80'unclean', or with some devices missing so that the kernel md driver
81can create appropriate redundancy (copying in raid1, parity
82calculation in raid4/5).
83
84When an array is assembled, it is first initialized with the
85SET_ARRAY_INFO ioctl. This contains, in particular, a major and minor
86version number. The major version number selects which superblock
87format is to be used. The minor number might be used to tune handling
88of the format, such as suggesting where on each device to look for the
89superblock.
90
91Then each device is added using the ADD_NEW_DISK ioctl. This
92provides, in particular, a major and minor number identifying the
93device to add.
94
95The array is started with the RUN_ARRAY ioctl.
96
97Once started, new devices can be added. They should have an
98appropriate superblock written to them, and then passed be in with
99ADD_NEW_DISK.
100
101Devices that have failed or are not yet active can be detached from an
102array using HOT_REMOVE_DISK.
103
104
105Specific Rules that apply to format-0 super block arrays, and
106 arrays with no superblock (non-persistent).
107-------------------------------------------------------------
108
109An array can be 'created' by describing the array (level, chunksize
110etc) in a SET_ARRAY_INFO ioctl. This must has major_version==0 and
111raid_disks != 0.
112
113Then uninitialized devices can be added with ADD_NEW_DISK. The
114structure passed to ADD_NEW_DISK must specify the state of the device
115and it's role in the array.
116
117Once started with RUN_ARRAY, uninitialized spares can be added with
118HOT_ADD_DISK.
NeilBrownbb636542005-11-08 21:39:45 -0800119
120
121
122MD devices in sysfs
123-------------------
124md devices appear in sysfs (/sys) as regular block devices,
125e.g.
126 /sys/block/md0
127
128Each 'md' device will contain a subdirectory called 'md' which
129contains further md-specific information about the device.
130
131All md devices contain:
132 level
133 a text file indicating the 'raid level'. This may be a standard
134 numerical level prefixed by "RAID-" - e.g. "RAID-5", or some
135 other name such as "linear" or "multipath".
136 If no raid level has been set yet (array is still being
137 assembled), this file will be empty.
138
139 raid_disks
140 a text file with a simple number indicating the number of devices
141 in a fully functional array. If this is not yet known, the file
142 will be empty. If an array is being resized (not currently
143 possible) this will contain the larger of the old and new sizes.
144
145As component devices are added to an md array, they appear in the 'md'
146directory as new directories named
147 dev-XXX
148where XXX is a name that the kernel knows for the device, e.g. hdb1.
149Each directory contains:
150
151 block
152 a symlink to the block device in /sys/block, e.g.
153 /sys/block/md0/md/dev-hdb1/block -> ../../../../block/hdb/hdb1
154
155 super
156 A file containing an image of the superblock read from, or
157 written to, that device.
158
159 state
160 A file recording the current state of the device in the array
161 which can be a comma separated list of
162 faulty - device has been kicked from active use due to
163 a detected fault
164 in_sync - device is a fully in-sync member of the array
165 spare - device is working, but not a full member.
166 This includes spares that are in the process
167 of being recoverred to
168 This list make grow in future.
169
170
171An active md device will also contain and entry for each active device
172in the array. These are named
173
174 rdNN
175
176where 'NN' is the possition in the array, starting from 0.
177So for a 3 drive array there will be rd0, rd1, rd2.
178These are symbolic links to the appropriate 'dev-XXX' entry.
179Thus, for example,
180 cat /sys/block/md*/md/rd*/state
181will show 'in_sync' on every line.
182
183
184
185Active md devices for levels that support data redundancy (1,4,5,6)
186also have
187
188 sync_action
189 a text file that can be used to monitor and control the rebuild
190 process. It contains one word which can be one of:
191 resync - redundancy is being recalculated after unclean
192 shutdown or creation
193 recover - a hot spare is being built to replace a
194 failed/missing device
195 idle - nothing is happening
196 check - A full check of redundancy was requested and is
197 happening. This reads all block and checks
198 them. A repair may also happen for some raid
199 levels.
200 repair - A full check and repair is happening. This is
201 similar to 'resync', but was requested by the
202 user, and the write-intent bitmap is NOT used to
203 optimise the process.
204
205 This file is writable, and each of the strings that could be
206 read are meaningful for writing.
207
208 'idle' will stop an active resync/recovery etc. There is no
209 guarantee that another resync/recovery may not be automatically
210 started again, though some event will be needed to trigger
211 this.
212 'resync' or 'recovery' can be used to restart the
213 corresponding operation if it was stopped with 'idle'.
214 'check' and 'repair' will start the appropriate process
215 providing the current state is 'idle'.
216
217 mismatch_count
218 When performing 'check' and 'repair', and possibly when
219 performing 'resync', md will count the number of errors that are
220 found. The count in 'mismatch_cnt' is the number of sectors
221 that were re-written, or (for 'check') would have been
222 re-written. As most raid levels work in units of pages rather
223 than sectors, this my be larger than the number of actual errors
224 by a factor of the number of sectors in a page.
225
226Each active md device may also have attributes specific to the
227personality module that manages it.
228These are specific to the implementation of the module and could
229change substantially if the implementation changes.
230
231These currently include
232
233 stripe_cache_size (currently raid5 only)
234 number of entries in the stripe cache. This is writable, but
235 there are upper and lower limits (32768, 16). Default is 128.
236 strip_cache_active (currently raid5 only)
237 number of active entries in the stripe cache