2 .\" Copyright Neil Brown and others.
3 .\" This program is free software; you can redistribute it and/or modify
4 .\" it under the terms of the GNU General Public License as published by
5 .\" the Free Software Foundation; either version 2 of the License, or
6 .\" (at your option) any later version.
7 .\" See file COPYING in distribution for details.
10 mdadm \- manage MD devices
16 .BI mdadm " [mode] <raiddevice> [options] <component-devices>"
19 RAID devices are virtual devices created from two or more
20 real block devices. This allows multiple devices (typically disk
21 drives or partitions thereof) to be combined into a single device to
22 hold (for example) a single filesystem.
23 Some RAID levels include redundancy and so can survive some degree of
26 Linux Software RAID devices are implemented through the md (Multiple
27 Devices) device driver.
29 Currently, Linux supports
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 RAID
126 levels 1/4/5/6, changing the RAID level between 1, 5, and 6, changing
127 the chunk size and layout for RAID5 and RAID5, as well as adding or
128 removing a write-intent bitmap.
131 .B "Incremental Assembly"
132 Add a single device to an appropriate array. If the addition of the
133 device makes the array runnable, the array will be started.
134 This provides a convenient interface to a
136 system. As each device is detected,
138 has a chance to include it in some array as appropriate.
144 in this mode, then any arrays within that container will be assembled
149 This is for doing things to specific components of an array such as
150 adding new spares and removing faulty devices.
154 This is an 'everything else' mode that supports operations on active
155 arrays, operations on component devices such as erasing old superblocks, and
156 information gathering operations.
157 .\"This mode allows operations on independent devices such as examine MD
158 .\"superblocks, erasing old superblocks and stopping active arrays.
162 This mode does not act on a specific device or array, but rather it
163 requests the Linux Kernel to activate any auto-detected arrays.
166 .SH Options for selecting a mode are:
169 .BR \-A ", " \-\-assemble
170 Assemble a pre-existing array.
173 .BR \-B ", " \-\-build
174 Build a legacy array without superblocks.
177 .BR \-C ", " \-\-create
181 .BR \-F ", " \-\-follow ", " \-\-monitor
187 .BR \-G ", " \-\-grow
188 Change the size or shape of an active array.
191 .BR \-I ", " \-\-incremental
192 Add a single device into an appropriate array, and possibly start the array.
196 Request that the kernel starts any auto-detected arrays. This can only
199 is compiled into the kernel \(em not if it is a module.
200 Arrays can be auto-detected by the kernel if all the components are in
201 primary MS-DOS partitions with partition type
203 and all use v0.90 metadata.
204 In-kernel autodetect is not recommended for new installations. Using
206 to detect and assemble arrays \(em possibly in an
208 \(em is substantially more flexible and should be preferred.
211 If a device is given before any options, or if the first option is
216 then the MANAGE mode is assumed.
217 Anything other than these will cause the
221 .SH Options that are not mode-specific are:
224 .BR \-h ", " \-\-help
225 Display general help message or, after one of the above options, a
226 mode-specific help message.
230 Display more detailed help about command line parsing and some commonly
234 .BR \-V ", " \-\-version
235 Print version information for mdadm.
238 .BR \-v ", " \-\-verbose
239 Be more verbose about what is happening. This can be used twice to be
241 The extra verbosity currently only affects
242 .B \-\-detail \-\-scan
244 .BR "\-\-examine \-\-scan" .
247 .BR \-q ", " \-\-quiet
248 Avoid printing purely informative messages. With this,
250 will be silent unless there is something really important to report.
253 .BR \-f ", " \-\-force
254 Be more forceful about certain operations. See the various modes for
255 the exact meaning of this option in different contexts.
258 .BR \-c ", " \-\-config=
259 Specify the config file. Default is to use
260 .BR /etc/mdadm.conf ,
261 or if that is missing then
262 .BR /etc/mdadm/mdadm.conf .
263 If the config file given is
265 then nothing will be read, but
267 will act as though the config file contained exactly
268 .B "DEVICE partitions containers"
271 to find a list of devices to scan, and
273 to find a list of containers to examine.
276 is given for the config file, then
278 will act as though the config file were empty.
281 .BR \-s ", " \-\-scan
284 for missing information.
285 In general, this option gives
287 permission to get any missing information (like component devices,
288 array devices, array identities, and alert destination) from the
289 configuration file (see previous option);
290 one exception is MISC mode when using
296 says to get a list of array devices from
300 .BR \-e ", " \-\-metadata=
301 Declare the style of RAID metadata (superblock) to be used. The
304 and to guess for other operations.
305 The default can be overridden by setting the
315 Use the original 0.90 format superblock. This format limits arrays to
316 28 component devices and limits component devices of levels 1 and
317 greater to 2 terabytes.
318 .IP "1, 1.0, 1.1, 1.2 default"
319 Use the new version-1 format superblock. This has few restrictions.
320 The different sub-versions store the superblock at different locations
321 on the device, either at the end (for 1.0), at the start (for 1.1) or
322 4K from the start (for 1.2). '1' is equivalent to '1.0', 'default' is
325 Use the "Industry Standard" DDF (Disk Data Format) format defined by
327 When creating a DDF array a
329 will be created, and normal arrays can be created in that container.
331 Use the Intel(R) Matrix Storage Manager metadata format. This creates a
333 which is managed in a similar manner to DDF, and is supported by an
334 option-rom on some platforms:
336 .B http://www.intel.com/design/chipsets/matrixstorage_sb.htm
342 This will override any
344 setting in the config file and provides the identity of the host which
345 should be considered the home for any arrays.
347 When creating an array, the
349 will be recorded in the metadata. For version-1 superblocks, it will
350 be prefixed to the array name. For version-0.90 superblocks, part of
351 the SHA1 hash of the hostname will be stored in the later half of the
354 When reporting information about an array, any array which is tagged
355 for the given homehost will be reported as such.
357 When using Auto-Assemble, only arrays tagged for the given homehost
358 will be allowed to use 'local' names (i.e. not ending in '_' followed
359 by a digit string). See below under
360 .BR "Auto Assembly" .
362 .SH For create, build, or grow:
365 .BR \-n ", " \-\-raid\-devices=
366 Specify the number of active devices in the array. This, plus the
367 number of spare devices (see below) must equal the number of
369 (including "\fBmissing\fP" devices)
370 that are listed on the command line for
372 Setting a value of 1 is probably
373 a mistake and so requires that
375 be specified first. A value of 1 will then be allowed for linear,
376 multipath, RAID0 and RAID1. It is never allowed for RAID4, RAID5 or RAID6.
378 This number can only be changed using
380 for RAID1, RAID4, RAID5 and RAID6 arrays, and only on kernels which provide
381 the necessary support.
384 .BR \-x ", " \-\-spare\-devices=
385 Specify the number of spare (eXtra) devices in the initial array.
386 Spares can also be added
387 and removed later. The number of component devices listed
388 on the command line must equal the number of RAID devices plus the
389 number of spare devices.
392 .BR \-z ", " \-\-size=
393 Amount (in Kibibytes) of space to use from each drive in RAID levels 1/4/5/6.
394 This must be a multiple of the chunk size, and must leave about 128Kb
395 of space at the end of the drive for the RAID superblock.
396 If this is not specified
397 (as it normally is not) the smallest drive (or partition) sets the
398 size, though if there is a variance among the drives of greater than 1%, a warning is
401 This value can be set with
403 for RAID level 1/4/5/6. If the array was created with a size smaller
404 than the currently active drives, the extra space can be accessed
407 The size can be given as
409 which means to choose the largest size that fits on all current drives.
411 This value can not be used with
413 metadata such as DDF and IMSM.
416 .BR \-Z ", " \-\-array-size=
417 This is only meaningful with
419 and its effect is not persistent: when the array is stopped an
420 restarted the default array size will be restored.
422 Setting the array-size causes the array to appear smaller to programs
423 that access the data. This is particularly needed before reshaping an
424 array so that it will be smaller. As the reshape is not reversible,
425 but setting the size with
427 is, it is required that the array size is reduced as appropriate
428 before the number of devices in the array is reduced.
431 .BR \-c ", " \-\-chunk=
432 Specify chunk size of kibibytes. The default when creating an
433 array is 512KB. To ensure compatibility with earlier versions, the
434 default when Building and array with no persistent metadata is 64KB.
435 This is only meaningful for RAID0, RAID4, RAID5, RAID6, and RAID10.
439 Specify rounding factor for a Linear array. The size of each
440 component will be rounded down to a multiple of this size.
441 This is a synonym for
443 but highlights the different meaning for Linear as compared to other
444 RAID levels. The default is 64K if a kernel earlier than 2.6.16 is in
445 use, and is 0K (i.e. no rounding) in later kernels.
448 .BR \-l ", " \-\-level=
449 Set RAID level. When used with
451 options are: linear, raid0, 0, stripe, raid1, 1, mirror, raid4, 4,
452 raid5, 5, raid6, 6, raid10, 10, multipath, mp, faulty, container.
453 Obviously some of these are synonymous.
457 metadata type is requested, only the
459 level is permitted, and it does not need to be explicitly given.
463 only linear, stripe, raid0, 0, raid1, multipath, mp, and faulty are valid.
465 Not yet supported with
469 .BR \-p ", " \-\-layout=
470 This option configures the fine details of data layout for RAID5, RAID6,
471 and RAID10 arrays, and controls the failure modes for
474 The layout of the RAID5 parity block can be one of
475 .BR left\-asymmetric ,
476 .BR left\-symmetric ,
477 .BR right\-asymmetric ,
478 .BR right\-symmetric ,
479 .BR la ", " ra ", " ls ", " rs .
481 .BR left\-symmetric .
483 It is also possibly to cause RAID5 to use a RAID4-like layout by
489 Finally for RAID5 there are DDF\-compatible layouts,
490 .BR ddf\-zero\-restart ,
491 .BR ddf\-N\-restart ,
493 .BR ddf\-N\-continue .
495 These same layouts are available for RAID6. There are also 4 layouts
496 that will provide an intermediate stage for converting between RAID5
497 and RAID6. These provide a layout which is identical to the
498 corresponding RAID5 layout on the first N\-1 devices, and has the 'Q'
499 syndrome (the second 'parity' block used by RAID6) on the last device.
501 .BR left\-symmetric\-6 ,
502 .BR right\-symmetric\-6 ,
503 .BR left\-asymmetric\-6 ,
504 .BR right\-asymmetric\-6 ,
506 .BR parity\-first\-6 .
508 When setting the failure mode for level
511 .BR write\-transient ", " wt ,
512 .BR read\-transient ", " rt ,
513 .BR write\-persistent ", " wp ,
514 .BR read\-persistent ", " rp ,
516 .BR read\-fixable ", " rf ,
517 .BR clear ", " flush ", " none .
519 Each failure mode can be followed by a number, which is used as a period
520 between fault generation. Without a number, the fault is generated
521 once on the first relevant request. With a number, the fault will be
522 generated after that many requests, and will continue to be generated
523 every time the period elapses.
525 Multiple failure modes can be current simultaneously by using the
527 option to set subsequent failure modes.
529 "clear" or "none" will remove any pending or periodic failure modes,
530 and "flush" will clear any persistent faults.
532 Finally, the layout options for RAID10 are one of 'n', 'o' or 'f' followed
533 by a small number. The default is 'n2'. The supported options are:
536 signals 'near' copies. Multiple copies of one data block are at
537 similar offsets in different devices.
540 signals 'offset' copies. Rather than the chunks being duplicated
541 within a stripe, whole stripes are duplicated but are rotated by one
542 device so duplicate blocks are on different devices. Thus subsequent
543 copies of a block are in the next drive, and are one chunk further
548 (multiple copies have very different offsets).
549 See md(4) for more detail about 'near', 'offset', and 'far'.
551 The number is the number of copies of each datablock. 2 is normal, 3
552 can be useful. This number can be at most equal to the number of
553 devices in the array. It does not need to divide evenly into that
554 number (e.g. it is perfectly legal to have an 'n2' layout for an array
555 with an odd number of devices).
557 When an array is converted between RAID5 and RAID6 an intermediate
558 RAID6 layout is used in which the second parity block (Q) is always on
559 the last device. To convert a RAID5 to RAID6 and leave it in this new
560 layout (which does not require re-striping) use
561 .BR \-\-layout=preserve .
562 This will try to avoid any restriping.
564 The converse of this is
565 .B \-\-layout=normalise
566 which will change a non-standard RAID6 layout into a more standard
573 (thus explaining the p of
577 .BR \-b ", " \-\-bitmap=
578 Specify a file to store a write-intent bitmap in. The file should not
581 is also given. The same file should be provided
582 when assembling the array. If the word
584 is given, then the bitmap is stored with the metadata on the array,
585 and so is replicated on all devices. If the word
589 mode, then any bitmap that is present is removed.
591 To help catch typing errors, the filename must contain at least one
592 slash ('/') if it is a real file (not 'internal' or 'none').
594 Note: external bitmaps are only known to work on ext2 and ext3.
595 Storing bitmap files on other filesystems may result in serious problems.
598 .BR \-\-bitmap\-chunk=
599 Set the chunksize of the bitmap. Each bit corresponds to that many
600 Kilobytes of storage.
601 When using a file based bitmap, the default is to use the smallest
602 size that is at-least 4 and requires no more than 2^21 chunks.
605 bitmap, the chunksize defaults to 64Meg, or larger if necessary to
606 fit the bitmap into the available space.
609 .BR \-W ", " \-\-write\-mostly
610 subsequent devices listed in a
615 command will be flagged as 'write-mostly'. This is valid for RAID1
616 only and means that the 'md' driver will avoid reading from these
617 devices if at all possible. This can be useful if mirroring over a
621 .BR \-\-write\-behind=
622 Specify that write-behind mode should be enabled (valid for RAID1
623 only). If an argument is specified, it will set the maximum number
624 of outstanding writes allowed. The default value is 256.
625 A write-intent bitmap is required in order to use write-behind
626 mode, and write-behind is only attempted on drives marked as
630 .BR \-\-assume\-clean
633 that the array pre-existed and is known to be clean. It can be useful
634 when trying to recover from a major failure as you can be sure that no
635 data will be affected unless you actually write to the array. It can
636 also be used when creating a RAID1 or RAID10 if you want to avoid the
637 initial resync, however this practice \(em while normally safe \(em is not
638 recommended. Use this only if you really know what you are doing.
640 When the devices that will be part of a new array were filled
641 with zeros before creation the operator knows the array is
642 actually clean. If that is the case, such as after running
643 badblocks, this argument can be used to tell mdadm the
644 facts the operator knows.
647 .BR \-\-backup\-file=
650 is used to increase the number of
651 raid-devices in a RAID5 if there are no spare devices available.
652 See the GROW MODE section below on RAID\-DEVICES CHANGES. The file
653 should be stored on a separate device, not on the RAID array being
657 .BR \-\-array-size= ", " \-Z
658 Set the size of the array which is seen by users of the device such as
659 filesystems. This can be less that the real size, but never greater.
660 The size set this way does not persist across restarts of the array.
662 This is most useful when reducing the number of devices in a RAID5 or
663 RAID6. Such arrays require the array-size to be reduced before a
664 reshape can be performed that reduces the real size.
668 restores the apparent size of the array to be whatever the real
669 amount of available space is.
672 .BR \-N ", " \-\-name=
675 for the array. This is currently only effective when creating an
676 array with a version-1 superblock, or an array in a DDF container.
677 The name is a simple textual string that can be used to identify array
678 components when assembling. If name is needed but not specified, it
679 is taken from the basename of the device that is being created.
691 run the array, even if some of the components
692 appear to be active in another array or filesystem. Normally
694 will ask for confirmation before including such components in an
695 array. This option causes that question to be suppressed.
698 .BR \-f ", " \-\-force
701 accept the geometry and layout specified without question. Normally
703 will not allow creation of an array with only one device, and will try
704 to create a RAID5 array with one missing drive (as this makes the
705 initial resync work faster). With
708 will not try to be so clever.
711 .BR \-a ", " "\-\-auto{=yes,md,mdp,part,p}{NN}"
712 Instruct mdadm how to create the device file if needed, possibly allocating
713 an unused minor number. "md" causes a non-partitionable array
714 to be used (though since Linux 2.6.28, these array devices are in fact
715 partitionable). "mdp", "part" or "p" causes a partitionable array (2.6 and
716 later) to be used. "yes" requires the named md device to have
717 a 'standard' format, and the type and minor number will be determined
718 from this. With mdadm 3.0, device creation is normally left up to
720 so this option is unlikely to be needed.
721 See DEVICE NAMES below.
723 The argument can also come immediately after
728 is not given on the command line or in the config file, then
734 is also given, then any
736 entries in the config file will override the
738 instruction given on the command line.
740 For partitionable arrays,
742 will create the device file for the whole array and for the first 4
743 partitions. A different number of partitions can be specified at the
744 end of this option (e.g.
746 If the device name ends with a digit, the partition names add a 'p',
748 .IR /dev/md/home1p3 .
749 If there is no trailing digit, then the partition names just have a
751 .IR /dev/md/scratch3 .
753 If the md device name is in a 'standard' format as described in DEVICE
754 NAMES, then it will be created, if necessary, with the appropriate
755 device number based on that name. If the device name is not in one of these
756 formats, then a unused device number will be allocated. The device
757 number will be considered unused if there is no active array for that
758 number, and there is no entry in /dev for that number and with a
759 non-standard name. Names that are not in 'standard' format are only
760 allowed in "/dev/md/".
764 .\".BR \-\-symlink = no
769 .\"to create devices in
771 .\"it will also create symlinks from
773 .\"with names starting with
779 .\"to suppress this, or
780 .\".B \-\-symlink=yes
781 .\"to enforce this even if it is suppressing
789 .BR \-u ", " \-\-uuid=
790 uuid of array to assemble. Devices which don't have this uuid are
794 .BR \-m ", " \-\-super\-minor=
795 Minor number of device that array was created for. Devices which
796 don't have this minor number are excluded. If you create an array as
797 /dev/md1, then all superblocks will contain the minor number 1, even if
798 the array is later assembled as /dev/md2.
800 Giving the literal word "dev" for
804 to use the minor number of the md device that is being assembled.
807 .B \-\-super\-minor=dev
808 will look for super blocks with a minor number of 0.
811 is only relevant for v0.90 metadata, and should not normally be used.
817 .BR \-N ", " \-\-name=
818 Specify the name of the array to assemble. This must be the name
819 that was specified when creating the array. It must either match
820 the name stored in the superblock exactly, or it must match
823 prefixed to the start of the given name.
826 .BR \-f ", " \-\-force
827 Assemble the array even if the metadata on some devices appears to be
830 cannot find enough working devices to start the array, but can find
831 some devices that are recorded as having failed, then it will mark
832 those devices as working so that the array can be started.
833 An array which requires
835 to be started may contain data corruption. Use it carefully.
839 Attempt to start the array even if fewer drives were given than were
840 present last time the array was active. Normally if not all the
841 expected drives are found and
843 is not used, then the array will be assembled but not started.
846 an attempt will be made to start it anyway.
850 This is the reverse of
852 in that it inhibits the startup of array unless all expected drives
853 are present. This is only needed with
855 and can be used if the physical connections to devices are
856 not as reliable as you would like.
859 .BR \-a ", " "\-\-auto{=no,yes,md,mdp,part}"
860 See this option under Create and Build options.
863 .BR \-b ", " \-\-bitmap=
864 Specify the bitmap file that was given when the array was created. If
867 bitmap, there is no need to specify this when assembling the array.
870 .BR \-\-backup\-file=
873 was used to grow the number of raid-devices in a RAID5, and the system
874 crashed during the critical section, then the same
878 to allow possibly corrupted data to be restored.
881 .BR \-U ", " \-\-update=
882 Update the superblock on each device while assembling the array. The
883 argument given to this flag can be one of
897 option will adjust the superblock of an array what was created on a Sparc
898 machine running a patched 2.2 Linux kernel. This kernel got the
899 alignment of part of the superblock wrong. You can use the
900 .B "\-\-examine \-\-sparc2.2"
903 to see what effect this would have.
907 option will update the
909 field on each superblock to match the minor number of the array being
911 This can be useful if
913 reports a different "Preferred Minor" to
915 In some cases this update will be performed automatically
916 by the kernel driver. In particular the update happens automatically
917 at the first write to an array with redundancy (RAID level 1 or
918 greater) on a 2.6 (or later) kernel.
922 option will change the uuid of the array. If a UUID is given with the
924 option that UUID will be used as a new UUID and will
926 be used to help identify the devices in the array.
929 is given, a random UUID is chosen.
933 option will change the
935 of the array as stored in the superblock. This is only supported for
936 version-1 superblocks.
940 option will change the
942 as recorded in the superblock. For version-0 superblocks, this is the
943 same as updating the UUID.
944 For version-1 superblocks, this involves updating the name.
948 option will cause the array to be marked
950 meaning that any redundancy in the array (e.g. parity for RAID5,
951 copies for RAID1) may be incorrect. This will cause the RAID system
952 to perform a "resync" pass to make sure that all redundant information
957 option allows arrays to be moved between machines with different
959 When assembling such an array for the first time after a move, giving
960 .B "\-\-update=byteorder"
963 to expect superblocks to have their byteorder reversed, and will
964 correct that order before assembling the array. This is only valid
965 with original (Version 0.90) superblocks.
969 option will correct the summaries in the superblock. That is the
970 counts of total, working, active, failed, and spare devices.
974 will rarely be of use. It applies to version 1.1 and 1.2 metadata
975 only (where the metadata is at the start of the device) and is only
976 useful when the component device has changed size (typically become
977 larger). The version 1 metadata records the amount of the device that
978 can be used to store data, so if a device in a version 1.1 or 1.2
979 array becomes larger, the metadata will still be visible, but the
980 extra space will not. In this case it might be useful to assemble the
982 .BR \-\-update=devicesize .
985 to determine the maximum usable amount of space on each device and
986 update the relevant field in the metadata.
990 .B \-\-auto\-update\-homehost
991 This flag is only meaningful with auto-assembly (see discussion below).
992 In that situation, if no suitable arrays are found for this homehost,
994 will rescan for any arrays at all and will assemble them and update the
995 homehost to match the current host.
1001 .BR \-a ", " \-\-add
1002 hot-add listed devices. For arrays with redundancy, the listed
1003 devices become available as spares. If the array is degraded, it will
1004 immediately start recovering data on to one of these spares.
1008 re-add a device that was recently removed from an array. This is only
1009 needed for arrays that have be built (i.e. with
1011 For created arrays, devices are always re-added if that is possible.
1012 When re-adding a device, if nothing has changed on the array since the
1013 device was removed, no recovery is performed. Also, if the array has
1014 a write-intent bitmap, then the recovery performed after a re-add will
1015 be limited to those blocks which, according to the bitmap, might have
1016 changed since the device was removed.
1019 .BR \-r ", " \-\-remove
1020 remove listed devices. They must not be active. i.e. they should
1021 be failed or spare devices. As well as the name of a device file
1030 The first causes all failed device to be removed. The second causes
1031 any device which is no longer connected to the system (i.e an 'open'
1034 to be removed. This will only succeed for devices that are spares or
1035 have already been marked as failed.
1038 .BR \-f ", " \-\-fail
1039 mark listed devices as faulty.
1040 As well as the name of a device file, the word
1042 can be given. This will cause any device that has been detached from
1043 the system to be marked as failed. It can then be removed.
1051 .BR \-\-write\-mostly
1052 Subsequent devices that are added or re-added will have the 'write-mostly'
1053 flag set. This is only valid for RAID1 and means that the 'md' driver
1054 will avoid reading from these devices if possible.
1057 Subsequent devices that are added or re-added will have the 'write-mostly'
1061 Each of these options requires that the first device listed is the array
1062 to be acted upon, and the remainder are component devices to be added,
1063 removed, marked as faulty, etc. Several different operations can be
1064 specified for different devices, e.g.
1066 mdadm /dev/md0 \-\-add /dev/sda1 \-\-fail /dev/sdb1 \-\-remove /dev/sdb1
1068 Each operation applies to all devices listed until the next
1071 If an array is using a write-intent bitmap, then devices which have
1072 been removed can be re-added in a way that avoids a full
1073 reconstruction but instead just updates the blocks that have changed
1074 since the device was removed. For arrays with persistent metadata
1075 (superblocks) this is done automatically. For arrays created with
1077 mdadm needs to be told that this device we removed recently with
1080 Devices can only be removed from an array if they are not in active
1081 use, i.e. that must be spares or failed devices. To remove an active
1082 device, it must first be marked as
1088 .BR \-Q ", " \-\-query
1089 Examine a device to see
1090 (1) if it is an md device and (2) if it is a component of an md
1092 Information about what is discovered is presented.
1095 .BR \-D ", " \-\-detail
1096 Print details of one or more md devices.
1099 .BR \-\-detail\-platform
1100 Print details of the platform's RAID capabilities (firmware / hardware
1101 topology) for a given metadata format.
1104 .BR \-Y ", " \-\-export
1109 output will be formatted as
1111 pairs for easy import into the environment.
1114 .BR \-E ", " \-\-examine
1115 Print contents of the metadata stored on the named device(s).
1116 Note the contrast between
1121 applies to devices which are components of an array, while
1123 applies to a whole array which is currently active.
1126 If an array was created on a SPARC machine with a 2.2 Linux kernel
1127 patched with RAID support, the superblock will have been created
1128 incorrectly, or at least incompatibly with 2.4 and later kernels.
1133 will fix the superblock before displaying it. If this appears to do
1134 the right thing, then the array can be successfully assembled using
1135 .BR "\-\-assemble \-\-update=sparc2.2" .
1138 .BR \-X ", " \-\-examine\-bitmap
1139 Report information about a bitmap file.
1140 The argument is either an external bitmap file or an array component
1141 in case of an internal bitmap. Note that running this on an array
1144 does not report the bitmap for that array.
1147 .BR \-R ", " \-\-run
1148 start a partially assembled array. If
1150 did not find enough devices to fully start the array, it might leaving
1151 it partially assembled. If you wish, you can then use
1153 to start the array in degraded mode.
1156 .BR \-S ", " \-\-stop
1157 deactivate array, releasing all resources.
1160 .BR \-o ", " \-\-readonly
1161 mark array as readonly.
1164 .BR \-w ", " \-\-readwrite
1165 mark array as readwrite.
1168 .B \-\-zero\-superblock
1169 If the device contains a valid md superblock, the block is
1170 overwritten with zeros. With
1172 the block where the superblock would be is overwritten even if it
1173 doesn't appear to be valid.
1176 .B \-\-kill\-subarray=
1177 If the device is a container and the argument to \-\-kill\-subarray
1178 specifies an inactive subarray in the container, then the subarray is
1179 deleted. Deleting all subarrays will leave an 'empty-container' or
1180 spare superblock on the drives. See \-\-zero\-superblock for completely
1181 removing a superblock. Note that some formats depend on the subarray
1182 index for generating a UUID, this command will fail if it would change
1183 the UUID of an active subarray.
1186 .B \-\-update\-subarray=
1187 If the device is a container and the argument to \-\-update\-subarray
1188 specifies a subarray in the container, then attempt to update the given
1189 superblock field in the subarray. See below in
1194 .BR \-t ", " \-\-test
1199 is set to reflect the status of the device. See below in
1204 .BR \-W ", " \-\-wait
1205 For each md device given, wait for any resync, recovery, or reshape
1206 activity to finish before returning.
1208 will return with success if it actually waited for every device
1209 listed, otherwise it will return failure.
1213 For each md device given, or each device in /proc/mdstat if
1215 is given, arrange for the array to be marked clean as soon as possible.
1216 Also, quiesce resync so that the monitor for external metadata arrays
1217 (mdmon) has an opportunity to checkpoint the resync position.
1219 will return with success if the array uses external metadata and we
1220 successfully waited. For native arrays this returns immediately as the
1221 kernel handles both dirty-clean transitions and resync checkpointing in
1222 the kernel at shutdown. No action is taken if safe-mode handling is
1225 .SH For Incremental Assembly mode:
1227 .BR \-\-rebuild\-map ", " \-r
1228 Rebuild the map file
1229 .RB ( /var/run/mdadm/map )
1232 uses to help track which arrays are currently being assembled.
1235 .BR \-\-run ", " \-R
1236 Run any array assembled as soon as a minimal number of devices are
1237 available, rather than waiting until all expected devices are present.
1241 This allows the hot-plug system to prevent arrays from running when it knows
1242 that more disks may arrive later in the discovery process.
1245 .BR \-\-scan ", " \-s
1246 Only meaningful with
1250 file for arrays that are being incrementally assembled and will try to
1251 start any that are not already started. If any such array is listed
1254 as requiring an external bitmap, that bitmap will be attached first.
1256 .SH For Monitor mode:
1258 .BR \-m ", " \-\-mail
1259 Give a mail address to send alerts to.
1262 .BR \-p ", " \-\-program ", " \-\-alert
1263 Give a program to be run whenever an event is detected.
1266 .BR \-y ", " \-\-syslog
1267 Cause all events to be reported through 'syslog'. The messages have
1268 facility of 'daemon' and varying priorities.
1271 .BR \-d ", " \-\-delay
1272 Give a delay in seconds.
1274 polls the md arrays and then waits this many seconds before polling
1275 again. The default is 60 seconds. Since 2.6.16, there is no need to
1276 reduce this as the kernel alerts
1278 immediately when there is any change.
1281 .BR \-r ", " \-\-increment
1282 Give a percentage increment.
1284 will generate RebuildNN events with the given percentage increment.
1287 .BR \-f ", " \-\-daemonise
1290 to run as a background daemon if it decides to monitor anything. This
1291 causes it to fork and run in the child, and to disconnect from the
1292 terminal. The process id of the child is written to stdout.
1295 which will only continue monitoring if a mail address or alert program
1296 is found in the config file.
1299 .BR \-i ", " \-\-pid\-file
1302 is running in daemon mode, write the pid of the daemon process to
1303 the specified file, instead of printing it on standard output.
1306 .BR \-1 ", " \-\-oneshot
1307 Check arrays only once. This will generate
1309 events and more significantly
1315 .B " mdadm \-\-monitor \-\-scan \-1"
1317 from a cron script will ensure regular notification of any degraded arrays.
1320 .BR \-t ", " \-\-test
1323 alert for every array found at startup. This alert gets mailed and
1324 passed to the alert program. This can be used for testing that alert
1325 message do get through successfully.
1331 .B mdadm \-\-assemble
1332 .I md-device options-and-component-devices...
1335 .B mdadm \-\-assemble \-\-scan
1336 .I md-devices-and-options...
1339 .B mdadm \-\-assemble \-\-scan
1343 This usage assembles one or more RAID arrays from pre-existing components.
1344 For each array, mdadm needs to know the md device, the identity of the
1345 array, and a number of component-devices. These can be found in a number of ways.
1347 In the first usage example (without the
1349 the first device given is the md device.
1350 In the second usage example, all devices listed are treated as md
1351 devices and assembly is attempted.
1352 In the third (where no devices are listed) all md devices that are
1353 listed in the configuration file are assembled. If not arrays are
1354 described by the configuration file, then any arrays that
1355 can be found on unused devices will be assembled.
1357 If precisely one device is listed, but
1363 was given and identity information is extracted from the configuration file.
1365 The identity can be given with the
1371 option, will be taken from the md-device record in the config file, or
1372 will be taken from the super block of the first component-device
1373 listed on the command line.
1375 Devices can be given on the
1377 command line or in the config file. Only devices which have an md
1378 superblock which contains the right identity will be considered for
1381 The config file is only used if explicitly named with
1383 or requested with (a possibly implicit)
1388 .B /etc/mdadm/mdadm.conf
1393 is not given, then the config file will only be used to find the
1394 identity of md arrays.
1396 Normally the array will be started after it is assembled. However if
1398 is not given and not all expected drives were listed, then the array
1399 is not started (to guard against usage errors). To insist that the
1400 array be started in this case (as may work for RAID1, 4, 5, 6, or 10),
1409 does not create any entries in
1413 It does record information in
1414 .B /var/run/mdadm/map
1417 to choose the correct name.
1421 detects that udev is not configured, it will create the devices in
1425 In Linux kernels prior to version 2.6.28 there were two distinctly
1426 different types of md devices that could be created: one that could be
1427 partitioned using standard partitioning tools and one that could not.
1428 Since 2.6.28 that distinction is no longer relevant as both type of
1429 devices can be partitioned.
1431 will normally create the type that originally could not be partitioned
1432 as it has a well defined major number (9).
1434 Prior to 2.6.28, it is important that mdadm chooses the correct type
1435 of array device to use. This can be controlled with the
1437 option. In particular, a value of "mdp" or "part" or "p" tells mdadm
1438 to use a partitionable device rather than the default.
1440 In the no-udev case, the value given to
1442 can be suffixed by a number. This tells
1444 to create that number of partition devices rather than the default of 4.
1448 can also be given in the configuration file as a word starting
1450 on the ARRAY line for the relevant array.
1457 and no devices are listed,
1459 will first attempt to assemble all the arrays listed in the config
1462 In no array at listed in the config (other than those marked
1464 it will look through the available devices for possible arrays and
1465 will try to assemble anything that it finds. Arrays which are tagged
1466 as belonging to the given homehost will be assembled and started
1467 normally. Arrays which do not obviously belong to this host are given
1468 names that are expected not to conflict with anything local, and are
1469 started "read-auto" so that nothing is written to any device until the
1470 array is written to. i.e. automatic resync etc is delayed.
1474 finds a consistent set of devices that look like they should comprise
1475 an array, and if the superblock is tagged as belonging to the given
1476 home host, it will automatically choose a device name and try to
1477 assemble the array. If the array uses version-0.90 metadata, then the
1479 number as recorded in the superblock is used to create a name in
1483 If the array uses version-1 metadata, then the
1485 from the superblock is used to similarly create a name in
1487 (the name will have any 'host' prefix stripped first).
1489 This behaviour can be modified by the
1493 configuration file. This line can indicate that specific metadata
1494 type should, or should not, be automatically assembled. If an array
1495 is found which is not listed in
1497 and has a metadata format that is denied by the
1499 line, then it will not be assembled.
1502 line can also request that all arrays identified as being for this
1503 homehost should be assembled regardless of their metadata type.
1506 for further details.
1511 cannot find any array for the given host at all, and if
1512 .B \-\-auto\-update\-homehost
1515 will search again for any array (not just an array created for this
1516 host) and will assemble each assuming
1517 .BR \-\-update=homehost .
1518 This will change the host tag in the superblock so that on the next run,
1519 these arrays will be found without the second pass. The intention of
1520 this feature is to support transitioning a set of md arrays to using
1523 The reason for requiring arrays to be tagged with the homehost for
1524 auto assembly is to guard against problems that can arise when moving
1525 devices from one host to another.
1536 .BI \-\-raid\-devices= Z
1540 This usage is similar to
1542 The difference is that it creates an array without a superblock. With
1543 these arrays there is no difference between initially creating the array and
1544 subsequently assembling the array, except that hopefully there is useful
1545 data there in the second case.
1547 The level may raid0, linear, raid1, raid10, multipath, or faulty, or
1548 one of their synonyms. All devices must be listed and the array will
1549 be started once complete. It will often be appropriate to use
1550 .B \-\-assume\-clean
1551 with levels raid1 or raid10.
1562 .BI \-\-raid\-devices= Z
1566 This usage will initialise a new md array, associate some devices with
1567 it, and activate the array.
1569 The named device will normally not exist when
1570 .I "mdadm \-\-create"
1571 is run, but will be created by
1573 once the array becomes active.
1575 As devices are added, they are checked to see if they contain RAID
1576 superblocks or filesystems. They are also checked to see if the variance in
1577 device size exceeds 1%.
1579 If any discrepancy is found, the array will not automatically be run, though
1582 can override this caution.
1584 To create a "degraded" array in which some devices are missing, simply
1585 give the word "\fBmissing\fP"
1586 in place of a device name. This will cause
1588 to leave the corresponding slot in the array empty.
1589 For a RAID4 or RAID5 array at most one slot can be
1590 "\fBmissing\fP"; for a RAID6 array at most two slots.
1591 For a RAID1 array, only one real device needs to be given. All of the
1595 When creating a RAID5 array,
1597 will automatically create a degraded array with an extra spare drive.
1598 This is because building the spare into a degraded array is in general
1599 faster than resyncing the parity on a non-degraded, but not clean,
1600 array. This feature can be overridden with the
1604 When creating an array with version-1 metadata a name for the array is
1606 If this is not given with the
1610 will choose a name based on the last component of the name of the
1611 device being created. So if
1613 is being created, then the name
1618 is being created, then the name
1622 When creating a partition based array, using
1624 with version-1.x metadata, the partition type should be set to
1626 (non fs-data). This type selection allows for greater precision since
1627 using any other [RAID auto-detect (0xFD) or a GNU/Linux partition (0x83)],
1628 might create problems in the event of array recovery through a live cdrom.
1630 A new array will normally get a randomly assigned 128bit UUID which is
1631 very likely to be unique. If you have a specific need, you can choose
1632 a UUID for the array by giving the
1634 option. Be warned that creating two arrays with the same UUID is a
1635 recipe for disaster. Also, using
1637 when creating a v0.90 array will silently override any
1642 .\"option is given, it is not necessary to list any component-devices in this command.
1643 .\"They can be added later, before a
1647 .\"is given, the apparent size of the smallest drive given is used.
1649 When creating an array within a
1652 can be given either the list of devices to use, or simply the name of
1653 the container. The former case gives control over which devices in
1654 the container will be used for the array. The latter case allows
1656 to automatically choose which devices to use based on how much spare
1659 The General Management options that are valid with
1664 insist on running the array even if some devices look like they might
1669 start the array readonly \(em not supported yet.
1676 .I options... devices...
1679 This usage will allow individual devices in an array to be failed,
1680 removed or added. It is possible to perform multiple operations with
1681 on command. For example:
1683 .B " mdadm /dev/md0 \-f /dev/hda1 \-r /dev/hda1 \-a /dev/hda1"
1689 and will then remove it from the array and finally add it back
1690 in as a spare. However only one md array can be affected by a single
1693 When a device is added to an active array, mdadm checks to see if it
1694 has metadata on it which suggests that it was recently a member of the
1695 array. If it does, it tried to "re-add" the device. If there have
1696 been no changes since the device was removed, or if the array has a
1697 write-intent bitmap which has recorded whatever changes there were,
1698 then the device will immediately become a full member of the array and
1699 those differences recorded in the bitmap will be resolved.
1709 MISC mode includes a number of distinct operations that
1710 operate on distinct devices. The operations are:
1713 The device is examined to see if it is
1714 (1) an active md array, or
1715 (2) a component of an md array.
1716 The information discovered is reported.
1720 The device should be an active md device.
1722 will display a detailed description of the array.
1726 will cause the output to be less detailed and the format to be
1727 suitable for inclusion in
1728 .BR /etc/mdadm.conf .
1731 will normally be 0 unless
1733 failed to get useful information about the device(s); however, if the
1735 option is given, then the exit status will be:
1739 The array is functioning normally.
1742 The array has at least one failed device.
1745 The array has multiple failed devices such that it is unusable.
1748 There was an error while trying to get information about the device.
1752 .B \-\-detail\-platform
1753 Print detail of the platform's RAID capabilities (firmware / hardware
1754 topology). If the metadata is specified with
1758 then the return status will be:
1762 metadata successfully enumerated its platform components on this system
1765 metadata is platform independent
1768 metadata failed to find its platform components on this system
1772 .B \-\-update\-subarray=
1773 If the device is a container and the argument to \-\-update\-subarray
1774 specifies a subarray in the container, then attempt to update the given
1775 superblock field in the subarray. Similar to updating an array in
1776 "assemble" mode, the field to update is selected by
1780 option. Currently only
1786 option updates the subarray name in the metadata, it may not affect the
1787 device node name or the device node symlink until the subarray is
1788 re\-assembled. If updating
1790 would change the UUID of an active subarray this operation is blocked,
1791 and the command will end in an error.
1795 The device should be a component of an md array.
1797 will read the md superblock of the device and display the contents.
1802 is given, then multiple devices that are components of the one array
1803 are grouped together and reported in a single entry suitable
1805 .BR /etc/mdadm.conf .
1809 without listing any devices will cause all devices listed in the
1810 config file to be examined.
1814 The devices should be active md arrays which will be deactivated, as
1815 long as they are not currently in use.
1819 This will fully activate a partially assembled md array.
1823 This will mark an active array as read-only, providing that it is
1824 not currently being used.
1830 array back to being read/write.
1834 For all operations except
1837 will cause the operation to be applied to all arrays listed in
1842 causes all devices listed in the config file to be examined.
1845 .BR \-b ", " \-\-brief
1846 Be less verbose. This is used with
1854 gives an intermediate level of verbosity.
1860 .B mdadm \-\-monitor
1861 .I options... devices...
1866 to periodically poll a number of md arrays and to report on any events
1869 will never exit once it decides that there are arrays to be checked,
1870 so it should normally be run in the background.
1872 As well as reporting events,
1874 may move a spare drive from one array to another if they are in the
1877 and if the destination array has a failed drive but no spares.
1879 If any devices are listed on the command line,
1881 will only monitor those devices. Otherwise all arrays listed in the
1882 configuration file will be monitored. Further, if
1884 is given, then any other md devices that appear in
1886 will also be monitored.
1888 The result of monitoring the arrays is the generation of events.
1889 These events are passed to a separate program (if specified) and may
1890 be mailed to a given E-mail address.
1892 When passing events to a program, the program is run once for each event,
1893 and is given 2 or 3 command-line arguments: the first is the
1894 name of the event (see below), the second is the name of the
1895 md device which is affected, and the third is the name of a related
1896 device if relevant (such as a component device that has failed).
1900 is given, then a program or an E-mail address must be specified on the
1901 command line or in the config file. If neither are available, then
1903 will not monitor anything.
1907 will continue monitoring as long as something was found to monitor. If
1908 no program or email is given, then each event is reported to
1911 The different events are:
1915 .B DeviceDisappeared
1916 An md array which previously was configured appears to no longer be
1917 configured. (syslog priority: Critical)
1921 was told to monitor an array which is RAID0 or Linear, then it will
1923 .B DeviceDisappeared
1924 with the extra information
1926 This is because RAID0 and Linear do not support the device-failed,
1927 hot-spare and resync operations which are monitored.
1931 An md array started reconstruction. (syslog priority: Warning)
1937 is a two-digit number (ie. 05, 48). This indicates that rebuild
1938 has passed that many percent of the total. The events are generated
1939 with fixed increment since 0. Increment size may be specified with
1940 a commandline option (default is 20). (syslog priority: Warning)
1944 An md array that was rebuilding, isn't any more, either because it
1945 finished normally or was aborted. (syslog priority: Warning)
1949 An active component device of an array has been marked as
1950 faulty. (syslog priority: Critical)
1954 A spare component device which was being rebuilt to replace a faulty
1955 device has failed. (syslog priority: Critical)
1959 A spare component device which was being rebuilt to replace a faulty
1960 device has been successfully rebuilt and has been made active.
1961 (syslog priority: Info)
1965 A new md array has been detected in the
1967 file. (syslog priority: Info)
1971 A newly noticed array appears to be degraded. This message is not
1974 notices a drive failure which causes degradation, but only when
1976 notices that an array is degraded when it first sees the array.
1977 (syslog priority: Critical)
1981 A spare drive has been moved from one array in a
1983 to another to allow a failed drive to be replaced.
1984 (syslog priority: Info)
1990 has been told, via the config file, that an array should have a certain
1991 number of spare devices, and
1993 detects that it has fewer than this number when it first sees the
1994 array, it will report a
1997 (syslog priority: Warning)
2001 An array was found at startup, and the
2004 (syslog priority: Info)
2014 cause Email to be sent. All events cause the program to be run.
2015 The program is run with two or three arguments: the event
2016 name, the array device and possibly a second device.
2018 Each event has an associated array device (e.g.
2020 and possibly a second device. For
2025 the second device is the relevant component device.
2028 the second device is the array that the spare was moved from.
2032 to move spares from one array to another, the different arrays need to
2033 be labeled with the same
2035 in the configuration file. The
2037 name can be any string; it is only necessary that different spare
2038 groups use different names.
2042 detects that an array in a spare group has fewer active
2043 devices than necessary for the complete array, and has no spare
2044 devices, it will look for another array in the same spare group that
2045 has a full complement of working drive and a spare. It will then
2046 attempt to remove the spare from the second drive and add it to the
2048 If the removal succeeds but the adding fails, then it is added back to
2052 The GROW mode is used for changing the size or shape of an active
2054 For this to work, the kernel must support the necessary change.
2055 Various types of growth are being added during 2.6 development,
2056 including restructuring a RAID5 array to have more active devices.
2058 Currently the only support available is to
2060 change the "size" attribute
2061 for RAID1, RAID5 and RAID6.
2063 increase or decrease the "raid\-devices" attribute of RAID1, RAID5,
2066 change the chunk-size and layout of RAID5 and RAID6.
2068 convert between RAID1 and RAID5, and between RAID5 and RAID6.
2070 add a write-intent bitmap to any array which supports these bitmaps, or
2071 remove a write-intent bitmap from such an array.
2074 GROW mode is not currently supported for
2076 or arrays inside containers.
2079 Normally when an array is built the "size" it taken from the smallest
2080 of the drives. If all the small drives in an arrays are, one at a
2081 time, removed and replaced with larger drives, then you could have an
2082 array of large drives with only a small amount used. In this
2083 situation, changing the "size" with "GROW" mode will allow the extra
2084 space to start being used. If the size is increased in this way, a
2085 "resync" process will start to make sure the new parts of the array
2088 Note that when an array changes size, any filesystem that may be
2089 stored in the array will not automatically grow to use the space. The
2090 filesystem will need to be explicitly told to use the extra space.
2092 Also the size of an array cannot be changed while it has an active
2093 bitmap. If an array has a bitmap, it must be removed before the size
2094 can be changed. Once the change it complete a new bitmap can be created.
2096 .SS RAID\-DEVICES CHANGES
2098 A RAID1 array can work with any number of devices from 1 upwards
2099 (though 1 is not very useful). There may be times which you want to
2100 increase or decrease the number of active devices. Note that this is
2101 different to hot-add or hot-remove which changes the number of
2104 When reducing the number of devices in a RAID1 array, the slots which
2105 are to be removed from the array must already be vacant. That is, the
2106 devices which were in those slots must be failed and removed.
2108 When the number of devices is increased, any hot spares that are
2109 present will be activated immediately.
2111 Changing the number of active devices in a RAID5 or RAID6 is much more
2112 effort. Every block in the array will need to be read and written
2113 back to a new location. From 2.6.17, the Linux Kernel is able to
2114 increase the number of devices in a RAID5 safely, including restarting
2115 an interrupted "reshape". From 2.6.31, the Linux Kernel is able to
2116 increase or decrease the number of devices in a RAID5 or RAID6.
2118 When decreasing the number of devices, the size of the array will also
2119 decrease. If there was data in the array, it could get destroyed and
2120 this is not reversible. To help prevent accidents,
2122 requires that the size of the array be decreased first with
2123 .BR "mdadm --grow --array-size" .
2124 This is a reversible change which simply makes the end of the array
2125 inaccessible. The integrity of any data can then be checked before
2126 the non-reversible reduction in the number of devices is request.
2128 When relocating the first few stripes on a RAID5, it is not possible
2129 to keep the data on disk completely consistent and crash-proof. To
2130 provide the required safety, mdadm disables writes to the array while
2131 this "critical section" is reshaped, and takes a backup of the data
2132 that is in that section. This backup is normally stored in any spare
2133 devices that the array has, however it can also be stored in a
2134 separate file specified with the
2136 option. If this option is used, and the system does crash during the
2137 critical period, the same file must be passed to
2139 to restore the backup and reassemble the array.
2143 Changing the RAID level of any array happens instantaneously. However
2144 in the RAID to RAID6 case this requires a non-standard layout of the
2145 RAID6 data, and in the RAID6 to RAID5 case that non-standard layout is
2146 required before the change can be accomplish. So while the level
2147 change is instant, the accompanying layout change can take quite a
2150 .SS CHUNK-SIZE AND LAYOUT CHANGES
2152 Changing the chunk-size of layout without also changing the number of
2153 devices as the same time will involve re-writing all blocks in-place.
2154 To ensure against data loss in the case of a crash, a
2156 must be provided for these changes. Small sections of the array will
2157 be copied to the backup file while they are being rearranged.
2159 If the reshape is interrupted for any reason, this backup file must be
2161 .B "mdadm --assemble"
2162 so the array can be reassembled. Consequently the file cannot be
2163 stored on the device being reshaped.
2168 A write-intent bitmap can be added to, or removed from, an active
2169 array. Either internal bitmaps, or bitmaps stored in a separate file,
2170 can be added. Note that if you add a bitmap stored in a file which is
2171 in a filesystem that is on the RAID array being affected, the system
2172 will deadlock. The bitmap must be on a separate filesystem.
2174 .SH INCREMENTAL MODE
2178 .B mdadm \-\-incremental
2184 .B mdadm \-\-incremental \-\-rebuild
2187 .B mdadm \-\-incremental \-\-run \-\-scan
2190 This mode is designed to be used in conjunction with a device
2191 discovery system. As devices are found in a system, they can be
2193 .B "mdadm \-\-incremental"
2194 to be conditionally added to an appropriate array.
2196 If the device passed is a
2198 device created by a previous call to
2200 then rather than trying to add that device to an array, all the arrays
2201 described by the metadata of the container will be started.
2204 performs a number of tests to determine if the device is part of an
2205 array, and which array it should be part of. If an appropriate array
2206 is found, or can be created,
2208 adds the device to the array and conditionally starts the array.
2212 will only add devices to an array which were previously working
2213 (active or spare) parts of that array. It does not currently support
2214 automatic inclusion of a new drive as a spare in some array.
2218 makes are as follow:
2220 Is the device permitted by
2222 That is, is it listed in a
2224 line in that file. If
2226 is absent then the default it to allow any device. Similar if
2228 contains the special word
2230 then any device is allowed. Otherwise the device name given to
2232 must match one of the names or patterns in a
2237 Does the device have a valid md superblock. If a specific metadata
2238 version is request with
2242 then only that style of metadata is accepted, otherwise
2244 finds any known version of metadata. If no
2246 metadata is found, the device is rejected.
2250 Does the metadata match an expected array?
2251 The metadata can match in two ways. Either there is an array listed
2254 which identifies the array (either by UUID, by name, by device list,
2255 or by minor-number), or the array was created with a
2261 or on the command line.
2264 is not able to positively identify the array as belonging to the
2265 current host, the device will be rejected.
2269 keeps a list of arrays that it has partially assembled in
2270 .B /var/run/mdadm/map
2272 .B /var/run/mdadm.map
2273 if the directory doesn't exist. Or maybe even
2274 .BR /dev/.mdadm.map ).
2275 If no array exists which matches
2276 the metadata on the new device,
2278 must choose a device name and unit number. It does this based on any
2281 or any name information stored in the metadata. If this name
2282 suggests a unit number, that number will be used, otherwise a free
2283 unit number will be chosen. Normally
2285 will prefer to create a partitionable array, however if the
2289 suggests that a non-partitionable array is preferred, that will be
2292 If the array is not found in the config file and its metadata does not
2293 identify it as belonging to the "homehost", then
2295 will choose a name for the array which is certain not to conflict with
2296 any array which does belong to this host. It does this be adding an
2297 underscore and a small number to the name preferred by the metadata.
2299 Once an appropriate array is found or created and the device is added,
2301 must decide if the array is ready to be started. It will
2302 normally compare the number of available (non-spare) devices to the
2303 number of devices that the metadata suggests need to be active. If
2304 there are at least that many, the array will be started. This means
2305 that if any devices are missing the array will not be restarted.
2311 in which case the array will be run as soon as there are enough
2312 devices present for the data to be accessible. For a RAID1, that
2313 means one device will start the array. For a clean RAID5, the array
2314 will be started as soon as all but one drive is present.
2316 Note that neither of these approaches is really ideal. If it can
2317 be known that all device discovery has completed, then
2321 can be run which will try to start all arrays that are being
2322 incrementally assembled. They are started in "read-auto" mode in
2323 which they are read-only until the first write request. This means
2324 that no metadata updates are made and no attempt at resync or recovery
2325 happens. Further devices that are found before the first write can
2326 still be added safely.
2329 This section describes environment variables that affect how mdadm
2334 Setting this value to 1 will prevent mdadm from automatically launching
2335 mdmon. This variable is intended primarily for debugging mdadm/mdmon.
2341 does not create any device nodes in /dev, but leaves that task to
2345 appears not to be configured, or if this environment variable is set
2348 will create and devices that are needed.
2352 .B " mdadm \-\-query /dev/name-of-device"
2354 This will find out if a given device is a RAID array, or is part of
2355 one, and will provide brief information about the device.
2357 .B " mdadm \-\-assemble \-\-scan"
2359 This will assemble and start all arrays listed in the standard config
2360 file. This command will typically go in a system startup file.
2362 .B " mdadm \-\-stop \-\-scan"
2364 This will shut down all arrays that can be shut down (i.e. are not
2365 currently in use). This will typically go in a system shutdown script.
2367 .B " mdadm \-\-follow \-\-scan \-\-delay=120"
2369 If (and only if) there is an Email address or program given in the
2370 standard config file, then
2371 monitor the status of all arrays listed in that file by
2372 polling them ever 2 minutes.
2374 .B " mdadm \-\-create /dev/md0 \-\-level=1 \-\-raid\-devices=2 /dev/hd[ac]1"
2376 Create /dev/md0 as a RAID1 array consisting of /dev/hda1 and /dev/hdc1.
2379 .B " echo 'DEVICE /dev/hd*[0\-9] /dev/sd*[0\-9]' > mdadm.conf"
2381 .B " mdadm \-\-detail \-\-scan >> mdadm.conf"
2383 This will create a prototype config file that describes currently
2384 active arrays that are known to be made from partitions of IDE or SCSI drives.
2385 This file should be reviewed before being used as it may
2386 contain unwanted detail.
2388 .B " echo 'DEVICE /dev/hd[a\-z] /dev/sd*[a\-z]' > mdadm.conf"
2390 .B " mdadm \-\-examine \-\-scan \-\-config=mdadm.conf >> mdadm.conf"
2392 This will find arrays which could be assembled from existing IDE and
2393 SCSI whole drives (not partitions), and store the information in the
2394 format of a config file.
2395 This file is very likely to contain unwanted detail, particularly
2398 entries. It should be reviewed and edited before being used as an
2401 .B " mdadm \-\-examine \-\-brief \-\-scan \-\-config=partitions"
2403 .B " mdadm \-Ebsc partitions"
2405 Create a list of devices by reading
2406 .BR /proc/partitions ,
2407 scan these for RAID superblocks, and printout a brief listing of all
2410 .B " mdadm \-Ac partitions \-m 0 /dev/md0"
2412 Scan all partitions and devices listed in
2413 .BR /proc/partitions
2416 out of all such devices with a RAID superblock with a minor number of 0.
2418 .B " mdadm \-\-monitor \-\-scan \-\-daemonise > /var/run/mdadm"
2420 If config file contains a mail address or alert program, run mdadm in
2421 the background in monitor mode monitoring all md devices. Also write
2422 pid of mdadm daemon to
2423 .BR /var/run/mdadm .
2425 .B " mdadm \-Iq /dev/somedevice"
2427 Try to incorporate newly discovered device into some array as
2430 .B " mdadm \-\-incremental \-\-rebuild \-\-run \-\-scan"
2432 Rebuild the array map from any current arrays, and then start any that
2435 .B " mdadm /dev/md4 --fail detached --remove detached"
2437 Any devices which are components of /dev/md4 will be marked as faulty
2438 and then remove from the array.
2440 .B " mdadm --grow /dev/md4 --level=6 --backup-file=/root/backup-md4
2444 which is currently a RAID5 array will be converted to RAID6. There
2445 should normally already be a spare drive attached to the array as a
2446 RAID6 needs one more drive than a matching RAID5.
2448 .B " mdadm --create /dev/md/ddf --metadata=ddf --raid-disks 6 /dev/sd[a-f]"
2450 Create a DDF array over 6 devices.
2452 .B " mdadm --create /dev/md/home -n3 -l5 -z 30000000 /dev/md/ddf"
2454 Create a RAID5 array over any 3 devices in the given DDF set. Use
2455 only 30 gigabytes of each device.
2457 .B " mdadm -A /dev/md/ddf1 /dev/sd[a-f]"
2459 Assemble a pre-exist ddf array.
2461 .B " mdadm -I /dev/md/ddf1"
2463 Assemble all arrays contained in the ddf array, assigning names as
2466 .B " mdadm \-\-create \-\-help"
2468 Provide help about the Create mode.
2470 .B " mdadm \-\-config \-\-help"
2472 Provide help about the format of the config file.
2474 .B " mdadm \-\-help"
2476 Provide general help.
2486 lists all active md devices with information about them.
2488 uses this to find arrays when
2490 is given in Misc mode, and to monitor array reconstruction
2495 The config file lists which devices may be scanned to see if
2496 they contain MD super block, and gives identifying information
2497 (e.g. UUID) about known MD arrays. See
2501 .SS /var/run/mdadm/map
2504 mode is used, this file gets a list of arrays currently being created.
2507 does not exist as a directory, then
2508 .B /var/run/mdadm.map
2511 is not available (as may be the case during early boot),
2513 is used on the basis that
2515 is usually available very early in boot.
2520 understand two sorts of names for array devices.
2522 The first is the so-called 'standard' format name, which matches the
2523 names used by the kernel and which appear in
2526 The second sort can be freely chosen, but must reside in
2528 When giving a device name to
2530 to create or assemble an array, either full path name such as
2534 can be given, or just the suffix of the second sort of name, such as
2540 chooses device names during auto-assembly or incremental assembly, it
2541 will sometimes add a small sequence number to the end of the name to
2542 avoid conflicted between multiple arrays that have the same name. If
2544 can reasonably determine that the array really is meant for this host,
2545 either by a hostname in the metadata, or by the presence of the array
2546 in /etc/mdadm.conf, then it will leave off the suffix if possible.
2547 Also if the homehost is specified as
2550 will only use a suffix if a different array of the same name already
2551 exists or is listed in the config file.
2553 The standard names for non-partitioned arrays (the only sort of md
2554 array available in 2.4 and earlier) are of the form
2558 where NN is a number.
2559 The standard names for partitionable arrays (as available from 2.6
2560 onwards) are of the form
2564 Partition numbers should be indicated by added "pMM" to these, thus "/dev/md/d1p2".
2566 From kernel version, 2.6.28 the "non-partitioned array" can actually
2567 be partitioned. So the "md_dNN" names are no longer needed, and
2568 partitions such as "/dev/mdNNpXX" are possible.
2572 was previously known as
2576 is completely separate from the
2578 package, and does not use the
2580 configuration file at all.
2583 For further information on mdadm usage, MD and the various levels of
2586 .B http://linux\-raid.osdl.org/
2588 (based upon Jakob \(/Ostergaard's Software\-RAID.HOWTO)
2590 .\"for new releases of the RAID driver check out:
2593 .\".UR ftp://ftp.kernel.org/pub/linux/kernel/people/mingo/raid-patches
2594 .\"ftp://ftp.kernel.org/pub/linux/kernel/people/mingo/raid-patches
2599 .\".UR http://www.cse.unsw.edu.au/~neilb/patches/linux-stable/
2600 .\"http://www.cse.unsw.edu.au/~neilb/patches/linux-stable/
2603 The latest version of
2605 should always be available from
2607 .B http://www.kernel.org/pub/linux/utils/raid/mdadm/
projects / thirdparty / mdadm.git / blob