2 ''' Copyright Neil Brown and others.
3 ''' This program is free software; you can redistribute it and/or modify
4 ''' it under the terms of the GNU General Public License as published by
5 ''' the Free Software Foundation; either version 2 of the License, or
6 ''' (at your option) any later version.
7 ''' See file COPYING in distribution for details.
10 mdadm \- manage MD devices
16 .BI mdadm " [mode] <raiddevice> [options] <component-devices>"
19 RAID devices are virtual devices created from two or more
20 real block devices. This allows multiple devices (typically disk
21 drives or partitions thereof) to be combined into a single device to
22 hold (for example) a single filesystem.
23 Some RAID levels include redundancy and so can survive some degree of
26 Linux Software RAID devices are implemented through the md (Multiple
27 Devices) device driver.
29 Currently, Linux supports
45 is not a Software RAID mechanism, but does involve
47 each device is a path to one common physical storage device.
50 is also not true RAID, and it only involves one device. It
51 provides a layer over a true device that can be used to inject faults.
54 '''is a program that can be used to create, manage, and monitor
56 '''such it provides a similar set of functionality to the
59 '''The key differences between
66 '''is a single program and not a collection of programs.
69 '''can perform (almost) all of its functions without having a
70 '''configuration file and does not use one by default. Also
72 '''helps with management of the configuration
76 '''can provide information about your arrays (through Query, Detail, and Examine)
86 '''configuration file, at all. It has a different configuration file
87 '''with a different format and a different purpose.
90 mdadm has several major modes of operation:
93 Assemble the components of a previously created
94 array into an active array. Components can be explicitly given
95 or can be searched for.
97 checks that the components
98 do form a bona fide array, and can, on request, fiddle superblock
99 information so as to assemble a faulty array.
103 Build an array that doesn't have per-device superblocks. For these
106 cannot differentiate between initial creation and subsequent assembly
107 of an array. It also cannot perform any checks that appropriate
108 components have been requested. Because of this, the
110 mode should only be used together with a complete understanding of
115 Create a new array with per-device superblocks.
117 '''in several step create-add-add-run or it can all happen with one command.
120 .B "Follow or Monitor"
121 Monitor one or more md devices and act on any state changes. This is
122 only meaningful for raid1, 4, 5, 6, 10 or multipath arrays, as
123 only these have interesting state. raid0 or linear never have
124 missing, spare, or failed drives, so there is nothing to monitor.
128 Grow (or shrink) an array, or otherwise reshape it in some way.
129 Currently supported growth options including changing the active size
130 of component devices and changing the number of active devices in RAID
131 levels 1/4/5/6, as well as adding or removing a write-intent bitmap.
134 .B "Incremental Assembly"
135 Add a single device to an appropriate array. If the addition of the
136 device makes the array runnable, the array will be started.
137 This provides a convenient interface to a
139 system. As each device is detected,
141 has a chance to include it in some array as appropriate.
145 This is for doing things to specific components of an array such as
146 adding new spares and removing faulty devices.
150 This is an 'everything else' mode that supports operations on active
151 arrays, operations on component devices such as erasing old superblocks, and
152 information gathering operations.
153 '''This mode allows operations on independent devices such as examine MD
154 '''superblocks, erasing old superblocks and stopping active arrays.
158 This mode does not act on a specific device or array, but rather it
159 requests the Linux Kernel to activate any auto-detected arrays.
162 .SH Options for selecting a mode are:
165 .BR \-A ", " \-\-assemble
166 Assemble a pre-existing array.
169 .BR \-B ", " \-\-build
170 Build a legacy array without superblocks.
173 .BR \-C ", " \-\-create
177 .BR \-F ", " \-\-follow ", " \-\-monitor
183 .BR \-G ", " \-\-grow
184 Change the size or shape of an active array.
187 .BR \-I ", " \-\-incremental
188 Add a single device into an appropriate array, and possibly start the array.
192 Request that the kernel starts any auto-detected arrays. This can only
195 is compiled into the kernel \(em not if it is a module.
196 Arrays can be auto-detected by the kernel if all the components are in
197 primary MS-DOS partitions with partition type
199 In-kernel autodetect is not recommended for new installations. Using
201 to detect and assemble arrays \(em possibly in an
203 \(em is substantially more flexible and should be preferred.
206 If a device is given before any options, or if the first option is
211 then the MANAGE mode is assume.
212 Anything other than these will cause the
216 .SH Options that are not mode-specific are:
219 .BR \-h ", " \-\-help
220 Display general help message or, after one of the above options, a
221 mode-specific help message.
225 Display more detailed help about command line parsing and some commonly
229 .BR \-V ", " \-\-version
230 Print version information for mdadm.
233 .BR \-v ", " \-\-verbose
234 Be more verbose about what is happening. This can be used twice to be
236 The extra verbosity currently only affects
237 .B \-\-detail \-\-scan
239 .BR "\-\-examine \-\-scan" .
242 .BR \-q ", " \-\-quiet
243 Avoid printing purely informative messages. With this,
245 will be silent unless there is something really important to report.
248 .BR \-b ", " \-\-brief
249 Be less verbose. This is used with
257 gives an intermediate level of verbosity.
260 .BR \-f ", " \-\-force
261 Be more forceful about certain operations. See the various modes for
262 the exact meaning of this option in different contexts.
265 .BR \-c ", " \-\-config=
266 Specify the config file. Default is to use
267 .BR /etc/mdadm.conf ,
268 or if that is missing then
269 .BR /etc/mdadm/mdadm.conf .
270 If the config file given is
272 then nothing will be read, but
274 will act as though the config file contained exactly
275 .B "DEVICE partitions"
278 to find a list of devices to scan.
281 is given for the config file, then
283 will act as though the config file were empty.
286 .BR \-s ", " \-\-scan
289 for missing information.
290 In general, this option gives
292 permission to get any missing information (like component devices,
293 array devices, array identities, and alert destination) from the
294 configuration file (see previous option);
295 one exception is MISC mode when using
301 says to get a list of array devices from
305 .B \-e ", " \-\-metadata=
306 Declare the style of superblock (raid metadata) to be used. The
309 and to guess for other operations.
310 The default can be overridden by setting the
319 .IP "0, 0.90, default"
320 Use the original 0.90 format superblock. This format limits arrays to
321 28 component devices and limits component devices of levels 1 and
322 greater to 2 terabytes.
323 .IP "1, 1.0, 1.1, 1.2"
324 Use the new version-1 format superblock. This has few restrictions.
325 The different sub-versions store the superblock at different locations
326 on the device, either at the end (for 1.0), at the start (for 1.1) or
327 4K from the start (for 1.2).
332 This will override any
334 setting in the config file and provides the identity of the host which
335 should be considered the home for any arrays.
337 When creating an array, the
339 will be recorded in the superblock. For version-1 superblocks, it will
340 be prefixed to the array name. For version-0.90 superblocks, part of
341 the SHA1 hash of the hostname will be stored in the later half of the
344 When reporting information about an array, any array which is tagged
345 for the given homehost will be reported as such.
347 When using Auto-Assemble, only arrays tagged for the given homehost
350 .SH For create, build, or grow:
353 .BR \-n ", " \-\-raid\-devices=
354 Specify the number of active devices in the array. This, plus the
355 number of spare devices (see below) must equal the number of
357 (including "\fBmissing\fP" devices)
358 that are listed on the command line for
360 Setting a value of 1 is probably
361 a mistake and so requires that
363 be specified first. A value of 1 will then be allowed for linear,
364 multipath, raid0 and raid1. It is never allowed for raid4 or raid5.
366 This number can only be changed using
368 for RAID1, RAID5 and RAID6 arrays, and only on kernels which provide
372 .BR \-x ", " \-\-spare\-devices=
373 Specify the number of spare (eXtra) devices in the initial array.
374 Spares can also be added
375 and removed later. The number of component devices listed
376 on the command line must equal the number of raid devices plus the
377 number of spare devices.
381 .BR \-z ", " \-\-size=
382 Amount (in Kibibytes) of space to use from each drive in RAID level 1/4/5/6.
383 This must be a multiple of the chunk size, and must leave about 128Kb
384 of space at the end of the drive for the RAID superblock.
385 If this is not specified
386 (as it normally is not) the smallest drive (or partition) sets the
387 size, though if there is a variance among the drives of greater than 1%, a warning is
390 This value can be set with
392 for RAID level 1/4/5/6. If the array was created with a size smaller
393 than the currently active drives, the extra space can be accessed
396 The size can be given as
398 which means to choose the largest size that fits on all current drives.
401 .BR \-c ", " \-\-chunk=
402 Specify chunk size of kibibytes. The default is 64.
406 Specify rounding factor for linear array (==chunk size)
409 .BR \-l ", " \-\-level=
410 Set raid level. When used with
412 options are: linear, raid0, 0, stripe, raid1, 1, mirror, raid4, 4,
413 raid5, 5, raid6, 6, raid10, 10, multipath, mp, faulty. Obviously some of these are synonymous.
417 only linear, stripe, raid0, 0, raid1, multipath, mp, and faulty are valid.
419 Not yet supported with
423 .BR \-p ", " \-\-layout=
424 This option configures the fine details of data layout for raid5,
425 and raid10 arrays, and controls the failure modes for
428 The layout of the raid5 parity block can be one of
429 .BR left\-asymmetric ,
430 .BR left\-symmetric ,
431 .BR right\-asymmetric ,
432 .BR right\-symmetric ,
433 .BR la ", " ra ", " ls ", " rs .
435 .BR left\-symmetric .
437 When setting the failure mode for level
440 .BR write\-transient ", " wt ,
441 .BR read\-transient ", " rt ,
442 .BR write\-persistent ", " wp ,
443 .BR read\-persistent ", " rp ,
445 .BR read\-fixable ", " rf ,
446 .BR clear ", " flush ", " none .
448 Each failure mode can be followed by a number, which is used as a period
449 between fault generation. Without a number, the fault is generated
450 once on the first relevant request. With a number, the fault will be
451 generated after that many requests, and will continue to be generated
452 every time the period elapses.
454 Multiple failure modes can be current simultaneously by using the
456 option to set subsequent failure modes.
458 "clear" or "none" will remove any pending or periodic failure modes,
459 and "flush" will clear any persistent faults.
461 To set the parity with
463 the level of the array ("faulty")
464 must be specified before the fault mode is specified.
466 Finally, the layout options for RAID10 are one of 'n', 'o' or 'f' followed
467 by a small number. The default is 'n2'. The supported options are:
470 signals 'near' copies. Multiple copies of one data block are at
471 similar offsets in different devices.
474 signals 'offset' copies. Rather than the chunks being duplicated
475 within a stripe, whole stripes are duplicated but are rotated by one
476 device so duplicate blocks are on different devices. Thus subsequent
477 copies of a block are in the next drive, and are one chunk further
482 (multiple copies have very different offsets).
483 See md(4) for more detail about 'near' and 'far'.
485 The number is the number of copies of each datablock. 2 is normal, 3
486 can be useful. This number can be at most equal to the number of
487 devices in the array. It does not need to divide evenly into that
488 number (e.g. it is perfectly legal to have an 'n2' layout for an array
489 with an odd number of devices).
495 (thus explaining the p of
499 .BR \-b ", " \-\-bitmap=
500 Specify a file to store a write-intent bitmap in. The file should not
503 is also given. The same file should be provided
504 when assembling the array. If the word
506 is given, then the bitmap is stored with the metadata on the array,
507 and so is replicated on all devices. If the word
511 mode, then any bitmap that is present is removed.
513 To help catch typing errors, the filename must contain at least one
514 slash ('/') if it is a real file (not 'internal' or 'none').
516 Note: external bitmaps are only known to work on ext2 and ext3.
517 Storing bitmap files on other filesystems may result in serious problems.
520 .BR \-\-bitmap\-chunk=
521 Set the chunksize of the bitmap. Each bit corresponds to that many
522 Kilobytes of storage.
523 When using a file based bitmap, the default is to use the smallest
524 size that is at-least 4 and requires no more than 2^21 chunks.
527 bitmap, the chunksize is automatically determined to make best use of
532 .BR \-W ", " \-\-write\-mostly
533 subsequent devices lists in a
538 command will be flagged as 'write-mostly'. This is valid for RAID1
539 only and means that the 'md' driver will avoid reading from these
540 devices if at all possible. This can be useful if mirroring over a
544 .BR \-\-write\-behind=
545 Specify that write-behind mode should be enabled (valid for RAID1
546 only). If an argument is specified, it will set the maximum number
547 of outstanding writes allowed. The default value is 256.
548 A write-intent bitmap is required in order to use write-behind
549 mode, and write-behind is only attempted on drives marked as
553 .BR \-\-assume\-clean
556 that the array pre-existed and is known to be clean. It can be useful
557 when trying to recover from a major failure as you can be sure that no
558 data will be affected unless you actually write to the array. It can
559 also be used when creating a RAID1 or RAID10 if you want to avoid the
560 initial resync, however this practice \(em while normally safe \(em is not
561 recommended. Use this only if you really know what you are doing.
564 .BR \-\-backup\-file=
567 is used to increase the number of
568 raid-devices in a RAID5 if there are no spare devices available.
569 See the section below on RAID_DEVICE CHANGES. The file should be
570 stored on a separate device, not on the raid array being reshaped.
573 .BR \-N ", " \-\-name=
576 for the array. This is currently only effective when creating an
577 array with a version-1 superblock. The name is a simple textual
578 string that can be used to identify array components when assembling.
584 run the array, even if some of the components
585 appear to be active in another array or filesystem. Normally
587 will ask for confirmation before including such components in an
588 array. This option causes that question to be suppressed.
591 .BR \-f ", " \-\-force
594 accept the geometry and layout specified without question. Normally
596 will not allow creation of an array with only one device, and will try
597 to create a raid5 array with one missing drive (as this makes the
598 initial resync work faster). With
601 will not try to be so clever.
604 .BR \-a ", " "\-\-auto{=no,yes,md,mdp,part,p}{NN}"
605 Instruct mdadm to create the device file if needed, possibly allocating
606 an unused minor number. "md" causes a non-partitionable array
607 to be used. "mdp", "part" or "p" causes a partitionable array (2.6 and
608 later) to be used. "yes" requires the named md device to have
609 a 'standard' format, and the type and minor number will be determined
610 from this. See DEVICE NAMES below.
612 The argument can also come immediately after
617 is not given on the command line or in the config file, then
623 is also given, then any
625 entries in the config file will override the
627 instruction given on the command line.
629 For partitionable arrays,
631 will create the device file for the whole array and for the first 4
632 partitions. A different number of partitions can be specified at the
633 end of this option (e.g.
635 If the device name ends with a digit, the partition names add a 'p',
636 and a number, e.g. "/dev/home1p3". If there is no
637 trailing digit, then the partition names just have a number added,
638 e.g. "/dev/scratch3".
640 If the md device name is in a 'standard' format as described in DEVICE
641 NAMES, then it will be created, if necessary, with the appropriate
642 number based on that name. If the device name is not in one of these
643 formats, then a unused minor number will be allocated. The minor
644 number will be considered unused if there is no active array for that
645 number, and there is no entry in /dev for that number and with a
656 it will also create symlinks from
658 with names starting with
666 to enforce this even if it is suppressing
673 .BR \-u ", " \-\-uuid=
674 uuid of array to assemble. Devices which don't have this uuid are
678 .BR \-m ", " \-\-super\-minor=
679 Minor number of device that array was created for. Devices which
680 don't have this minor number are excluded. If you create an array as
681 /dev/md1, then all superblocks will contain the minor number 1, even if
682 the array is later assembled as /dev/md2.
684 Giving the literal word "dev" for
688 to use the minor number of the md device that is being assembled.
691 .M \-\-super\-minor=dev
692 will look for super blocks with a minor number of 0.
695 .BR \-N ", " \-\-name=
696 Specify the name of the array to assemble. This must be the name
697 that was specified when creating the array. It must either match
698 the name stored in the superblock exactly, or it must match
701 prefixed to the start of the given name.
704 .BR \-f ", " \-\-force
705 Assemble the array even if some superblocks appear out-of-date
709 Attempt to start the array even if fewer drives were given than were
710 present last time the array was active. Normally if not all the
711 expected drives are found and
713 is not used, then the array will be assembled but not started.
716 an attempt will be made to start it anyway.
720 This is the reverse of
722 in that it inhibits the startup of array unless all expected drives
723 are present. This is only needed with
725 and can be used if the physical connections to devices are
726 not as reliable as you would like.
729 .BR \-a ", " "\-\-auto{=no,yes,md,mdp,part}"
730 See this option under Create and Build options.
733 .BR \-b ", " \-\-bitmap=
734 Specify the bitmap file that was given when the array was created. If
737 bitmap, there is no need to specify this when assembling the array.
740 .BR \-\-backup\-file=
743 was used to grow the number of raid-devices in a RAID5, and the system
744 crashed during the critical section, then the same
748 to allow possibly corrupted data to be restored.
751 .BR \-U ", " \-\-update=
752 Update the superblock on each device while assembling the array. The
753 argument given to this flag can be one of
767 option will adjust the superblock of an array what was created on a Sparc
768 machine running a patched 2.2 Linux kernel. This kernel got the
769 alignment of part of the superblock wrong. You can use the
770 .B "\-\-examine \-\-sparc2.2"
773 to see what effect this would have.
777 option will update the
779 field on each superblock to match the minor number of the array being
781 This can be useful if
783 reports a different "Preferred Minor" to
785 In some cases this update will be performed automatically
786 by the kernel driver. In particular the update happens automatically
787 at the first write to an array with redundancy (RAID level 1 or
788 greater) on a 2.6 (or later) kernel.
792 option will change the uuid of the array. If a UUID is given with the
794 option that UUID will be used as a new UUID and will
796 be used to help identify the devices in the array.
799 is given, a random UUID is chosen.
803 option will change the
805 of the array as stored in the superblock. This is only supported for
806 version-1 superblocks.
810 option will change the
812 as recorded in the superblock. For version-0 superblocks, this is the
813 same as updating the UUID.
814 For version-1 superblocks, this involves updating the name.
818 option will cause the array to be marked
820 meaning that any redundancy in the array (e.g. parity for raid5,
821 copies for raid1) may be incorrect. This will cause the raid system
822 to perform a "resync" pass to make sure that all redundant information
827 option allows arrays to be moved between machines with different
829 When assembling such an array for the first time after a move, giving
830 .B "\-\-update=byteorder"
833 to expect superblocks to have their byteorder reversed, and will
834 correct that order before assembling the array. This is only valid
835 with original (Version 0.90) superblocks.
839 option will correct the summaries in the superblock. That is the
840 counts of total, working, active, failed, and spare devices.
844 will rarely be of use. It applies to version 1.1 and 1.2 metadata
845 only (where the metadata is at the start of the device) and is only
846 useful when the component device has changed size (typically become
847 larger). The version 1 metadata records the amount of the device that
848 can be used to store data, so if a device in a version 1.1 or 1.2
849 array becomes larger, the metadata will still be visible, but the
850 extra space will not. In this case it might be useful to assemble the
852 .BR \-\-update=devicesize .
855 to determine the maximum usable amount of space on each device and
856 update the relevant field in the metadata.
859 .B \-\-auto\-update\-homehost
860 This flag is only meaningful with auto-assembly (see discussion below).
861 In that situation, if no suitable arrays are found for this homehost,
863 will rescan for any arrays at all and will assemble them and update the
864 homehost to match the current host.
870 hot-add listed devices.
874 re-add a device that was recently removed from an array.
877 .BR \-r ", " \-\-remove
878 remove listed devices. They must not be active. i.e. they should
879 be failed or spare devices. As well as the name of a device file
888 The first causes all failed device to be removed. The second causes
889 any device which is no longer connected to the system (i.e an 'open'
892 to be removed. This will only succeed for devices that are spares or
893 have already been marked as failed.
896 .BR \-f ", " \-\-fail
897 mark listed devices as faulty.
898 As well as the name of a device file, the word
900 can be given. This will cause any device that has been detached from
901 the system to be marked as failed. It can then be removed.
909 Each of these options require that the first device listed is the array
910 to be acted upon, and the remainder are component devices to be added,
911 removed, or marked as faulty. Several different operations can be
912 specified for different devices, e.g.
914 mdadm /dev/md0 \-\-add /dev/sda1 \-\-fail /dev/sdb1 \-\-remove /dev/sdb1
916 Each operation applies to all devices listed until the next
919 If an array is using a write-intent bitmap, then devices which have
920 been removed can be re-added in a way that avoids a full
921 reconstruction but instead just updates the blocks that have changed
922 since the device was removed. For arrays with persistent metadata
923 (superblocks) this is done automatically. For arrays created with
925 mdadm needs to be told that this device we removed recently with
928 Devices can only be removed from an array if they are not in active
929 use, i.e. that must be spares or failed devices. To remove an active
930 device, it must first be marked as
936 .BR \-Q ", " \-\-query
937 Examine a device to see
938 (1) if it is an md device and (2) if it is a component of an md
940 Information about what is discovered is presented.
943 .BR \-D ", " \-\-detail
944 Print detail of one or more md devices.
947 .BR \-Y ", " \-\-export
950 output will be formatted as
952 pairs for easy import into the environment.
955 .BR \-E ", " \-\-examine
956 Print content of md superblock on device(s).
959 If an array was created on a 2.2 Linux kernel patched with RAID
960 support, the superblock will have been created incorrectly, or at
961 least incompatibly with 2.4 and later kernels. Using the
965 will fix the superblock before displaying it. If this appears to do
966 the right thing, then the array can be successfully assembled using
967 .BR "\-\-assemble \-\-update=sparc2.2" .
970 .BR \-X ", " \-\-examine\-bitmap
971 Report information about a bitmap file.
972 The argument is either an external bitmap file or an array component
973 in case of an internal bitmap.
977 start a partially built array.
980 .BR \-S ", " \-\-stop
981 deactivate array, releasing all resources.
984 .BR \-o ", " \-\-readonly
985 mark array as readonly.
988 .BR \-w ", " \-\-readwrite
989 mark array as readwrite.
992 .B \-\-zero\-superblock
993 If the device contains a valid md superblock, the block is
994 overwritten with zeros. With
996 the block where the superblock would be is overwritten even if it
997 doesn't appear to be valid.
1000 .BR \-t ", " \-\-test
1005 is set to reflect the status of the device.
1008 .BR \-W ", " \-\-wait
1009 For each md device given, wait for any resync, recovery, or reshape
1010 activity to finish before returning.
1012 will return with success if it actually waited for every device
1013 listed, otherwise it will return failure.
1015 .SH For Incremental Assembly mode:
1017 .BR \-\-rebuild\-map ", " \-r
1018 Rebuild the map file
1019 .RB ( /var/run/mdadm/map )
1022 uses to help track which arrays are currently being assembled.
1025 .BR \-\-run ", " \-R
1026 Run any array assembled as soon as a minimal number of devices are
1027 available, rather than waiting until all expected devices are present.
1030 .BR \-\-scan ", " \-s
1031 Only meaningful with
1035 file for arrays that are being incrementally assembled and will try to
1036 start any that are not already started. If any such array is listed
1039 as requiring an external bitmap, that bitmap will be attached first.
1041 .SH For Monitor mode:
1043 .BR \-m ", " \-\-mail
1044 Give a mail address to send alerts to.
1047 .BR \-p ", " \-\-program ", " \-\-alert
1048 Give a program to be run whenever an event is detected.
1051 .BR \-y ", " \-\-syslog
1052 Cause all events to be reported through 'syslog'. The messages have
1053 facility of 'daemon' and varying priorities.
1056 .BR \-d ", " \-\-delay
1057 Give a delay in seconds.
1059 polls the md arrays and then waits this many seconds before polling
1060 again. The default is 60 seconds.
1063 .BR \-f ", " \-\-daemonise
1066 to run as a background daemon if it decides to monitor anything. This
1067 causes it to fork and run in the child, and to disconnect form the
1068 terminal. The process id of the child is written to stdout.
1071 which will only continue monitoring if a mail address or alert program
1072 is found in the config file.
1075 .BR \-i ", " \-\-pid\-file
1078 is running in daemon mode, write the pid of the daemon process to
1079 the specified file, instead of printing it on standard output.
1082 .BR \-1 ", " \-\-oneshot
1083 Check arrays only once. This will generate
1085 events and more significantly
1091 .B " mdadm \-\-monitor \-\-scan \-1"
1093 from a cron script will ensure regular notification of any degraded arrays.
1096 .BR \-t ", " \-\-test
1099 alert for every array found at startup. This alert gets mailed and
1100 passed to the alert program. This can be used for testing that alert
1101 message do get through successfully.
1107 .B mdadm \-\-assemble
1108 .I md-device options-and-component-devices...
1111 .B mdadm \-\-assemble \-\-scan
1112 .I md-devices-and-options...
1115 .B mdadm \-\-assemble \-\-scan
1119 This usage assembles one or more raid arrays from pre-existing components.
1120 For each array, mdadm needs to know the md device, the identity of the
1121 array, and a number of component-devices. These can be found in a number of ways.
1123 In the first usage example (without the
1125 the first device given is the md device.
1126 In the second usage example, all devices listed are treated as md
1127 devices and assembly is attempted.
1128 In the third (where no devices are listed) all md devices that are
1129 listed in the configuration file are assembled.
1131 If precisely one device is listed, but
1137 was given and identity information is extracted from the configuration file.
1139 The identity can be given with the
1143 option, will be taken from the md-device record in the config file, or
1144 will be taken from the super block of the first component-device
1145 listed on the command line.
1147 Devices can be given on the
1149 command line or in the config file. Only devices which have an md
1150 superblock which contains the right identity will be considered for
1153 The config file is only used if explicitly named with
1155 or requested with (a possibly implicit)
1163 is not given, then the config file will only be used to find the
1164 identity of md arrays.
1166 Normally the array will be started after it is assembled. However if
1168 is not given and insufficient drives were listed to start a complete
1169 (non-degraded) array, then the array is not started (to guard against
1170 usage errors). To insist that the array be started in this case (as
1171 may work for RAID1, 4, 5, 6, or 10), give the
1175 If the md device does not exist, then it will be created providing the
1176 intent is clear. i.e. the name must be in a standard form, or the
1178 option must be given to clarify how and whether the device should be
1180 This can be useful for handling partitioned devices (which don't have
1181 a stable device number \(em it can change after a reboot) and when using
1182 "udev" to manage your
1184 tree (udev cannot handle md devices because of the unusual device
1185 initialisation conventions).
1187 If the option to "auto" is "mdp" or "part" or (on the command line
1188 only) "p", then mdadm will create a partitionable array, using the
1189 first free one that is not in use and does not already have an entry
1190 in /dev (apart from numeric /dev/md* entries).
1192 If the option to "auto" is "yes" or "md" or (on the command line)
1193 nothing, then mdadm will create a traditional, non-partitionable md
1196 It is expected that the "auto" functionality will be used to create
1197 device entries with meaningful names such as "/dev/md/home" or
1198 "/dev/md/root", rather than names based on the numerical array number.
1200 When using option "auto" to create a partitionable array, the device
1201 files for the first 4 partitions are also created. If a different
1202 number is required it can be simply appended to the auto option.
1203 e.g. "auto=part8". Partition names are created by appending a digit
1204 string to the device name, with an intervening "p" if the device name
1209 option is also available in Build and Create modes. As those modes do
1210 not use a config file, the "auto=" config option does not apply to
1218 and no devices are listed,
1220 will first attempt to assemble all the arrays listed in the config
1225 has been specified (either in the config file or on the command line),
1227 will look further for possible arrays and will try to assemble
1228 anything that it finds which is tagged as belonging to the given
1229 homehost. This is the only situation where
1231 will assemble arrays without being given specific device name or
1232 identity information for the array.
1236 finds a consistent set of devices that look like they should comprise
1237 an array, and if the superblock is tagged as belonging to the given
1238 home host, it will automatically choose a device name and try to
1239 assemble the array. If the array uses version-0.90 metadata, then the
1241 number as recorded in the superblock is used to create a name in
1245 If the array uses version-1 metadata, then the
1247 from the superblock is used to similarly create a name in
1249 (the name will have any 'host' prefix stripped first).
1253 cannot find any array for the given host at all, and if
1254 .B \-\-auto\-update\-homehost
1257 will search again for any array (not just an array created for this
1258 host) and will assemble each assuming
1259 .BR \-\-update=homehost .
1260 This will change the host tag in the superblock so that on the next run,
1261 these arrays will be found without the second pass. The intention of
1262 this feature is to support transitioning a set of md arrays to using
1265 The reason for requiring arrays to be tagged with the homehost for
1266 auto assembly is to guard against problems that can arise when moving
1267 devices from one host to another.
1277 .BI \-\-raid\-devices= Z
1281 This usage is similar to
1283 The difference is that it creates an array without a superblock. With
1284 these arrays there is no difference between initially creating the array and
1285 subsequently assembling the array, except that hopefully there is useful
1286 data there in the second case.
1288 The level may raid0, linear, multipath, or faulty, or one of their
1289 synonyms. All devices must be listed and the array will be started
1301 .BI \-\-raid\-devices= Z
1305 This usage will initialise a new md array, associate some devices with
1306 it, and activate the array.
1310 option is given (as described in more detail in the section on
1311 Assemble mode), then the md device will be created with a suitable
1312 device number if necessary.
1314 As devices are added, they are checked to see if they contain raid
1315 superblocks or filesystems. They are also checked to see if the variance in
1316 device size exceeds 1%.
1318 If any discrepancy is found, the array will not automatically be run, though
1321 can override this caution.
1323 To create a "degraded" array in which some devices are missing, simply
1324 give the word "\fBmissing\fP"
1325 in place of a device name. This will cause
1327 to leave the corresponding slot in the array empty.
1328 For a RAID4 or RAID5 array at most one slot can be
1329 "\fBmissing\fP"; for a RAID6 array at most two slots.
1330 For a RAID1 array, only one real device needs to be given. All of the
1334 When creating a RAID5 array,
1336 will automatically create a degraded array with an extra spare drive.
1337 This is because building the spare into a degraded array is in general faster than resyncing
1338 the parity on a non-degraded, but not clean, array. This feature can
1339 be overridden with the
1343 When creating an array with version-1 metadata a name for the host is
1345 If this is not given with the
1349 will chose a name based on the last component of the name of the
1350 device being created. So if
1352 is being created, then the name
1357 is being created, then the name
1361 A new array will normally get a randomly assigned 128bit UUID which is
1362 very likely to be unique. If you have a specific need, you can choose
1363 a UUID for the array by giving the
1365 option. Be warned that creating two arrays with the same UUID is a
1366 recipe for disaster. Also, using
1368 when creating a v0.90 array will silently override any
1373 '''option is given, it is not necessary to list any component-devices in this command.
1374 '''They can be added later, before a
1378 '''is given, the apparent size of the smallest drive given is used.
1380 The General Management options that are valid with
1385 insist on running the array even if some devices look like they might
1390 start the array readonly \(em not supported yet.
1398 .I options... devices...
1401 This usage will allow individual devices in an array to be failed,
1402 removed or added. It is possible to perform multiple operations with
1403 on command. For example:
1405 .B " mdadm /dev/md0 \-f /dev/hda1 \-r /dev/hda1 \-a /dev/hda1"
1411 and will then remove it from the array and finally add it back
1412 in as a spare. However only one md array can be affected by a single
1423 MISC mode includes a number of distinct operations that
1424 operate on distinct devices. The operations are:
1427 The device is examined to see if it is
1428 (1) an active md array, or
1429 (2) a component of an md array.
1430 The information discovered is reported.
1434 The device should be an active md device.
1436 will display a detailed description of the array.
1440 will cause the output to be less detailed and the format to be
1441 suitable for inclusion in
1442 .BR /etc/mdadm.conf .
1445 will normally be 0 unless
1447 failed to get useful information about the device(s); however, if the
1449 option is given, then the exit status will be:
1453 The array is functioning normally.
1456 The array has at least one failed device.
1459 The array has multiple failed devices such that it is unusable.
1462 There was an error while trying to get information about the device.
1467 The device should be a component of an md array.
1469 will read the md superblock of the device and display the contents.
1474 is given, then multiple devices that are components of the one array
1475 are grouped together and reported in a single entry suitable
1477 .BR /etc/mdadm.conf .
1481 without listing any devices will cause all devices listed in the
1482 config file to be examined.
1486 The devices should be active md arrays which will be deactivated, as
1487 long as they are not currently in use.
1491 This will fully activate a partially assembled md array.
1495 This will mark an active array as read-only, providing that it is
1496 not currently being used.
1502 array back to being read/write.
1506 For all operations except
1509 will cause the operation to be applied to all arrays listed in
1514 causes all devices listed in the config file to be examined.
1521 .B mdadm \-\-monitor
1522 .I options... devices...
1527 to periodically poll a number of md arrays and to report on any events
1530 will never exit once it decides that there are arrays to be checked,
1531 so it should normally be run in the background.
1533 As well as reporting events,
1535 may move a spare drive from one array to another if they are in the
1538 and if the destination array has a failed drive but no spares.
1540 If any devices are listed on the command line,
1542 will only monitor those devices. Otherwise all arrays listed in the
1543 configuration file will be monitored. Further, if
1545 is given, then any other md devices that appear in
1547 will also be monitored.
1549 The result of monitoring the arrays is the generation of events.
1550 These events are passed to a separate program (if specified) and may
1551 be mailed to a given E-mail address.
1553 When passing events to a program, the program is run once for each event,
1554 and is given 2 or 3 command-line arguments: the first is the
1555 name of the event (see below), the second is the name of the
1556 md device which is affected, and the third is the name of a related
1557 device if relevant (such as a component device that has failed).
1561 is given, then a program or an E-mail address must be specified on the
1562 command line or in the config file. If neither are available, then
1564 will not monitor anything.
1568 will continue monitoring as long as something was found to monitor. If
1569 no program or email is given, then each event is reported to
1572 The different events are:
1576 .B DeviceDisappeared
1577 An md array which previously was configured appears to no longer be
1578 configured. (syslog priority: Critical)
1582 was told to monitor an array which is RAID0 or Linear, then it will
1584 .B DeviceDisappeared
1585 with the extra information
1587 This is because RAID0 and Linear do not support the device-failed,
1588 hot-spare and resync operations which are monitored.
1592 An md array started reconstruction. (syslog priority: Warning)
1598 is 20, 40, 60, or 80, this indicates that rebuild has passed that many
1599 percentage of the total. (syslog priority: Warning)
1603 An md array that was rebuilding, isn't any more, either because it
1604 finished normally or was aborted. (syslog priority: Warning)
1608 An active component device of an array has been marked as
1609 faulty. (syslog priority: Critical)
1613 A spare component device which was being rebuilt to replace a faulty
1614 device has failed. (syslog priority: Critical)
1618 A spare component device which was being rebuilt to replace a faulty
1619 device has been successfully rebuilt and has been made active.
1620 (syslog priority: Info)
1624 A new md array has been detected in the
1626 file. (syslog priority: Info)
1630 A newly noticed array appears to be degraded. This message is not
1633 notices a drive failure which causes degradation, but only when
1635 notices that an array is degraded when it first sees the array.
1636 (syslog priority: Critical)
1640 A spare drive has been moved from one array in a
1642 to another to allow a failed drive to be replaced.
1643 (syslog priority: Info)
1649 has been told, via the config file, that an array should have a certain
1650 number of spare devices, and
1652 detects that it has fewer than this number when it first sees the
1653 array, it will report a
1656 (syslog priority: Warning)
1660 An array was found at startup, and the
1663 (syslog priority: Info)
1673 cause Email to be sent. All events cause the program to be run.
1674 The program is run with two or three arguments: the event
1675 name, the array device and possibly a second device.
1677 Each event has an associated array device (e.g.
1679 and possibly a second device. For
1684 the second device is the relevant component device.
1687 the second device is the array that the spare was moved from.
1691 to move spares from one array to another, the different arrays need to
1692 be labeled with the same
1694 in the configuration file. The
1696 name can be any string; it is only necessary that different spare
1697 groups use different names.
1701 detects that an array in a spare group has fewer active
1702 devices than necessary for the complete array, and has no spare
1703 devices, it will look for another array in the same spare group that
1704 has a full complement of working drive and a spare. It will then
1705 attempt to remove the spare from the second drive and add it to the
1707 If the removal succeeds but the adding fails, then it is added back to
1711 The GROW mode is used for changing the size or shape of an active
1713 For this to work, the kernel must support the necessary change.
1714 Various types of growth are being added during 2.6 development,
1715 including restructuring a raid5 array to have more active devices.
1717 Currently the only support available is to
1719 change the "size" attribute
1720 for RAID1, RAID5 and RAID6.
1722 increase the "raid-disks" attribute of RAID1, RAID5, and RAID6.
1724 add a write-intent bitmap to any array which supports these bitmaps, or
1725 remove a write-intent bitmap from such an array.
1729 Normally when an array is built the "size" it taken from the smallest
1730 of the drives. If all the small drives in an arrays are, one at a
1731 time, removed and replaced with larger drives, then you could have an
1732 array of large drives with only a small amount used. In this
1733 situation, changing the "size" with "GROW" mode will allow the extra
1734 space to start being used. If the size is increased in this way, a
1735 "resync" process will start to make sure the new parts of the array
1738 Note that when an array changes size, any filesystem that may be
1739 stored in the array will not automatically grow to use the space. The
1740 filesystem will need to be explicitly told to use the extra space.
1742 .SS RAID-DEVICES CHANGES
1744 A RAID1 array can work with any number of devices from 1 upwards
1745 (though 1 is not very useful). There may be times which you want to
1746 increase or decrease the number of active devices. Note that this is
1747 different to hot-add or hot-remove which changes the number of
1750 When reducing the number of devices in a RAID1 array, the slots which
1751 are to be removed from the array must already be vacant. That is, the
1752 devices which were in those slots must be failed and removed.
1754 When the number of devices is increased, any hot spares that are
1755 present will be activated immediately.
1757 Increasing the number of active devices in a RAID5 is much more
1758 effort. Every block in the array will need to be read and written
1759 back to a new location. From 2.6.17, the Linux Kernel is able to do
1760 this safely, including restart and interrupted "reshape".
1762 When relocating the first few stripes on a raid5, it is not possible
1763 to keep the data on disk completely consistent and crash-proof. To
1764 provide the required safety, mdadm disables writes to the array while
1765 this "critical section" is reshaped, and takes a backup of the data
1766 that is in that section. This backup is normally stored in any spare
1767 devices that the array has, however it can also be stored in a
1768 separate file specified with the
1770 option. If this option is used, and the system does crash during the
1771 critical period, the same file must be passed to
1773 to restore the backup and reassemble the array.
1777 A write-intent bitmap can be added to, or removed from, an active
1778 array. Either internal bitmaps, or bitmaps stored in a separate file,
1779 can be added. Note that if you add a bitmap stored in a file which is
1780 in a filesystem that is on the raid array being affected, the system
1781 will deadlock. The bitmap must be on a separate filesystem.
1783 .SH INCREMENTAL MODE
1787 .B mdadm \-\-incremental
1793 .B mdadm \-\-incremental \-\-rebuild
1796 .B mdadm \-\-incremental \-\-run \-\-scan
1800 This mode is designed to be used in conjunction with a device
1801 discovery system. As devices are found in a system, they can be
1803 .B "mdadm \-\-incremental"
1804 to be conditionally added to an appropriate array.
1807 performs a number of tests to determine if the device is part of an
1808 array, and which array it should be part of. If an appropriate array
1809 is found, or can be created,
1811 adds the device to the array and conditionally starts the array.
1815 will only add devices to an array which were previously working
1816 (active or spare) parts of that array. It does not currently support
1817 automatic inclusion of a new drive as a spare in some array.
1819 .B "mdadm \-\-incremental"
1820 requires a bug-fix in all kernels through 2.6.19.
1821 Hopefully, this will be fixed in 2.6.20; alternately, apply the patch
1822 which is included with the mdadm source distribution. If
1824 detects that this bug is present, it will abort any attempt to use
1825 .BR \-\-incremental .
1829 makes are as follow:
1831 Is the device permitted by
1833 That is, is it listed in a
1835 line in that file. If
1837 is absent then the default it to allow any device. Similar if
1839 contains the special word
1841 then any device is allowed. Otherwise the device name given to
1843 must match one of the names or patterns in a
1848 Does the device have a valid md superblock. If a specific metadata
1849 version is request with
1853 then only that style of metadata is accepted, otherwise
1855 finds any known version of metadata. If no
1857 metadata is found, the device is rejected.
1860 Does the metadata match an expected array?
1861 The metadata can match in two ways. Either there is an array listed
1864 which identifies the array (either by UUID, by name, by device list,
1865 or by minor-number), or the array was created with a
1871 or on the command line.
1874 is not able to positively identify the array as belonging to the
1875 current host, the device will be rejected.
1879 keeps a list of arrays that it has partially assembled in
1880 .B /var/run/mdadm/map
1882 .B /var/run/mdadm.map
1883 if the directory doesn't exist). If no array exists which matches
1884 the metadata on the new device,
1886 must choose a device name and unit number. It does this based on any
1889 or any name information stored in the metadata. If this name
1890 suggests a unit number, that number will be used, otherwise a free
1891 unit number will be chosen. Normally
1893 will prefer to create a partitionable array, however if the
1897 suggests that a non-partitionable array is preferred, that will be
1901 Once an appropriate array is found or created and the device is added,
1903 must decide if the array is ready to be started. It will
1904 normally compare the number of available (non-spare) devices to the
1905 number of devices that the metadata suggests need to be active. If
1906 there are at least that many, the array will be started. This means
1907 that if any devices are missing the array will not be restarted.
1913 in which case the array will be run as soon as there are enough
1914 devices present for the data to be accessible. For a raid1, that
1915 means one device will start the array. For a clean raid5, the array
1916 will be started as soon as all but one drive is present.
1918 Note that neither of these approaches is really ideal. If it can
1919 be known that all device discovery has completed, then
1923 can be run which will try to start all arrays that are being
1924 incrementally assembled. They are started in "read-auto" mode in
1925 which they are read-only until the first write request. This means
1926 that no metadata updates are made and no attempt at resync or recovery
1927 happens. Further devices that are found before the first write can
1928 still be added safely.
1932 .B " mdadm \-\-query /dev/name-of-device"
1934 This will find out if a given device is a raid array, or is part of
1935 one, and will provide brief information about the device.
1937 .B " mdadm \-\-assemble \-\-scan"
1939 This will assemble and start all arrays listed in the standard config
1940 file. This command will typically go in a system startup file.
1942 .B " mdadm \-\-stop \-\-scan"
1944 This will shut down all arrays that can be shut down (i.e. are not
1945 currently in use). This will typically go in a system shutdown script.
1947 .B " mdadm \-\-follow \-\-scan \-\-delay=120"
1949 If (and only if) there is an Email address or program given in the
1950 standard config file, then
1951 monitor the status of all arrays listed in that file by
1952 polling them ever 2 minutes.
1954 .B " mdadm \-\-create /dev/md0 \-\-level=1 \-\-raid\-devices=2 /dev/hd[ac]1"
1956 Create /dev/md0 as a RAID1 array consisting of /dev/hda1 and /dev/hdc1.
1959 .B " echo 'DEVICE /dev/hd*[0\-9] /dev/sd*[0\-9]' > mdadm.conf"
1961 .B " mdadm \-\-detail \-\-scan >> mdadm.conf"
1963 This will create a prototype config file that describes currently
1964 active arrays that are known to be made from partitions of IDE or SCSI drives.
1965 This file should be reviewed before being used as it may
1966 contain unwanted detail.
1968 .B " echo 'DEVICE /dev/hd[a\-z] /dev/sd*[a\-z]' > mdadm.conf"
1970 .B " mdadm \-\-examine \-\-scan \-\-config=mdadm.conf >> mdadm.conf"
1972 This will find arrays which could be assembled from existing IDE and
1973 SCSI whole drives (not partitions), and store the information in the
1974 format of a config file.
1975 This file is very likely to contain unwanted detail, particularly
1978 entries. It should be reviewed and edited before being used as an
1981 .B " mdadm \-\-examine \-\-brief \-\-scan \-\-config=partitions"
1983 .B " mdadm \-Ebsc partitions"
1985 Create a list of devices by reading
1986 .BR /proc/partitions ,
1987 scan these for RAID superblocks, and printout a brief listing of all
1990 .B " mdadm \-Ac partitions \-m 0 /dev/md0"
1992 Scan all partitions and devices listed in
1993 .BR /proc/partitions
1996 out of all such devices with a RAID superblock with a minor number of 0.
1998 .B " mdadm \-\-monitor \-\-scan \-\-daemonise > /var/run/mdadm"
2000 If config file contains a mail address or alert program, run mdadm in
2001 the background in monitor mode monitoring all md devices. Also write
2002 pid of mdadm daemon to
2003 .BR /var/run/mdadm .
2005 .B " mdadm \-Iq /dev/somedevice"
2007 Try to incorporate newly discovered device into some array as
2010 .B " mdadm \-\-incremental \-\-rebuild \-\-run \-\-scan"
2012 Rebuild the array map from any current arrays, and then start any that
2015 .B " mdadm /dev/md4 --fail detached --remove detached"
2017 Any devices which are components of /dev/md4 will be marked as faulty
2018 and then remove from the array.
2020 .B " mdadm \-\-create \-\-help"
2022 Provide help about the Create mode.
2024 .B " mdadm \-\-config \-\-help"
2026 Provide help about the format of the config file.
2028 .B " mdadm \-\-help"
2030 Provide general help.
2041 lists all active md devices with information about them.
2043 uses this to find arrays when
2045 is given in Misc mode, and to monitor array reconstruction
2051 The config file lists which devices may be scanned to see if
2052 they contain MD super block, and gives identifying information
2053 (e.g. UUID) about known MD arrays. See
2057 .SS /var/run/mdadm/map
2060 mode is used, this file gets a list of arrays currently being created.
2063 does not exist as a directory, then
2064 .B /var/run/mdadm.map
2069 While entries in the /dev directory can have any format you like,
2071 has an understanding of 'standard' formats which it uses to guide its
2072 behaviour when creating device files via the
2076 The standard names for non-partitioned arrays (the only sort of md
2077 array available in 2.4 and earlier) are either of
2083 where NN is a number.
2084 The standard names for partitionable arrays (as available from 2.6
2085 onwards) are either of
2091 Partition numbers should be indicated by added "pMM" to these, thus "/dev/md/d1p2".
2095 was previously known as
2099 is completely separate from the
2101 package, and does not use the
2103 configuration file at all.
2106 For information on the various levels of
2110 .UR http://ostenfeld.dk/~jakob/Software\-RAID.HOWTO/
2111 http://ostenfeld.dk/~jakob/Software\-RAID.HOWTO/
2114 '''for new releases of the RAID driver check out:
2117 '''.UR ftp://ftp.kernel.org/pub/linux/kernel/people/mingo/raid-patches
2118 '''ftp://ftp.kernel.org/pub/linux/kernel/people/mingo/raid-patches
2123 '''.UR http://www.cse.unsw.edu.au/~neilb/patches/linux-stable/
2124 '''http://www.cse.unsw.edu.au/~neilb/patches/linux-stable/
2127 The latest version of
2129 should always be available from
2131 .UR http://www.kernel.org/pub/linux/utils/raid/mdadm/
2132 http://www.kernel.org/pub/linux/utils/raid/mdadm/