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.
8 .TH MDADM 8 "" v4.1-rc2
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
46 is not a Software RAID mechanism, but does involve
48 each device is a path to one common physical storage device.
49 New installations should not use md/multipath as it is not well
50 supported and has no ongoing development. Use the Device Mapper based
51 multipath-tools instead.
54 is also not true RAID, and it only involves one device. It
55 provides a layer over a true device that can be used to inject faults.
60 is a collection of devices that are
61 managed as a set. This is similar to the set of devices connected to
62 a hardware RAID controller. The set of devices may contain a number
63 of different RAID arrays each utilising some (or all) of the blocks from a
64 number of the devices in the set. For example, two devices in a 5-device set
65 might form a RAID1 using the whole devices. The remaining three might
66 have a RAID5 over the first half of each device, and a RAID0 over the
71 there is one set of metadata that describes all of
72 the arrays in the container. So when
76 device, the device just represents the metadata. Other normal arrays (RAID1
77 etc) can be created inside the container.
80 mdadm has several major modes of operation:
83 Assemble the components of a previously created
84 array into an active array. Components can be explicitly given
85 or can be searched for.
87 checks that the components
88 do form a bona fide array, and can, on request, fiddle superblock
89 information so as to assemble a faulty array.
93 Build an array that doesn't have per-device metadata (superblocks). For these
96 cannot differentiate between initial creation and subsequent assembly
97 of an array. It also cannot perform any checks that appropriate
98 components have been requested. Because of this, the
100 mode should only be used together with a complete understanding of
105 Create a new array with per-device metadata (superblocks).
106 Appropriate metadata is written to each device, and then the array
107 comprising those devices is activated. A 'resync' process is started
108 to make sure that the array is consistent (e.g. both sides of a mirror
109 contain the same data) but the content of the device is left otherwise
111 The array can be used as soon as it has been created. There is no
112 need to wait for the initial resync to finish.
115 .B "Follow or Monitor"
116 Monitor one or more md devices and act on any state changes. This is
117 only meaningful for RAID1, 4, 5, 6, 10 or multipath arrays, as
118 only these have interesting state. RAID0 or Linear never have
119 missing, spare, or failed drives, so there is nothing to monitor.
123 Grow (or shrink) an array, or otherwise reshape it in some way.
124 Currently supported growth options including changing the active size
125 of component devices and changing the number of active devices in
126 Linear and RAID levels 0/1/4/5/6,
127 changing the RAID level between 0, 1, 5, and 6, and between 0 and 10,
128 changing the chunk size and layout for RAID 0,4,5,6,10 as well as adding or
129 removing a write-intent bitmap and changing the array's consistency policy.
132 .B "Incremental Assembly"
133 Add a single device to an appropriate array. If the addition of the
134 device makes the array runnable, the array will be started.
135 This provides a convenient interface to a
137 system. As each device is detected,
139 has a chance to include it in some array as appropriate.
142 flag is passed in we will remove the device from any active array
143 instead of adding it.
149 in this mode, then any arrays within that container will be assembled
154 This is for doing things to specific components of an array such as
155 adding new spares and removing faulty devices.
159 This is an 'everything else' mode that supports operations on active
160 arrays, operations on component devices such as erasing old superblocks, and
161 information gathering operations.
162 .\"This mode allows operations on independent devices such as examine MD
163 .\"superblocks, erasing old superblocks and stopping active arrays.
167 This mode does not act on a specific device or array, but rather it
168 requests the Linux Kernel to activate any auto-detected arrays.
171 .SH Options for selecting a mode are:
174 .BR \-A ", " \-\-assemble
175 Assemble a pre-existing array.
178 .BR \-B ", " \-\-build
179 Build a legacy array without superblocks.
182 .BR \-C ", " \-\-create
186 .BR \-F ", " \-\-follow ", " \-\-monitor
192 .BR \-G ", " \-\-grow
193 Change the size or shape of an active array.
196 .BR \-I ", " \-\-incremental
197 Add/remove a single device to/from an appropriate array, and possibly start the array.
201 Request that the kernel starts any auto-detected arrays. This can only
204 is compiled into the kernel \(em not if it is a module.
205 Arrays can be auto-detected by the kernel if all the components are in
206 primary MS-DOS partitions with partition type
208 and all use v0.90 metadata.
209 In-kernel autodetect is not recommended for new installations. Using
211 to detect and assemble arrays \(em possibly in an
213 \(em is substantially more flexible and should be preferred.
216 If a device is given before any options, or if the first option is
225 then the MANAGE mode is assumed.
226 Anything other than these will cause the
230 .SH Options that are not mode-specific are:
233 .BR \-h ", " \-\-help
234 Display general help message or, after one of the above options, a
235 mode-specific help message.
239 Display more detailed help about command line parsing and some commonly
243 .BR \-V ", " \-\-version
244 Print version information for mdadm.
247 .BR \-v ", " \-\-verbose
248 Be more verbose about what is happening. This can be used twice to be
250 The extra verbosity currently only affects
251 .B \-\-detail \-\-scan
253 .BR "\-\-examine \-\-scan" .
256 .BR \-q ", " \-\-quiet
257 Avoid printing purely informative messages. With this,
259 will be silent unless there is something really important to report.
263 .BR \-f ", " \-\-force
264 Be more forceful about certain operations. See the various modes for
265 the exact meaning of this option in different contexts.
268 .BR \-c ", " \-\-config=
269 Specify the config file or directory. Default is to use
272 .BR /etc/mdadm.conf.d ,
273 or if those are missing then
274 .B /etc/mdadm/mdadm.conf
276 .BR /etc/mdadm/mdadm.conf.d .
277 If the config file given is
279 then nothing will be read, but
281 will act as though the config file contained exactly
283 .B " DEVICE partitions containers"
287 to find a list of devices to scan, and
289 to find a list of containers to examine.
292 is given for the config file, then
294 will act as though the config file were empty.
296 If the name given is of a directory, then
298 will collect all the files contained in the directory with a name ending
301 sort them lexically, and process all of those files as config files.
304 .BR \-s ", " \-\-scan
307 for missing information.
308 In general, this option gives
310 permission to get any missing information (like component devices,
311 array devices, array identities, and alert destination) from the
312 configuration file (see previous option);
313 one exception is MISC mode when using
319 says to get a list of array devices from
323 .BR \-e ", " \-\-metadata=
324 Declare the style of RAID metadata (superblock) to be used. The
325 default is {DEFAULT_METADATA} for
327 and to guess for other operations.
328 The default can be overridden by setting the
337 .ie '{DEFAULT_METADATA}'0.90'
338 .IP "0, 0.90, default"
341 Use the original 0.90 format superblock. This format limits arrays to
342 28 component devices and limits component devices of levels 1 and
343 greater to 2 terabytes. It is also possible for there to be confusion
344 about whether the superblock applies to a whole device or just the
345 last partition, if that partition starts on a 64K boundary.
346 .ie '{DEFAULT_METADATA}'0.90'
347 .IP "1, 1.0, 1.1, 1.2"
349 .IP "1, 1.0, 1.1, 1.2 default"
350 Use the new version-1 format superblock. This has fewer restrictions.
351 It can easily be moved between hosts with different endian-ness, and a
352 recovery operation can be checkpointed and restarted. The different
353 sub-versions store the superblock at different locations on the
354 device, either at the end (for 1.0), at the start (for 1.1) or 4K from
355 the start (for 1.2). "1" is equivalent to "1.2" (the commonly
356 preferred 1.x format).
357 'if '{DEFAULT_METADATA}'1.2' "default" is equivalent to "1.2".
359 Use the "Industry Standard" DDF (Disk Data Format) format defined by
361 When creating a DDF array a
363 will be created, and normal arrays can be created in that container.
365 Use the Intel(R) Matrix Storage Manager metadata format. This creates a
367 which is managed in a similar manner to DDF, and is supported by an
368 option-rom on some platforms:
370 .B http://www.intel.com/design/chipsets/matrixstorage_sb.htm
376 This will override any
378 setting in the config file and provides the identity of the host which
379 should be considered the home for any arrays.
381 When creating an array, the
383 will be recorded in the metadata. For version-1 superblocks, it will
384 be prefixed to the array name. For version-0.90 superblocks, part of
385 the SHA1 hash of the hostname will be stored in the later half of the
388 When reporting information about an array, any array which is tagged
389 for the given homehost will be reported as such.
391 When using Auto-Assemble, only arrays tagged for the given homehost
392 will be allowed to use 'local' names (i.e. not ending in '_' followed
393 by a digit string). See below under
394 .BR "Auto Assembly" .
396 The special name "\fBany\fP" can be used as a wild card. If an array
399 then the name "\fBany\fP" will be stored in the array and it can be
400 assembled in the same way on any host. If an array is assembled with
401 this option, then the homehost recorded on the array will be ignored.
407 needs to print the name for a device it normally finds the name in
409 which refers to the device and is shortest. When a path component is
413 will prefer a longer name if it contains that component. For example
414 .B \-\-prefer=by-uuid
415 will prefer a name in a subdirectory of
420 This functionality is currently only provided by
426 .B \-\-home\-cluster=
427 specifies the cluster name for the md device. The md device can be assembled
428 only on the cluster which matches the name specified. If this option is not
429 provided, mdadm tries to detect the cluster name automatically.
431 .SH For create, build, or grow:
434 .BR \-n ", " \-\-raid\-devices=
435 Specify the number of active devices in the array. This, plus the
436 number of spare devices (see below) must equal the number of
438 (including "\fBmissing\fP" devices)
439 that are listed on the command line for
441 Setting a value of 1 is probably
442 a mistake and so requires that
444 be specified first. A value of 1 will then be allowed for linear,
445 multipath, RAID0 and RAID1. It is never allowed for RAID4, RAID5 or RAID6.
447 This number can only be changed using
449 for RAID1, RAID4, RAID5 and RAID6 arrays, and only on kernels which provide
450 the necessary support.
453 .BR \-x ", " \-\-spare\-devices=
454 Specify the number of spare (eXtra) devices in the initial array.
455 Spares can also be added
456 and removed later. The number of component devices listed
457 on the command line must equal the number of RAID devices plus the
458 number of spare devices.
461 .BR \-z ", " \-\-size=
462 Amount (in Kilobytes) of space to use from each drive in RAID levels 1/4/5/6.
463 This must be a multiple of the chunk size, and must leave about 128Kb
464 of space at the end of the drive for the RAID superblock.
465 If this is not specified
466 (as it normally is not) the smallest drive (or partition) sets the
467 size, though if there is a variance among the drives of greater than 1%, a warning is
470 A suffix of 'K', 'M' or 'G' can be given to indicate Kilobytes, Megabytes or
471 Gigabytes respectively.
473 Sometimes a replacement drive can be a little smaller than the
474 original drives though this should be minimised by IDEMA standards.
475 Such a replacement drive will be rejected by
477 To guard against this it can be useful to set the initial size
478 slightly smaller than the smaller device with the aim that it will
479 still be larger than any replacement.
481 This value can be set with
483 for RAID level 1/4/5/6 though
485 based arrays such as those with IMSM metadata may not be able to
487 If the array was created with a size smaller than the currently
488 active drives, the extra space can be accessed using
490 The size can be given as
492 which means to choose the largest size that fits on all current drives.
494 Before reducing the size of the array (with
495 .BR "\-\-grow \-\-size=" )
496 you should make sure that space isn't needed. If the device holds a
497 filesystem, you would need to resize the filesystem to use less space.
499 After reducing the array size you should check that the data stored in
500 the device is still available. If the device holds a filesystem, then
501 an 'fsck' of the filesystem is a minimum requirement. If there are
502 problems the array can be made bigger again with no loss with another
503 .B "\-\-grow \-\-size="
506 This value cannot be used when creating a
508 such as with DDF and IMSM metadata, though it perfectly valid when
509 creating an array inside a container.
512 .BR \-Z ", " \-\-array\-size=
513 This is only meaningful with
515 and its effect is not persistent: when the array is stopped and
516 restarted the default array size will be restored.
518 Setting the array-size causes the array to appear smaller to programs
519 that access the data. This is particularly needed before reshaping an
520 array so that it will be smaller. As the reshape is not reversible,
521 but setting the size with
523 is, it is required that the array size is reduced as appropriate
524 before the number of devices in the array is reduced.
526 Before reducing the size of the array you should make sure that space
527 isn't needed. If the device holds a filesystem, you would need to
528 resize the filesystem to use less space.
530 After reducing the array size you should check that the data stored in
531 the device is still available. If the device holds a filesystem, then
532 an 'fsck' of the filesystem is a minimum requirement. If there are
533 problems the array can be made bigger again with no loss with another
534 .B "\-\-grow \-\-array\-size="
537 A suffix of 'K', 'M' or 'G' can be given to indicate Kilobytes, Megabytes or
538 Gigabytes respectively.
541 restores the apparent size of the array to be whatever the real
542 amount of available space is.
544 Clustered arrays do not support this parameter yet.
547 .BR \-c ", " \-\-chunk=
548 Specify chunk size of kilobytes. The default when creating an
549 array is 512KB. To ensure compatibility with earlier versions, the
550 default when building an array with no persistent metadata is 64KB.
551 This is only meaningful for RAID0, RAID4, RAID5, RAID6, and RAID10.
553 RAID4, RAID5, RAID6, and RAID10 require the chunk size to be a power
554 of 2. In any case it must be a multiple of 4KB.
556 A suffix of 'K', 'M' or 'G' can be given to indicate Kilobytes, Megabytes or
557 Gigabytes respectively.
561 Specify rounding factor for a Linear array. The size of each
562 component will be rounded down to a multiple of this size.
563 This is a synonym for
565 but highlights the different meaning for Linear as compared to other
566 RAID levels. The default is 64K if a kernel earlier than 2.6.16 is in
567 use, and is 0K (i.e. no rounding) in later kernels.
570 .BR \-l ", " \-\-level=
571 Set RAID level. When used with
573 options are: linear, raid0, 0, stripe, raid1, 1, mirror, raid4, 4,
574 raid5, 5, raid6, 6, raid10, 10, multipath, mp, faulty, container.
575 Obviously some of these are synonymous.
579 metadata type is requested, only the
581 level is permitted, and it does not need to be explicitly given.
585 only linear, stripe, raid0, 0, raid1, multipath, mp, and faulty are valid.
589 to change the RAID level in some cases. See LEVEL CHANGES below.
592 .BR \-p ", " \-\-layout=
593 This option configures the fine details of data layout for RAID5, RAID6,
594 and RAID10 arrays, and controls the failure modes for
596 It can also be used for working around a kernel bug with RAID0, but generally
597 doesn't need to be used explicitly.
599 The layout of the RAID5 parity block can be one of
600 .BR left\-asymmetric ,
601 .BR left\-symmetric ,
602 .BR right\-asymmetric ,
603 .BR right\-symmetric ,
604 .BR la ", " ra ", " ls ", " rs .
606 .BR left\-symmetric .
608 It is also possible to cause RAID5 to use a RAID4-like layout by
614 Finally for RAID5 there are DDF\-compatible layouts,
615 .BR ddf\-zero\-restart ,
616 .BR ddf\-N\-restart ,
618 .BR ddf\-N\-continue .
620 These same layouts are available for RAID6. There are also 4 layouts
621 that will provide an intermediate stage for converting between RAID5
622 and RAID6. These provide a layout which is identical to the
623 corresponding RAID5 layout on the first N\-1 devices, and has the 'Q'
624 syndrome (the second 'parity' block used by RAID6) on the last device.
626 .BR left\-symmetric\-6 ,
627 .BR right\-symmetric\-6 ,
628 .BR left\-asymmetric\-6 ,
629 .BR right\-asymmetric\-6 ,
631 .BR parity\-first\-6 .
633 When setting the failure mode for level
636 .BR write\-transient ", " wt ,
637 .BR read\-transient ", " rt ,
638 .BR write\-persistent ", " wp ,
639 .BR read\-persistent ", " rp ,
641 .BR read\-fixable ", " rf ,
642 .BR clear ", " flush ", " none .
644 Each failure mode can be followed by a number, which is used as a period
645 between fault generation. Without a number, the fault is generated
646 once on the first relevant request. With a number, the fault will be
647 generated after that many requests, and will continue to be generated
648 every time the period elapses.
650 Multiple failure modes can be current simultaneously by using the
652 option to set subsequent failure modes.
654 "clear" or "none" will remove any pending or periodic failure modes,
655 and "flush" will clear any persistent faults.
657 The layout options for RAID10 are one of 'n', 'o' or 'f' followed
658 by a small number. The default is 'n2'. The supported options are:
661 signals 'near' copies. Multiple copies of one data block are at
662 similar offsets in different devices.
665 signals 'offset' copies. Rather than the chunks being duplicated
666 within a stripe, whole stripes are duplicated but are rotated by one
667 device so duplicate blocks are on different devices. Thus subsequent
668 copies of a block are in the next drive, and are one chunk further
673 (multiple copies have very different offsets).
674 See md(4) for more detail about 'near', 'offset', and 'far'.
676 The number is the number of copies of each datablock. 2 is normal, 3
677 can be useful. This number can be at most equal to the number of
678 devices in the array. It does not need to divide evenly into that
679 number (e.g. it is perfectly legal to have an 'n2' layout for an array
680 with an odd number of devices).
682 A bug introduced in Linux 3.14 means that RAID0 arrays
683 .B "with devices of differing sizes"
684 started using a different layout. This could lead to
685 data corruption. Since Linux 5.4 (and various stable releases that received
686 backports), the kernel will not accept such an array unless
687 a layout is explictly set. It can be set to
691 When creating a new array,
695 by default, so the layout does not normally need to be set.
696 An array created for either
700 will not be recognized by an (unpatched) kernel prior to 5.4. To create
701 a RAID0 array with devices of differing sizes that can be used on an
702 older kernel, you can set the layout to
704 This will use whichever layout the running kernel supports, so the data
705 on the array may become corrupt when changing kernel from pre-3.14 to a
708 When an array is converted between RAID5 and RAID6 an intermediate
709 RAID6 layout is used in which the second parity block (Q) is always on
710 the last device. To convert a RAID5 to RAID6 and leave it in this new
711 layout (which does not require re-striping) use
712 .BR \-\-layout=preserve .
713 This will try to avoid any restriping.
715 The converse of this is
716 .B \-\-layout=normalise
717 which will change a non-standard RAID6 layout into a more standard
724 (thus explaining the p of
728 .BR \-b ", " \-\-bitmap=
729 Specify a file to store a write-intent bitmap in. The file should not
732 is also given. The same file should be provided
733 when assembling the array. If the word
735 is given, then the bitmap is stored with the metadata on the array,
736 and so is replicated on all devices. If the word
740 mode, then any bitmap that is present is removed. If the word
742 is given, the array is created for a clustered environment. One bitmap
743 is created for each node as defined by the
745 parameter and are stored internally.
747 To help catch typing errors, the filename must contain at least one
748 slash ('/') if it is a real file (not 'internal' or 'none').
750 Note: external bitmaps are only known to work on ext2 and ext3.
751 Storing bitmap files on other filesystems may result in serious problems.
753 When creating an array on devices which are 100G or larger,
755 automatically adds an internal bitmap as it will usually be
756 beneficial. This can be suppressed with
758 or by selecting a different consistency policy with
759 .BR \-\-consistency\-policy .
762 .BR \-\-bitmap\-chunk=
763 Set the chunksize of the bitmap. Each bit corresponds to that many
764 Kilobytes of storage.
765 When using a file based bitmap, the default is to use the smallest
766 size that is at-least 4 and requires no more than 2^21 chunks.
769 bitmap, the chunksize defaults to 64Meg, or larger if necessary to
770 fit the bitmap into the available space.
772 A suffix of 'K', 'M' or 'G' can be given to indicate Kilobytes, Megabytes or
773 Gigabytes respectively.
776 .BR \-W ", " \-\-write\-mostly
777 subsequent devices listed in a
782 command will be flagged as 'write\-mostly'. This is valid for RAID1
783 only and means that the 'md' driver will avoid reading from these
784 devices if at all possible. This can be useful if mirroring over a
788 .BR \-\-write\-behind=
789 Specify that write-behind mode should be enabled (valid for RAID1
790 only). If an argument is specified, it will set the maximum number
791 of outstanding writes allowed. The default value is 256.
792 A write-intent bitmap is required in order to use write-behind
793 mode, and write-behind is only attempted on drives marked as
798 subsequent devices listed in a
802 command will be flagged as 'failfast'. This is valid for RAID1 and
803 RAID10 only. IO requests to these devices will be encouraged to fail
804 quickly rather than cause long delays due to error handling. Also no
805 attempt is made to repair a read error on these devices.
807 If an array becomes degraded so that the 'failfast' device is the only
808 usable device, the 'failfast' flag will then be ignored and extended
809 delays will be preferred to complete failure.
811 The 'failfast' flag is appropriate for storage arrays which have a
812 low probability of true failure, but which may sometimes
813 cause unacceptable delays due to internal maintenance functions.
816 .BR \-\-assume\-clean
819 that the array pre-existed and is known to be clean. It can be useful
820 when trying to recover from a major failure as you can be sure that no
821 data will be affected unless you actually write to the array. It can
822 also be used when creating a RAID1 or RAID10 if you want to avoid the
823 initial resync, however this practice \(em while normally safe \(em is not
824 recommended. Use this only if you really know what you are doing.
826 When the devices that will be part of a new array were filled
827 with zeros before creation the operator knows the array is
828 actually clean. If that is the case, such as after running
829 badblocks, this argument can be used to tell mdadm the
830 facts the operator knows.
832 When an array is resized to a larger size with
833 .B "\-\-grow \-\-size="
834 the new space is normally resynced in that same way that the whole
835 array is resynced at creation. From Linux version 3.0,
837 can be used with that command to avoid the automatic resync.
840 .BR \-\-backup\-file=
843 is used to increase the number of raid-devices in a RAID5 or RAID6 if
844 there are no spare devices available, or to shrink, change RAID level
845 or layout. See the GROW MODE section below on RAID\-DEVICES CHANGES.
846 The file must be stored on a separate device, not on the RAID array
851 Arrays with 1.x metadata can leave a gap between the start of the
852 device and the start of array data. This gap can be used for various
853 metadata. The start of data is known as the
855 Normally an appropriate data offset is computed automatically.
856 However it can be useful to set it explicitly such as when re-creating
857 an array which was originally created using a different version of
859 which computed a different offset.
861 Setting the offset explicitly over-rides the default. The value given
862 is in Kilobytes unless a suffix of 'K', 'M' or 'G' is used to explicitly
863 indicate Kilobytes, Megabytes or Gigabytes respectively.
867 can also be used with
869 for some RAID levels (initially on RAID10). This allows the
870 data\-offset to be changed as part of the reshape process. When the
871 data offset is changed, no backup file is required as the difference
872 in offsets is used to provide the same functionality.
874 When the new offset is earlier than the old offset, the number of
875 devices in the array cannot shrink. When it is after the old offset,
876 the number of devices in the array cannot increase.
878 When creating an array,
882 In the case each member device is expected to have a offset appended
883 to the name, separated by a colon. This makes it possible to recreate
884 exactly an array which has varying data offsets (as can happen when
885 different versions of
887 are used to add different devices).
891 This option is complementary to the
892 .B \-\-freeze-reshape
893 option for assembly. It is needed when
895 operation is interrupted and it is not restarted automatically due to
896 .B \-\-freeze-reshape
897 usage during array assembly. This option is used together with
901 ) command and device for a pending reshape to be continued.
902 All parameters required for reshape continuation will be read from array metadata.
906 .BR \-\-backup\-file=
907 option to be set, continuation option will require to have exactly the same
908 backup file given as well.
910 Any other parameter passed together with
912 option will be ignored.
915 .BR \-N ", " \-\-name=
918 for the array. This is currently only effective when creating an
919 array with a version-1 superblock, or an array in a DDF container.
920 The name is a simple textual string that can be used to identify array
921 components when assembling. If name is needed but not specified, it
922 is taken from the basename of the device that is being created.
934 run the array, even if some of the components
935 appear to be active in another array or filesystem. Normally
937 will ask for confirmation before including such components in an
938 array. This option causes that question to be suppressed.
941 .BR \-f ", " \-\-force
944 accept the geometry and layout specified without question. Normally
946 will not allow creation of an array with only one device, and will try
947 to create a RAID5 array with one missing drive (as this makes the
948 initial resync work faster). With
951 will not try to be so clever.
954 .BR \-o ", " \-\-readonly
957 rather than read-write as normal. No writes will be allowed to the
958 array, and no resync, recovery, or reshape will be started. It works with
959 Create, Assemble, Manage and Misc mode.
962 .BR \-a ", " "\-\-auto{=yes,md,mdp,part,p}{NN}"
963 Instruct mdadm how to create the device file if needed, possibly allocating
964 an unused minor number. "md" causes a non-partitionable array
965 to be used (though since Linux 2.6.28, these array devices are in fact
966 partitionable). "mdp", "part" or "p" causes a partitionable array (2.6 and
967 later) to be used. "yes" requires the named md device to have
968 a 'standard' format, and the type and minor number will be determined
969 from this. With mdadm 3.0, device creation is normally left up to
971 so this option is unlikely to be needed.
972 See DEVICE NAMES below.
974 The argument can also come immediately after
979 is not given on the command line or in the config file, then
985 is also given, then any
987 entries in the config file will override the
989 instruction given on the command line.
991 For partitionable arrays,
993 will create the device file for the whole array and for the first 4
994 partitions. A different number of partitions can be specified at the
995 end of this option (e.g.
997 If the device name ends with a digit, the partition names add a 'p',
999 .IR /dev/md/home1p3 .
1000 If there is no trailing digit, then the partition names just have a
1002 .IR /dev/md/scratch3 .
1004 If the md device name is in a 'standard' format as described in DEVICE
1005 NAMES, then it will be created, if necessary, with the appropriate
1006 device number based on that name. If the device name is not in one of these
1007 formats, then a unused device number will be allocated. The device
1008 number will be considered unused if there is no active array for that
1009 number, and there is no entry in /dev for that number and with a
1010 non-standard name. Names that are not in 'standard' format are only
1011 allowed in "/dev/md/".
1013 This is meaningful with
1019 .BR \-a ", " "\-\-add"
1020 This option can be used in Grow mode in two cases.
1022 If the target array is a Linear array, then
1024 can be used to add one or more devices to the array. They
1025 are simply catenated on to the end of the array. Once added, the
1026 devices cannot be removed.
1030 option is being used to increase the number of devices in an array,
1033 can be used to add some extra devices to be included in the array.
1034 In most cases this is not needed as the extra devices can be added as
1035 spares first, and then the number of raid-disks can be changed.
1036 However for RAID0, it is not possible to add spares. So to increase
1037 the number of devices in a RAID0, it is necessary to set the new
1038 number of devices, and to add the new devices, in the same command.
1042 Only works when the array is for clustered environment. It specifies
1043 the maximum number of nodes in the cluster that will use this device
1044 simultaneously. If not specified, this defaults to 4.
1047 .BR \-\-write-journal
1048 Specify journal device for the RAID-4/5/6 array. The journal device
1049 should be a SSD with reasonable lifetime.
1053 Auto creation of symlinks in /dev to /dev/md, option --symlinks must
1054 be 'no' or 'yes' and work with --create and --build.
1057 .BR \-k ", " \-\-consistency\-policy=
1058 Specify how the array maintains consistency in case of unexpected shutdown.
1059 Only relevant for RAID levels with redundancy.
1060 Currently supported options are:
1065 Full resync is performed and all redundancy is regenerated when the array is
1066 started after unclean shutdown.
1070 Resync assisted by a write-intent bitmap. Implicitly selected when using
1075 For RAID levels 4/5/6, journal device is used to log transactions and replay
1076 after unclean shutdown. Implicitly selected when using
1077 .BR \-\-write\-journal .
1081 For RAID5 only, Partial Parity Log is used to close the write hole and
1082 eliminate resync. PPL is stored in the metadata region of RAID member drives,
1083 no additional journal drive is needed.
1086 Can be used with \-\-grow to change the consistency policy of an active array
1087 in some cases. See CONSISTENCY POLICY CHANGES below.
1094 .BR \-u ", " \-\-uuid=
1095 uuid of array to assemble. Devices which don't have this uuid are
1099 .BR \-m ", " \-\-super\-minor=
1100 Minor number of device that array was created for. Devices which
1101 don't have this minor number are excluded. If you create an array as
1102 /dev/md1, then all superblocks will contain the minor number 1, even if
1103 the array is later assembled as /dev/md2.
1105 Giving the literal word "dev" for
1109 to use the minor number of the md device that is being assembled.
1110 e.g. when assembling
1112 .B \-\-super\-minor=dev
1113 will look for super blocks with a minor number of 0.
1116 is only relevant for v0.90 metadata, and should not normally be used.
1122 .BR \-N ", " \-\-name=
1123 Specify the name of the array to assemble. This must be the name
1124 that was specified when creating the array. It must either match
1125 the name stored in the superblock exactly, or it must match
1128 prefixed to the start of the given name.
1131 .BR \-f ", " \-\-force
1132 Assemble the array even if the metadata on some devices appears to be
1135 cannot find enough working devices to start the array, but can find
1136 some devices that are recorded as having failed, then it will mark
1137 those devices as working so that the array can be started.
1138 An array which requires
1140 to be started may contain data corruption. Use it carefully.
1143 .BR \-R ", " \-\-run
1144 Attempt to start the array even if fewer drives were given than were
1145 present last time the array was active. Normally if not all the
1146 expected drives are found and
1148 is not used, then the array will be assembled but not started.
1151 an attempt will be made to start it anyway.
1155 This is the reverse of
1157 in that it inhibits the startup of array unless all expected drives
1158 are present. This is only needed with
1160 and can be used if the physical connections to devices are
1161 not as reliable as you would like.
1164 .BR \-a ", " "\-\-auto{=no,yes,md,mdp,part}"
1165 See this option under Create and Build options.
1168 .BR \-b ", " \-\-bitmap=
1169 Specify the bitmap file that was given when the array was created. If
1172 bitmap, there is no need to specify this when assembling the array.
1175 .BR \-\-backup\-file=
1178 was used while reshaping an array (e.g. changing number of devices or
1179 chunk size) and the system crashed during the critical section, then the same
1181 must be presented to
1183 to allow possibly corrupted data to be restored, and the reshape
1187 .BR \-\-invalid\-backup
1188 If the file needed for the above option is not available for any
1189 reason an empty file can be given together with this option to
1190 indicate that the backup file is invalid. In this case the data that
1191 was being rearranged at the time of the crash could be irrecoverably
1192 lost, but the rest of the array may still be recoverable. This option
1193 should only be used as a last resort if there is no way to recover the
1198 .BR \-U ", " \-\-update=
1199 Update the superblock on each device while assembling the array. The
1200 argument given to this flag can be one of
1216 .BR layout\-original ,
1217 .BR layout\-alternate ,
1224 option will adjust the superblock of an array what was created on a Sparc
1225 machine running a patched 2.2 Linux kernel. This kernel got the
1226 alignment of part of the superblock wrong. You can use the
1227 .B "\-\-examine \-\-sparc2.2"
1230 to see what effect this would have.
1234 option will update the
1235 .B "preferred minor"
1236 field on each superblock to match the minor number of the array being
1238 This can be useful if
1240 reports a different "Preferred Minor" to
1242 In some cases this update will be performed automatically
1243 by the kernel driver. In particular the update happens automatically
1244 at the first write to an array with redundancy (RAID level 1 or
1245 greater) on a 2.6 (or later) kernel.
1249 option will change the uuid of the array. If a UUID is given with the
1251 option that UUID will be used as a new UUID and will
1253 be used to help identify the devices in the array.
1256 is given, a random UUID is chosen.
1260 option will change the
1262 of the array as stored in the superblock. This is only supported for
1263 version-1 superblocks.
1267 option will change the
1269 of the array as stored in the bitmap superblock. This option only
1270 works for a clustered environment.
1274 option will change the
1276 as recorded in the superblock. For version-0 superblocks, this is the
1277 same as updating the UUID.
1278 For version-1 superblocks, this involves updating the name.
1282 option will change the cluster name as recorded in the superblock and
1283 bitmap. This option only works for clustered environment.
1287 option will cause the array to be marked
1289 meaning that any redundancy in the array (e.g. parity for RAID5,
1290 copies for RAID1) may be incorrect. This will cause the RAID system
1291 to perform a "resync" pass to make sure that all redundant information
1296 option allows arrays to be moved between machines with different
1297 byte-order, such as from a big-endian machine like a Sparc or some
1298 MIPS machines, to a little-endian x86_64 machine.
1299 When assembling such an array for the first time after a move, giving
1300 .B "\-\-update=byteorder"
1303 to expect superblocks to have their byteorder reversed, and will
1304 correct that order before assembling the array. This is only valid
1305 with original (Version 0.90) superblocks.
1309 option will correct the summaries in the superblock. That is the
1310 counts of total, working, active, failed, and spare devices.
1314 option will rarely be of use. It applies to version 1.1 and 1.2 metadata
1315 only (where the metadata is at the start of the device) and is only
1316 useful when the component device has changed size (typically become
1317 larger). The version 1 metadata records the amount of the device that
1318 can be used to store data, so if a device in a version 1.1 or 1.2
1319 array becomes larger, the metadata will still be visible, but the
1320 extra space will not. In this case it might be useful to assemble the
1322 .BR \-\-update=devicesize .
1325 to determine the maximum usable amount of space on each device and
1326 update the relevant field in the metadata.
1330 option only works on v0.90 metadata arrays and will convert them to
1331 v1.0 metadata. The array must not be dirty (i.e. it must not need a
1332 sync) and it must not have a write-intent bitmap.
1334 The old metadata will remain on the devices, but will appear older
1335 than the new metadata and so will usually be ignored. The old metadata
1336 (or indeed the new metadata) can be removed by giving the appropriate
1339 .BR \-\-zero\-superblock .
1343 option can be used when an array has an internal bitmap which is
1344 corrupt in some way so that assembling the array normally fails. It
1345 will cause any internal bitmap to be ignored.
1349 option will reserve space in each device for a bad block list. This
1350 will be 4K in size and positioned near the end of any free space
1351 between the superblock and the data.
1355 option will cause any reservation of space for a bad block list to be
1356 removed. If the bad block list contains entries, this will fail, as
1357 removing the list could cause data corruption.
1361 option will enable PPL for a RAID5 array and reserve space for PPL on each
1362 device. There must be enough free space between the data and superblock and a
1363 write-intent bitmap or journal must not be used.
1367 option will disable PPL in the superblock.
1372 .B layout\-alternate
1373 options are for RAID0 arrays in use before Linux 5.4. If the array was being
1374 used with Linux 3.13 or earlier, then to assemble the array on a new kernel,
1375 .B \-\-update=layout\-original
1376 must be given. If the array was created and used with a kernel from Linux 3.14 to
1378 .B \-\-update=layout\-alternate
1379 must be given. This only needs to be given once. Subsequent assembly of the array
1380 will happen normally.
1381 For more information, see
1385 .BR \-\-freeze\-reshape
1386 Option is intended to be used in start-up scripts during initrd boot phase.
1387 When array under reshape is assembled during initrd phase, this option
1388 stops reshape after reshape critical section is being restored. This happens
1389 before file system pivot operation and avoids loss of file system context.
1390 Losing file system context would cause reshape to be broken.
1392 Reshape can be continued later using the
1394 option for the grow command.
1398 See this option under Create and Build options.
1400 .SH For Manage mode:
1403 .BR \-t ", " \-\-test
1404 Unless a more serious error occurred,
1406 will exit with a status of 2 if no changes were made to the array and
1407 0 if at least one change was made.
1408 This can be useful when an indirect specifier such as
1413 is used in requesting an operation on the array.
1415 will report failure if these specifiers didn't find any match.
1418 .BR \-a ", " \-\-add
1419 hot-add listed devices.
1420 If a device appears to have recently been part of the array
1421 (possibly it failed or was removed) the device is re\-added as described
1423 If that fails or the device was never part of the array, the device is
1424 added as a hot-spare.
1425 If the array is degraded, it will immediately start to rebuild data
1428 Note that this and the following options are only meaningful on array
1429 with redundancy. They don't apply to RAID0 or Linear.
1433 re\-add a device that was previously removed from an array.
1434 If the metadata on the device reports that it is a member of the
1435 array, and the slot that it used is still vacant, then the device will
1436 be added back to the array in the same position. This will normally
1437 cause the data for that device to be recovered. However based on the
1438 event count on the device, the recovery may only require sections that
1439 are flagged a write-intent bitmap to be recovered or may not require
1440 any recovery at all.
1442 When used on an array that has no metadata (i.e. it was built with
1444 it will be assumed that bitmap-based recovery is enough to make the
1445 device fully consistent with the array.
1447 When used with v1.x metadata,
1449 can be accompanied by
1450 .BR \-\-update=devicesize ,
1451 .BR \-\-update=bbl ", or"
1452 .BR \-\-update=no\-bbl .
1453 See the description of these option when used in Assemble mode for an
1454 explanation of their use.
1456 If the device name given is
1460 will try to find any device that looks like it should be
1461 part of the array but isn't and will try to re\-add all such devices.
1463 If the device name given is
1467 will find all devices in the array that are marked
1469 remove them and attempt to immediately re\-add them. This can be
1470 useful if you are certain that the reason for failure has been
1475 Add a device as a spare. This is similar to
1477 except that it does not attempt
1479 first. The device will be added as a spare even if it looks like it
1480 could be an recent member of the array.
1483 .BR \-r ", " \-\-remove
1484 remove listed devices. They must not be active. i.e. they should
1485 be failed or spare devices.
1487 As well as the name of a device file
1497 The first causes all failed device to be removed. The second causes
1498 any device which is no longer connected to the system (i.e an 'open'
1502 The third will remove a set as describe below under
1506 .BR \-f ", " \-\-fail
1507 Mark listed devices as faulty.
1508 As well as the name of a device file, the word
1512 can be given. The former will cause any device that has been detached from
1513 the system to be marked as failed. It can then be removed.
1515 For RAID10 arrays where the number of copies evenly divides the number
1516 of devices, the devices can be conceptually divided into sets where
1517 each set contains a single complete copy of the data on the array.
1518 Sometimes a RAID10 array will be configured so that these sets are on
1519 separate controllers. In this case all the devices in one set can be
1520 failed by giving a name like
1526 The appropriate set names are reported by
1536 Mark listed devices as requiring replacement. As soon as a spare is
1537 available, it will be rebuilt and will replace the marked device.
1538 This is similar to marking a device as faulty, but the device remains
1539 in service during the recovery process to increase resilience against
1540 multiple failures. When the replacement process finishes, the
1541 replaced device will be marked as faulty.
1545 This can follow a list of
1547 devices. The devices listed after
1549 will be preferentially used to replace the devices listed after
1551 These device must already be spare devices in the array.
1554 .BR \-\-write\-mostly
1555 Subsequent devices that are added or re\-added will have the 'write-mostly'
1556 flag set. This is only valid for RAID1 and means that the 'md' driver
1557 will avoid reading from these devices if possible.
1560 Subsequent devices that are added or re\-added will have the 'write-mostly'
1563 .BR \-\-cluster\-confirm
1564 Confirm the existence of the device. This is issued in response to an \-\-add
1565 request by a node in a cluster. When a node adds a device it sends a message
1566 to all nodes in the cluster to look for a device with a UUID. This translates
1567 to a udev notification with the UUID of the device to be added and the slot
1568 number. The receiving node must acknowledge this message
1569 with \-\-cluster\-confirm. Valid arguments are <slot>:<devicename> in case
1570 the device is found or <slot>:missing in case the device is not found.
1574 Add journal to an existing array, or recreate journal for RAID-4/5/6 array
1575 that lost a journal device. To avoid interrupting on-going write opertions,
1577 only works for array in Read-Only state.
1581 Subsequent devices that are added or re\-added will have
1582 the 'failfast' flag set. This is only valid for RAID1 and RAID10 and
1583 means that the 'md' driver will avoid long timeouts on error handling
1587 Subsequent devices that are re\-added will be re\-added without
1588 the 'failfast' flag set.
1591 Each of these options requires that the first device listed is the array
1592 to be acted upon, and the remainder are component devices to be added,
1593 removed, marked as faulty, etc. Several different operations can be
1594 specified for different devices, e.g.
1596 mdadm /dev/md0 \-\-add /dev/sda1 \-\-fail /dev/sdb1 \-\-remove /dev/sdb1
1598 Each operation applies to all devices listed until the next
1601 If an array is using a write-intent bitmap, then devices which have
1602 been removed can be re\-added in a way that avoids a full
1603 reconstruction but instead just updates the blocks that have changed
1604 since the device was removed. For arrays with persistent metadata
1605 (superblocks) this is done automatically. For arrays created with
1607 mdadm needs to be told that this device we removed recently with
1610 Devices can only be removed from an array if they are not in active
1611 use, i.e. that must be spares or failed devices. To remove an active
1612 device, it must first be marked as
1618 .BR \-Q ", " \-\-query
1619 Examine a device to see
1620 (1) if it is an md device and (2) if it is a component of an md
1622 Information about what is discovered is presented.
1625 .BR \-D ", " \-\-detail
1626 Print details of one or more md devices.
1629 .BR \-\-detail\-platform
1630 Print details of the platform's RAID capabilities (firmware / hardware
1631 topology) for a given metadata format. If used without argument, mdadm
1632 will scan all controllers looking for their capabilities. Otherwise, mdadm
1633 will only look at the controller specified by the argument in form of an
1634 absolute filepath or a link, e.g.
1635 .IR /sys/devices/pci0000:00/0000:00:1f.2 .
1638 .BR \-Y ", " \-\-export
1641 .BR \-\-detail-platform ,
1645 output will be formatted as
1647 pairs for easy import into the environment.
1653 indicates whether an array was started
1655 or not, which may include a reason
1656 .RB ( unsafe ", " nothing ", " no ).
1659 indicates if the array is expected on this host
1661 or seems to be from elsewhere
1665 .BR \-E ", " \-\-examine
1666 Print contents of the metadata stored on the named device(s).
1667 Note the contrast between
1672 applies to devices which are components of an array, while
1674 applies to a whole array which is currently active.
1677 If an array was created on a SPARC machine with a 2.2 Linux kernel
1678 patched with RAID support, the superblock will have been created
1679 incorrectly, or at least incompatibly with 2.4 and later kernels.
1684 will fix the superblock before displaying it. If this appears to do
1685 the right thing, then the array can be successfully assembled using
1686 .BR "\-\-assemble \-\-update=sparc2.2" .
1689 .BR \-X ", " \-\-examine\-bitmap
1690 Report information about a bitmap file.
1691 The argument is either an external bitmap file or an array component
1692 in case of an internal bitmap. Note that running this on an array
1695 does not report the bitmap for that array.
1698 .B \-\-examine\-badblocks
1699 List the bad-blocks recorded for the device, if a bad-blocks list has
1700 been configured. Currently only
1702 metadata supports bad-blocks lists.
1705 .BI \-\-dump= directory
1707 .BI \-\-restore= directory
1708 Save metadata from lists devices, or restore metadata to listed devices.
1711 .BR \-R ", " \-\-run
1712 start a partially assembled array. If
1714 did not find enough devices to fully start the array, it might leaving
1715 it partially assembled. If you wish, you can then use
1717 to start the array in degraded mode.
1720 .BR \-S ", " \-\-stop
1721 deactivate array, releasing all resources.
1724 .BR \-o ", " \-\-readonly
1725 mark array as readonly.
1728 .BR \-w ", " \-\-readwrite
1729 mark array as readwrite.
1732 .B \-\-zero\-superblock
1733 If the device contains a valid md superblock, the block is
1734 overwritten with zeros. With
1736 the block where the superblock would be is overwritten even if it
1737 doesn't appear to be valid.
1740 Be careful to call \-\-zero\-superblock with clustered raid, make sure
1741 array isn't used or assembled in other cluster node before execute it.
1744 .B \-\-kill\-subarray=
1745 If the device is a container and the argument to \-\-kill\-subarray
1746 specifies an inactive subarray in the container, then the subarray is
1747 deleted. Deleting all subarrays will leave an 'empty-container' or
1748 spare superblock on the drives. See
1749 .B \-\-zero\-superblock
1751 removing a superblock. Note that some formats depend on the subarray
1752 index for generating a UUID, this command will fail if it would change
1753 the UUID of an active subarray.
1756 .B \-\-update\-subarray=
1757 If the device is a container and the argument to \-\-update\-subarray
1758 specifies a subarray in the container, then attempt to update the given
1759 superblock field in the subarray. See below in
1764 .BR \-t ", " \-\-test
1769 is set to reflect the status of the device. See below in
1774 .BR \-W ", " \-\-wait
1775 For each md device given, wait for any resync, recovery, or reshape
1776 activity to finish before returning.
1778 will return with success if it actually waited for every device
1779 listed, otherwise it will return failure.
1783 For each md device given, or each device in /proc/mdstat if
1785 is given, arrange for the array to be marked clean as soon as possible.
1787 will return with success if the array uses external metadata and we
1788 successfully waited. For native arrays this returns immediately as the
1789 kernel handles dirty-clean transitions at shutdown. No action is taken
1790 if safe-mode handling is disabled.
1794 Set the "sync_action" for all md devices given to one of
1801 will abort any currently running action though some actions will
1802 automatically restart.
1805 will abort any current action and ensure no other action starts
1815 .BR "SCRUBBING AND MISMATCHES" .
1817 .SH For Incremental Assembly mode:
1819 .BR \-\-rebuild\-map ", " \-r
1820 Rebuild the map file
1824 uses to help track which arrays are currently being assembled.
1827 .BR \-\-run ", " \-R
1828 Run any array assembled as soon as a minimal number of devices are
1829 available, rather than waiting until all expected devices are present.
1832 .BR \-\-scan ", " \-s
1833 Only meaningful with
1837 file for arrays that are being incrementally assembled and will try to
1838 start any that are not already started. If any such array is listed
1841 as requiring an external bitmap, that bitmap will be attached first.
1844 .BR \-\-fail ", " \-f
1845 This allows the hot-plug system to remove devices that have fully disappeared
1846 from the kernel. It will first fail and then remove the device from any
1847 array it belongs to.
1848 The device name given should be a kernel device name such as "sda",
1854 Only used with \-\-fail. The 'path' given will be recorded so that if
1855 a new device appears at the same location it can be automatically
1856 added to the same array. This allows the failed device to be
1857 automatically replaced by a new device without metadata if it appears
1858 at specified path. This option is normally only set by a
1862 .SH For Monitor mode:
1864 .BR \-m ", " \-\-mail
1865 Give a mail address to send alerts to.
1868 .BR \-p ", " \-\-program ", " \-\-alert
1869 Give a program to be run whenever an event is detected.
1872 .BR \-y ", " \-\-syslog
1873 Cause all events to be reported through 'syslog'. The messages have
1874 facility of 'daemon' and varying priorities.
1877 .BR \-d ", " \-\-delay
1878 Give a delay in seconds.
1880 polls the md arrays and then waits this many seconds before polling
1881 again. The default is 60 seconds. Since 2.6.16, there is no need to
1882 reduce this as the kernel alerts
1884 immediately when there is any change.
1887 .BR \-r ", " \-\-increment
1888 Give a percentage increment.
1890 will generate RebuildNN events with the given percentage increment.
1893 .BR \-f ", " \-\-daemonise
1896 to run as a background daemon if it decides to monitor anything. This
1897 causes it to fork and run in the child, and to disconnect from the
1898 terminal. The process id of the child is written to stdout.
1901 which will only continue monitoring if a mail address or alert program
1902 is found in the config file.
1905 .BR \-i ", " \-\-pid\-file
1908 is running in daemon mode, write the pid of the daemon process to
1909 the specified file, instead of printing it on standard output.
1912 .BR \-1 ", " \-\-oneshot
1913 Check arrays only once. This will generate
1915 events and more significantly
1921 .B " mdadm \-\-monitor \-\-scan \-1"
1923 from a cron script will ensure regular notification of any degraded arrays.
1926 .BR \-t ", " \-\-test
1929 alert for every array found at startup. This alert gets mailed and
1930 passed to the alert program. This can be used for testing that alert
1931 message do get through successfully.
1935 This inhibits the functionality for moving spares between arrays.
1936 Only one monitoring process started with
1938 but without this flag is allowed, otherwise the two could interfere
1945 .B mdadm \-\-assemble
1946 .I md-device options-and-component-devices...
1949 .B mdadm \-\-assemble \-\-scan
1950 .I md-devices-and-options...
1953 .B mdadm \-\-assemble \-\-scan
1957 This usage assembles one or more RAID arrays from pre-existing components.
1958 For each array, mdadm needs to know the md device, the identity of the
1959 array, and a number of component-devices. These can be found in a number of ways.
1961 In the first usage example (without the
1963 the first device given is the md device.
1964 In the second usage example, all devices listed are treated as md
1965 devices and assembly is attempted.
1966 In the third (where no devices are listed) all md devices that are
1967 listed in the configuration file are assembled. If no arrays are
1968 described by the configuration file, then any arrays that
1969 can be found on unused devices will be assembled.
1971 If precisely one device is listed, but
1977 was given and identity information is extracted from the configuration file.
1979 The identity can be given with the
1985 option, will be taken from the md-device record in the config file, or
1986 will be taken from the super block of the first component-device
1987 listed on the command line.
1989 Devices can be given on the
1991 command line or in the config file. Only devices which have an md
1992 superblock which contains the right identity will be considered for
1995 The config file is only used if explicitly named with
1997 or requested with (a possibly implicit)
2002 .B /etc/mdadm/mdadm.conf
2007 is not given, then the config file will only be used to find the
2008 identity of md arrays.
2010 Normally the array will be started after it is assembled. However if
2012 is not given and not all expected drives were listed, then the array
2013 is not started (to guard against usage errors). To insist that the
2014 array be started in this case (as may work for RAID1, 4, 5, 6, or 10),
2023 does not create any entries in
2027 It does record information in
2031 to choose the correct name.
2035 detects that udev is not configured, it will create the devices in
2039 In Linux kernels prior to version 2.6.28 there were two distinctly
2040 different types of md devices that could be created: one that could be
2041 partitioned using standard partitioning tools and one that could not.
2042 Since 2.6.28 that distinction is no longer relevant as both type of
2043 devices can be partitioned.
2045 will normally create the type that originally could not be partitioned
2046 as it has a well defined major number (9).
2048 Prior to 2.6.28, it is important that mdadm chooses the correct type
2049 of array device to use. This can be controlled with the
2051 option. In particular, a value of "mdp" or "part" or "p" tells mdadm
2052 to use a partitionable device rather than the default.
2054 In the no-udev case, the value given to
2056 can be suffixed by a number. This tells
2058 to create that number of partition devices rather than the default of 4.
2062 can also be given in the configuration file as a word starting
2064 on the ARRAY line for the relevant array.
2071 and no devices are listed,
2073 will first attempt to assemble all the arrays listed in the config
2076 If no arrays are listed in the config (other than those marked
2078 it will look through the available devices for possible arrays and
2079 will try to assemble anything that it finds. Arrays which are tagged
2080 as belonging to the given homehost will be assembled and started
2081 normally. Arrays which do not obviously belong to this host are given
2082 names that are expected not to conflict with anything local, and are
2083 started "read-auto" so that nothing is written to any device until the
2084 array is written to. i.e. automatic resync etc is delayed.
2088 finds a consistent set of devices that look like they should comprise
2089 an array, and if the superblock is tagged as belonging to the given
2090 home host, it will automatically choose a device name and try to
2091 assemble the array. If the array uses version-0.90 metadata, then the
2093 number as recorded in the superblock is used to create a name in
2097 If the array uses version-1 metadata, then the
2099 from the superblock is used to similarly create a name in
2101 (the name will have any 'host' prefix stripped first).
2103 This behaviour can be modified by the
2107 configuration file. This line can indicate that specific metadata
2108 type should, or should not, be automatically assembled. If an array
2109 is found which is not listed in
2111 and has a metadata format that is denied by the
2113 line, then it will not be assembled.
2116 line can also request that all arrays identified as being for this
2117 homehost should be assembled regardless of their metadata type.
2120 for further details.
2122 Note: Auto assembly cannot be used for assembling and activating some
2123 arrays which are undergoing reshape. In particular as the
2125 cannot be given, any reshape which requires a backup-file to continue
2126 cannot be started by auto assembly. An array which is growing to more
2127 devices and has passed the critical section can be assembled using
2138 .BI \-\-raid\-devices= Z
2142 This usage is similar to
2144 The difference is that it creates an array without a superblock. With
2145 these arrays there is no difference between initially creating the array and
2146 subsequently assembling the array, except that hopefully there is useful
2147 data there in the second case.
2149 The level may raid0, linear, raid1, raid10, multipath, or faulty, or
2150 one of their synonyms. All devices must be listed and the array will
2151 be started once complete. It will often be appropriate to use
2152 .B \-\-assume\-clean
2153 with levels raid1 or raid10.
2164 .BI \-\-raid\-devices= Z
2168 This usage will initialise a new md array, associate some devices with
2169 it, and activate the array.
2171 The named device will normally not exist when
2172 .I "mdadm \-\-create"
2173 is run, but will be created by
2175 once the array becomes active.
2177 As devices are added, they are checked to see if they contain RAID
2178 superblocks or filesystems. They are also checked to see if the variance in
2179 device size exceeds 1%.
2181 If any discrepancy is found, the array will not automatically be run, though
2184 can override this caution.
2186 To create a "degraded" array in which some devices are missing, simply
2187 give the word "\fBmissing\fP"
2188 in place of a device name. This will cause
2190 to leave the corresponding slot in the array empty.
2191 For a RAID4 or RAID5 array at most one slot can be
2192 "\fBmissing\fP"; for a RAID6 array at most two slots.
2193 For a RAID1 array, only one real device needs to be given. All of the
2197 When creating a RAID5 array,
2199 will automatically create a degraded array with an extra spare drive.
2200 This is because building the spare into a degraded array is in general
2201 faster than resyncing the parity on a non-degraded, but not clean,
2202 array. This feature can be overridden with the
2206 When creating an array with version-1 metadata a name for the array is
2208 If this is not given with the
2212 will choose a name based on the last component of the name of the
2213 device being created. So if
2215 is being created, then the name
2220 is being created, then the name
2224 When creating a partition based array, using
2226 with version-1.x metadata, the partition type should be set to
2228 (non fs-data). This type selection allows for greater precision since
2229 using any other [RAID auto-detect (0xFD) or a GNU/Linux partition (0x83)],
2230 might create problems in the event of array recovery through a live cdrom.
2232 A new array will normally get a randomly assigned 128bit UUID which is
2233 very likely to be unique. If you have a specific need, you can choose
2234 a UUID for the array by giving the
2236 option. Be warned that creating two arrays with the same UUID is a
2237 recipe for disaster. Also, using
2239 when creating a v0.90 array will silently override any
2244 .\"option is given, it is not necessary to list any component-devices in this command.
2245 .\"They can be added later, before a
2249 .\"is given, the apparent size of the smallest drive given is used.
2251 If the array type supports a write-intent bitmap, and if the devices
2252 in the array exceed 100G is size, an internal write-intent bitmap
2253 will automatically be added unless some other option is explicitly
2256 option or a different consistency policy is selected with the
2257 .B \-\-consistency\-policy
2258 option. In any case space for a bitmap will be reserved so that one
2259 can be added later with
2260 .BR "\-\-grow \-\-bitmap=internal" .
2262 If the metadata type supports it (currently only 1.x and IMSM metadata),
2263 space will be allocated to store a bad block list. This allows a modest
2264 number of bad blocks to be recorded, allowing the drive to remain in
2265 service while only partially functional.
2267 When creating an array within a
2270 can be given either the list of devices to use, or simply the name of
2271 the container. The former case gives control over which devices in
2272 the container will be used for the array. The latter case allows
2274 to automatically choose which devices to use based on how much spare
2277 The General Management options that are valid with
2282 insist on running the array even if some devices look like they might
2287 start the array in readonly mode.
2294 .I options... devices...
2297 This usage will allow individual devices in an array to be failed,
2298 removed or added. It is possible to perform multiple operations with
2299 on command. For example:
2301 .B " mdadm /dev/md0 \-f /dev/hda1 \-r /dev/hda1 \-a /dev/hda1"
2307 and will then remove it from the array and finally add it back
2308 in as a spare. However only one md array can be affected by a single
2311 When a device is added to an active array, mdadm checks to see if it
2312 has metadata on it which suggests that it was recently a member of the
2313 array. If it does, it tries to "re\-add" the device. If there have
2314 been no changes since the device was removed, or if the array has a
2315 write-intent bitmap which has recorded whatever changes there were,
2316 then the device will immediately become a full member of the array and
2317 those differences recorded in the bitmap will be resolved.
2327 MISC mode includes a number of distinct operations that
2328 operate on distinct devices. The operations are:
2331 The device is examined to see if it is
2332 (1) an active md array, or
2333 (2) a component of an md array.
2334 The information discovered is reported.
2338 The device should be an active md device.
2340 will display a detailed description of the array.
2344 will cause the output to be less detailed and the format to be
2345 suitable for inclusion in
2349 will normally be 0 unless
2351 failed to get useful information about the device(s); however, if the
2353 option is given, then the exit status will be:
2357 The array is functioning normally.
2360 The array has at least one failed device.
2363 The array has multiple failed devices such that it is unusable.
2366 There was an error while trying to get information about the device.
2370 .B \-\-detail\-platform
2371 Print detail of the platform's RAID capabilities (firmware / hardware
2372 topology). If the metadata is specified with
2376 then the return status will be:
2380 metadata successfully enumerated its platform components on this system
2383 metadata is platform independent
2386 metadata failed to find its platform components on this system
2390 .B \-\-update\-subarray=
2391 If the device is a container and the argument to \-\-update\-subarray
2392 specifies a subarray in the container, then attempt to update the given
2393 superblock field in the subarray. Similar to updating an array in
2394 "assemble" mode, the field to update is selected by
2398 option. The supported options are
2406 option updates the subarray name in the metadata, it may not affect the
2407 device node name or the device node symlink until the subarray is
2408 re\-assembled. If updating
2410 would change the UUID of an active subarray this operation is blocked,
2411 and the command will end in an error.
2417 options enable and disable PPL in the metadata. Currently supported only for
2422 The device should be a component of an md array.
2424 will read the md superblock of the device and display the contents.
2429 is given, then multiple devices that are components of the one array
2430 are grouped together and reported in a single entry suitable
2436 without listing any devices will cause all devices listed in the
2437 config file to be examined.
2440 .BI \-\-dump= directory
2441 If the device contains RAID metadata, a file will be created in the
2443 and the metadata will be written to it. The file will be the same
2444 size as the device and have the metadata written in the file at the
2445 same locate that it exists in the device. However the file will be "sparse" so
2446 that only those blocks containing metadata will be allocated. The
2447 total space used will be small.
2449 The file name used in the
2451 will be the base name of the device. Further if any links appear in
2453 which point to the device, then hard links to the file will be created
2460 Multiple devices can be listed and their metadata will all be stored
2461 in the one directory.
2464 .BI \-\-restore= directory
2465 This is the reverse of
2468 will locate a file in the directory that has a name appropriate for
2469 the given device and will restore metadata from it. Names that match
2471 names are preferred, however if two of those refer to different files,
2473 will not choose between them but will abort the operation.
2475 If a file name is given instead of a
2479 will restore from that file to a single device, always provided the
2480 size of the file matches that of the device, and the file contains
2484 The devices should be active md arrays which will be deactivated, as
2485 long as they are not currently in use.
2489 This will fully activate a partially assembled md array.
2493 This will mark an active array as read-only, providing that it is
2494 not currently being used.
2500 array back to being read/write.
2504 For all operations except
2507 will cause the operation to be applied to all arrays listed in
2512 causes all devices listed in the config file to be examined.
2515 .BR \-b ", " \-\-brief
2516 Be less verbose. This is used with
2524 gives an intermediate level of verbosity.
2530 .B mdadm \-\-monitor
2531 .I options... devices...
2536 to periodically poll a number of md arrays and to report on any events
2539 will never exit once it decides that there are arrays to be checked,
2540 so it should normally be run in the background.
2542 As well as reporting events,
2544 may move a spare drive from one array to another if they are in the
2549 and if the destination array has a failed drive but no spares.
2551 If any devices are listed on the command line,
2553 will only monitor those devices. Otherwise all arrays listed in the
2554 configuration file will be monitored. Further, if
2556 is given, then any other md devices that appear in
2558 will also be monitored.
2560 The result of monitoring the arrays is the generation of events.
2561 These events are passed to a separate program (if specified) and may
2562 be mailed to a given E-mail address.
2564 When passing events to a program, the program is run once for each event,
2565 and is given 2 or 3 command-line arguments: the first is the
2566 name of the event (see below), the second is the name of the
2567 md device which is affected, and the third is the name of a related
2568 device if relevant (such as a component device that has failed).
2572 is given, then a program or an E-mail address must be specified on the
2573 command line or in the config file. If neither are available, then
2575 will not monitor anything.
2579 will continue monitoring as long as something was found to monitor. If
2580 no program or email is given, then each event is reported to
2583 The different events are:
2587 .B DeviceDisappeared
2588 An md array which previously was configured appears to no longer be
2589 configured. (syslog priority: Critical)
2593 was told to monitor an array which is RAID0 or Linear, then it will
2595 .B DeviceDisappeared
2596 with the extra information
2598 This is because RAID0 and Linear do not support the device-failed,
2599 hot-spare and resync operations which are monitored.
2603 An md array started reconstruction (e.g. recovery, resync, reshape,
2604 check, repair). (syslog priority: Warning)
2610 is a two-digit number (ie. 05, 48). This indicates that rebuild
2611 has passed that many percent of the total. The events are generated
2612 with fixed increment since 0. Increment size may be specified with
2613 a commandline option (default is 20). (syslog priority: Warning)
2617 An md array that was rebuilding, isn't any more, either because it
2618 finished normally or was aborted. (syslog priority: Warning)
2622 An active component device of an array has been marked as
2623 faulty. (syslog priority: Critical)
2627 A spare component device which was being rebuilt to replace a faulty
2628 device has failed. (syslog priority: Critical)
2632 A spare component device which was being rebuilt to replace a faulty
2633 device has been successfully rebuilt and has been made active.
2634 (syslog priority: Info)
2638 A new md array has been detected in the
2640 file. (syslog priority: Info)
2644 A newly noticed array appears to be degraded. This message is not
2647 notices a drive failure which causes degradation, but only when
2649 notices that an array is degraded when it first sees the array.
2650 (syslog priority: Critical)
2654 A spare drive has been moved from one array in a
2658 to another to allow a failed drive to be replaced.
2659 (syslog priority: Info)
2665 has been told, via the config file, that an array should have a certain
2666 number of spare devices, and
2668 detects that it has fewer than this number when it first sees the
2669 array, it will report a
2672 (syslog priority: Warning)
2676 An array was found at startup, and the
2679 (syslog priority: Info)
2689 cause Email to be sent. All events cause the program to be run.
2690 The program is run with two or three arguments: the event
2691 name, the array device and possibly a second device.
2693 Each event has an associated array device (e.g.
2695 and possibly a second device. For
2700 the second device is the relevant component device.
2703 the second device is the array that the spare was moved from.
2707 to move spares from one array to another, the different arrays need to
2708 be labeled with the same
2710 or the spares must be allowed to migrate through matching POLICY domains
2711 in the configuration file. The
2713 name can be any string; it is only necessary that different spare
2714 groups use different names.
2718 detects that an array in a spare group has fewer active
2719 devices than necessary for the complete array, and has no spare
2720 devices, it will look for another array in the same spare group that
2721 has a full complement of working drive and a spare. It will then
2722 attempt to remove the spare from the second drive and add it to the
2724 If the removal succeeds but the adding fails, then it is added back to
2727 If the spare group for a degraded array is not defined,
2729 will look at the rules of spare migration specified by POLICY lines in
2731 and then follow similar steps as above if a matching spare is found.
2734 The GROW mode is used for changing the size or shape of an active
2736 For this to work, the kernel must support the necessary change.
2737 Various types of growth are being added during 2.6 development.
2739 Currently the supported changes include
2741 change the "size" attribute for RAID1, RAID4, RAID5 and RAID6.
2743 increase or decrease the "raid\-devices" attribute of RAID0, RAID1, RAID4,
2746 change the chunk-size and layout of RAID0, RAID4, RAID5, RAID6 and RAID10.
2748 convert between RAID1 and RAID5, between RAID5 and RAID6, between
2749 RAID0, RAID4, and RAID5, and between RAID0 and RAID10 (in the near-2 mode).
2751 add a write-intent bitmap to any array which supports these bitmaps, or
2752 remove a write-intent bitmap from such an array.
2754 change the array's consistency policy.
2757 Using GROW on containers is currently supported only for Intel's IMSM
2758 container format. The number of devices in a container can be
2759 increased - which affects all arrays in the container - or an array
2760 in a container can be converted between levels where those levels are
2761 supported by the container, and the conversion is on of those listed
2762 above. Resizing arrays in an IMSM container with
2764 is not yet supported.
2769 Intel's native checkpointing doesn't use
2771 option and it is transparent for assembly feature.
2773 Roaming between Windows(R) and Linux systems for IMSM metadata is not
2774 supported during grow process.
2777 Normally when an array is built the "size" is taken from the smallest
2778 of the drives. If all the small drives in an arrays are, one at a
2779 time, removed and replaced with larger drives, then you could have an
2780 array of large drives with only a small amount used. In this
2781 situation, changing the "size" with "GROW" mode will allow the extra
2782 space to start being used. If the size is increased in this way, a
2783 "resync" process will start to make sure the new parts of the array
2786 Note that when an array changes size, any filesystem that may be
2787 stored in the array will not automatically grow or shrink to use or
2788 vacate the space. The
2789 filesystem will need to be explicitly told to use the extra space
2790 after growing, or to reduce its size
2792 to shrinking the array.
2794 Also the size of an array cannot be changed while it has an active
2795 bitmap. If an array has a bitmap, it must be removed before the size
2796 can be changed. Once the change is complete a new bitmap can be created.
2801 is not yet supported for external file bitmap.
2803 .SS RAID\-DEVICES CHANGES
2805 A RAID1 array can work with any number of devices from 1 upwards
2806 (though 1 is not very useful). There may be times which you want to
2807 increase or decrease the number of active devices. Note that this is
2808 different to hot-add or hot-remove which changes the number of
2811 When reducing the number of devices in a RAID1 array, the slots which
2812 are to be removed from the array must already be vacant. That is, the
2813 devices which were in those slots must be failed and removed.
2815 When the number of devices is increased, any hot spares that are
2816 present will be activated immediately.
2818 Changing the number of active devices in a RAID5 or RAID6 is much more
2819 effort. Every block in the array will need to be read and written
2820 back to a new location. From 2.6.17, the Linux Kernel is able to
2821 increase the number of devices in a RAID5 safely, including restarting
2822 an interrupted "reshape". From 2.6.31, the Linux Kernel is able to
2823 increase or decrease the number of devices in a RAID5 or RAID6.
2825 From 2.6.35, the Linux Kernel is able to convert a RAID0 in to a RAID4
2828 uses this functionality and the ability to add
2829 devices to a RAID4 to allow devices to be added to a RAID0. When
2830 requested to do this,
2832 will convert the RAID0 to a RAID4, add the necessary disks and make
2833 the reshape happen, and then convert the RAID4 back to RAID0.
2835 When decreasing the number of devices, the size of the array will also
2836 decrease. If there was data in the array, it could get destroyed and
2837 this is not reversible, so you should firstly shrink the filesystem on
2838 the array to fit within the new size. To help prevent accidents,
2840 requires that the size of the array be decreased first with
2841 .BR "mdadm --grow --array-size" .
2842 This is a reversible change which simply makes the end of the array
2843 inaccessible. The integrity of any data can then be checked before
2844 the non-reversible reduction in the number of devices is request.
2846 When relocating the first few stripes on a RAID5 or RAID6, it is not
2847 possible to keep the data on disk completely consistent and
2848 crash-proof. To provide the required safety, mdadm disables writes to
2849 the array while this "critical section" is reshaped, and takes a
2850 backup of the data that is in that section. For grows, this backup may be
2851 stored in any spare devices that the array has, however it can also be
2852 stored in a separate file specified with the
2854 option, and is required to be specified for shrinks, RAID level
2855 changes and layout changes. If this option is used, and the system
2856 does crash during the critical period, the same file must be passed to
2858 to restore the backup and reassemble the array. When shrinking rather
2859 than growing the array, the reshape is done from the end towards the
2860 beginning, so the "critical section" is at the end of the reshape.
2864 Changing the RAID level of any array happens instantaneously. However
2865 in the RAID5 to RAID6 case this requires a non-standard layout of the
2866 RAID6 data, and in the RAID6 to RAID5 case that non-standard layout is
2867 required before the change can be accomplished. So while the level
2868 change is instant, the accompanying layout change can take quite a
2871 is required. If the array is not simultaneously being grown or
2872 shrunk, so that the array size will remain the same - for example,
2873 reshaping a 3-drive RAID5 into a 4-drive RAID6 - the backup file will
2874 be used not just for a "cricital section" but throughout the reshape
2875 operation, as described below under LAYOUT CHANGES.
2877 .SS CHUNK-SIZE AND LAYOUT CHANGES
2879 Changing the chunk-size of layout without also changing the number of
2880 devices as the same time will involve re-writing all blocks in-place.
2881 To ensure against data loss in the case of a crash, a
2883 must be provided for these changes. Small sections of the array will
2884 be copied to the backup file while they are being rearranged. This
2885 means that all the data is copied twice, once to the backup and once
2886 to the new layout on the array, so this type of reshape will go very
2889 If the reshape is interrupted for any reason, this backup file must be
2891 .B "mdadm --assemble"
2892 so the array can be reassembled. Consequently the file cannot be
2893 stored on the device being reshaped.
2898 A write-intent bitmap can be added to, or removed from, an active
2899 array. Either internal bitmaps, or bitmaps stored in a separate file,
2900 can be added. Note that if you add a bitmap stored in a file which is
2901 in a filesystem that is on the RAID array being affected, the system
2902 will deadlock. The bitmap must be on a separate filesystem.
2904 .SS CONSISTENCY POLICY CHANGES
2906 The consistency policy of an active array can be changed by using the
2907 .B \-\-consistency\-policy
2908 option in Grow mode. Currently this works only for the
2912 policies and allows to enable or disable the RAID5 Partial Parity Log (PPL).
2914 .SH INCREMENTAL MODE
2918 .B mdadm \-\-incremental
2922 .RI [ optional-aliases-for-device ]
2925 .B mdadm \-\-incremental \-\-fail
2929 .B mdadm \-\-incremental \-\-rebuild\-map
2932 .B mdadm \-\-incremental \-\-run \-\-scan
2935 This mode is designed to be used in conjunction with a device
2936 discovery system. As devices are found in a system, they can be
2938 .B "mdadm \-\-incremental"
2939 to be conditionally added to an appropriate array.
2941 Conversely, it can also be used with the
2943 flag to do just the opposite and find whatever array a particular device
2944 is part of and remove the device from that array.
2946 If the device passed is a
2948 device created by a previous call to
2950 then rather than trying to add that device to an array, all the arrays
2951 described by the metadata of the container will be started.
2954 performs a number of tests to determine if the device is part of an
2955 array, and which array it should be part of. If an appropriate array
2956 is found, or can be created,
2958 adds the device to the array and conditionally starts the array.
2962 will normally only add devices to an array which were previously working
2963 (active or spare) parts of that array. The support for automatic
2964 inclusion of a new drive as a spare in some array requires
2965 a configuration through POLICY in config file.
2969 makes are as follow:
2971 Is the device permitted by
2973 That is, is it listed in a
2975 line in that file. If
2977 is absent then the default it to allow any device. Similarly if
2979 contains the special word
2981 then any device is allowed. Otherwise the device name given to
2983 or one of the aliases given, or an alias found in the filesystem,
2984 must match one of the names or patterns in a
2988 This is the only context where the aliases are used. They are
2989 usually provided by a
2992 .BR $env{DEVLINKS} .
2995 Does the device have a valid md superblock? If a specific metadata
2996 version is requested with
3000 then only that style of metadata is accepted, otherwise
3002 finds any known version of metadata. If no
3004 metadata is found, the device may be still added to an array
3005 as a spare if POLICY allows.
3009 Does the metadata match an expected array?
3010 The metadata can match in two ways. Either there is an array listed
3013 which identifies the array (either by UUID, by name, by device list,
3014 or by minor-number), or the array was created with a
3020 or on the command line.
3023 is not able to positively identify the array as belonging to the
3024 current host, the device will be rejected.
3029 keeps a list of arrays that it has partially assembled in
3031 If no array exists which matches
3032 the metadata on the new device,
3034 must choose a device name and unit number. It does this based on any
3037 or any name information stored in the metadata. If this name
3038 suggests a unit number, that number will be used, otherwise a free
3039 unit number will be chosen. Normally
3041 will prefer to create a partitionable array, however if the
3045 suggests that a non-partitionable array is preferred, that will be
3048 If the array is not found in the config file and its metadata does not
3049 identify it as belonging to the "homehost", then
3051 will choose a name for the array which is certain not to conflict with
3052 any array which does belong to this host. It does this be adding an
3053 underscore and a small number to the name preferred by the metadata.
3055 Once an appropriate array is found or created and the device is added,
3057 must decide if the array is ready to be started. It will
3058 normally compare the number of available (non-spare) devices to the
3059 number of devices that the metadata suggests need to be active. If
3060 there are at least that many, the array will be started. This means
3061 that if any devices are missing the array will not be restarted.
3067 in which case the array will be run as soon as there are enough
3068 devices present for the data to be accessible. For a RAID1, that
3069 means one device will start the array. For a clean RAID5, the array
3070 will be started as soon as all but one drive is present.
3072 Note that neither of these approaches is really ideal. If it can
3073 be known that all device discovery has completed, then
3077 can be run which will try to start all arrays that are being
3078 incrementally assembled. They are started in "read-auto" mode in
3079 which they are read-only until the first write request. This means
3080 that no metadata updates are made and no attempt at resync or recovery
3081 happens. Further devices that are found before the first write can
3082 still be added safely.
3085 This section describes environment variables that affect how mdadm
3090 Setting this value to 1 will prevent mdadm from automatically launching
3091 mdmon. This variable is intended primarily for debugging mdadm/mdmon.
3097 does not create any device nodes in /dev, but leaves that task to
3101 appears not to be configured, or if this environment variable is set
3104 will create and devices that are needed.
3107 .B MDADM_NO_SYSTEMCTL
3112 is in use it will normally request
3114 to start various background tasks (particularly
3116 rather than forking and running them in the background. This can be
3117 suppressed by setting
3118 .BR MDADM_NO_SYSTEMCTL=1 .
3122 A key value of IMSM metadata is that it allows interoperability with
3123 boot ROMs on Intel platforms, and with other major operating systems.
3126 will only allow an IMSM array to be created or modified if detects
3127 that it is running on an Intel platform which supports IMSM, and
3128 supports the particular configuration of IMSM that is being requested
3129 (some functionality requires newer OROM support).
3131 These checks can be suppressed by setting IMSM_NO_PLATFORM=1 in the
3132 environment. This can be useful for testing or for disaster
3133 recovery. You should be aware that interoperability may be
3134 compromised by setting this value.
3137 .B MDADM_GROW_ALLOW_OLD
3138 If an array is stopped while it is performing a reshape and that
3139 reshape was making use of a backup file, then when the array is
3142 will sometimes complain that the backup file is too old. If this
3143 happens and you are certain it is the right backup file, you can
3144 over-ride this check by setting
3145 .B MDADM_GROW_ALLOW_OLD=1
3150 Any string given in this variable is added to the start of the
3152 line in the config file, or treated as the whole
3154 line if none is given. It can be used to disable certain metadata
3157 is called from a boot script. For example
3159 .B " export MDADM_CONF_AUTO='-ddf -imsm'
3163 does not automatically assemble any DDF or
3164 IMSM arrays that are found. This can be useful on systems configured
3165 to manage such arrays with
3171 .B " mdadm \-\-query /dev/name-of-device"
3173 This will find out if a given device is a RAID array, or is part of
3174 one, and will provide brief information about the device.
3176 .B " mdadm \-\-assemble \-\-scan"
3178 This will assemble and start all arrays listed in the standard config
3179 file. This command will typically go in a system startup file.
3181 .B " mdadm \-\-stop \-\-scan"
3183 This will shut down all arrays that can be shut down (i.e. are not
3184 currently in use). This will typically go in a system shutdown script.
3186 .B " mdadm \-\-follow \-\-scan \-\-delay=120"
3188 If (and only if) there is an Email address or program given in the
3189 standard config file, then
3190 monitor the status of all arrays listed in that file by
3191 polling them ever 2 minutes.
3193 .B " mdadm \-\-create /dev/md0 \-\-level=1 \-\-raid\-devices=2 /dev/hd[ac]1"
3195 Create /dev/md0 as a RAID1 array consisting of /dev/hda1 and /dev/hdc1.
3198 .B " echo 'DEVICE /dev/hd*[0\-9] /dev/sd*[0\-9]' > mdadm.conf"
3200 .B " mdadm \-\-detail \-\-scan >> mdadm.conf"
3202 This will create a prototype config file that describes currently
3203 active arrays that are known to be made from partitions of IDE or SCSI drives.
3204 This file should be reviewed before being used as it may
3205 contain unwanted detail.
3207 .B " echo 'DEVICE /dev/hd[a\-z] /dev/sd*[a\-z]' > mdadm.conf"
3209 .B " mdadm \-\-examine \-\-scan \-\-config=mdadm.conf >> mdadm.conf"
3211 This will find arrays which could be assembled from existing IDE and
3212 SCSI whole drives (not partitions), and store the information in the
3213 format of a config file.
3214 This file is very likely to contain unwanted detail, particularly
3217 entries. It should be reviewed and edited before being used as an
3220 .B " mdadm \-\-examine \-\-brief \-\-scan \-\-config=partitions"
3222 .B " mdadm \-Ebsc partitions"
3224 Create a list of devices by reading
3225 .BR /proc/partitions ,
3226 scan these for RAID superblocks, and printout a brief listing of all
3229 .B " mdadm \-Ac partitions \-m 0 /dev/md0"
3231 Scan all partitions and devices listed in
3232 .BR /proc/partitions
3235 out of all such devices with a RAID superblock with a minor number of 0.
3237 .B " mdadm \-\-monitor \-\-scan \-\-daemonise > /run/mdadm/mon.pid"
3239 If config file contains a mail address or alert program, run mdadm in
3240 the background in monitor mode monitoring all md devices. Also write
3241 pid of mdadm daemon to
3242 .BR /run/mdadm/mon.pid .
3244 .B " mdadm \-Iq /dev/somedevice"
3246 Try to incorporate newly discovered device into some array as
3249 .B " mdadm \-\-incremental \-\-rebuild\-map \-\-run \-\-scan"
3251 Rebuild the array map from any current arrays, and then start any that
3254 .B " mdadm /dev/md4 --fail detached --remove detached"
3256 Any devices which are components of /dev/md4 will be marked as faulty
3257 and then remove from the array.
3259 .B " mdadm --grow /dev/md4 --level=6 --backup-file=/root/backup-md4"
3263 which is currently a RAID5 array will be converted to RAID6. There
3264 should normally already be a spare drive attached to the array as a
3265 RAID6 needs one more drive than a matching RAID5.
3267 .B " mdadm --create /dev/md/ddf --metadata=ddf --raid-disks 6 /dev/sd[a-f]"
3269 Create a DDF array over 6 devices.
3271 .B " mdadm --create /dev/md/home -n3 -l5 -z 30000000 /dev/md/ddf"
3273 Create a RAID5 array over any 3 devices in the given DDF set. Use
3274 only 30 gigabytes of each device.
3276 .B " mdadm -A /dev/md/ddf1 /dev/sd[a-f]"
3278 Assemble a pre-exist ddf array.
3280 .B " mdadm -I /dev/md/ddf1"
3282 Assemble all arrays contained in the ddf array, assigning names as
3285 .B " mdadm \-\-create \-\-help"
3287 Provide help about the Create mode.
3289 .B " mdadm \-\-config \-\-help"
3291 Provide help about the format of the config file.
3293 .B " mdadm \-\-help"
3295 Provide general help.
3305 lists all active md devices with information about them.
3307 uses this to find arrays when
3309 is given in Misc mode, and to monitor array reconstruction
3314 The config file lists which devices may be scanned to see if
3315 they contain MD super block, and gives identifying information
3316 (e.g. UUID) about known MD arrays. See
3320 .SS /etc/mdadm.conf.d
3322 A directory containing configuration files which are read in lexical
3328 mode is used, this file gets a list of arrays currently being created.
3333 understand two sorts of names for array devices.
3335 The first is the so-called 'standard' format name, which matches the
3336 names used by the kernel and which appear in
3339 The second sort can be freely chosen, but must reside in
3341 When giving a device name to
3343 to create or assemble an array, either full path name such as
3347 can be given, or just the suffix of the second sort of name, such as
3353 chooses device names during auto-assembly or incremental assembly, it
3354 will sometimes add a small sequence number to the end of the name to
3355 avoid conflicted between multiple arrays that have the same name. If
3357 can reasonably determine that the array really is meant for this host,
3358 either by a hostname in the metadata, or by the presence of the array
3361 then it will leave off the suffix if possible.
3362 Also if the homehost is specified as
3365 will only use a suffix if a different array of the same name already
3366 exists or is listed in the config file.
3368 The standard names for non-partitioned arrays (the only sort of md
3369 array available in 2.4 and earlier) are of the form
3373 where NN is a number.
3374 The standard names for partitionable arrays (as available from 2.6
3375 onwards) are of the form:
3379 Partition numbers should be indicated by adding "pMM" to these, thus "/dev/md/d1p2".
3381 From kernel version 2.6.28 the "non-partitioned array" can actually
3382 be partitioned. So the "md_d\fBNN\fP"
3383 names are no longer needed, and
3384 partitions such as "/dev/md\fBNN\fPp\fBXX\fP"
3387 From kernel version 2.6.29 standard names can be non-numeric following
3394 is any string. These names are supported by
3396 since version 3.3 provided they are enabled in
3401 was previously known as
3405 For further information on mdadm usage, MD and the various levels of
3408 .B http://raid.wiki.kernel.org/
3410 (based upon Jakob \(/Ostergaard's Software\-RAID.HOWTO)
3412 The latest version of
3414 should always be available from
3416 .B http://www.kernel.org/pub/linux/utils/raid/mdadm/