1 Changes Prior to this release
2 - Fixed UUID printing in "--detail --brief" for version1 metadata.
3 - --update=resync did exactly the wrong thing for version1 metadata.
4 It caused a resync to not happen, rather than to happen.
5 - Allow --assemble --force to mark a raid6 clean when it has two
6 missing devices (which is needed else if won't assemble.
7 Without this fix it would only assemble if one or zero
9 - Support --update=devicesize for cases where the underlying device
11 - Default to --auto=yes so the array devices with 'standard' names
12 get created automatically, as this is almost always what is wanted.
13 - Give useful message if raid4/5/6 cannot be started because it is
14 not clean and is also degraded.
15 - Increase raid456 stripe cache size if needed to --grow the array.
16 The setting used unfortunately requires intimate knowledge of the
17 kernel, and it not reset when the reshape finishes.
18 - Change 'Device Size' to 'Used Dev Size' because it only shows how
19 much of each device is actually used, not how big they are.
20 - --wait or -W will wait for resync activity to finish on the given
22 - Fix some problems with --update=uuid and add a test.
23 - If two drives in a raid5 disappear at the same time, then "-Af"
24 will add them both in rather than just one and forcing the array
25 to 'clean'. This is slightly safer in some cases.
26 - Check device is large enough before hot-add: this improves quality
28 - Don't hold md device open for so long in --monitor mode - map_dev
29 can be slow and interferes with trying to stop the array.
31 Changes Prior to 2.5.6 release
32 - Fix bug which meant "bitmap=xxx" in mdadm.conf was not handled
34 - Documentation updates.
35 - Fix bug that caused infinite loop when doing auto-assembly,
36 in certain cases where arrays couldn't be assembled.
38 Changes Prior to 2.5.5 release
39 - Don't #include linux/blkpg.h as that isn't safe. Just
40 include the content literally.
41 - Reduce maximum bitmap usage when working with bitmap files,
42 so that a only single-page allocations are made, even on
43 64bit hosts with 4K pages.
44 - Close stray fd in mdassemble so that it can assemble stacked
46 - If mdassemble finds an array already assembled, it marks it
48 - Remove error in md_open if array is already active. This isn't
49 needed and gets in the ways if an array was created e.g. in
50 initramfs, but device doesn't yet exist in /dev.
51 - When --assemble --scan is run, if all arrays that could be found
52 have already been started, don't report an error.
53 - Fix a couple of bugs related to raid10 and the new 'offset' layout.
54 - Improve error message when a wrong '--update' option is given.
56 Changes Prior to 2.5.4 release
57 - When creating devices in /dev/md/ create matching symlinks
58 from /dev. e.g. /dev/md0 -> /dev/md/0.
59 Allow this to be disabled in mdadm.conf or on command line.
60 - Fix some endian-ness issues with version-1 superblocks (affects
62 - Fix endian problem with 'bitmap' metadata
63 - Allow a number (of partitions) after the 'yes' option to --auto=
64 This is particularly useful in the 'create' line in mdadm.conf.
65 - Remove partitions from any whole device that is made part of
66 an md array. This is a work-around for annoying messages
67 when the first block on some drive accidentally looks like a
70 Changes Prior to 2.5.3 release
71 - Document v0.91 superblocks in md.4
72 - Make GPL explicit in man pages.
73 - Fix recent breakage of starting degraded arrays.
74 - Tidyup automatic name choice for v-1 arrays:
75 /dev/md_d0 now becomes '0', not '_d0'.
77 Changes Prior to 2.5.2 release
78 - Fix problem with compiling with gcc-2 compilers
79 - Fix compile problem of post-incrmenting a variable in a macro arg.
80 - Stop map_dev from returning [0:0], as that breaks things.
81 - Add 'Array Slot' line to --examine for version-1 superblocks
82 to make it a bit easier to see what is happening.
83 - Work around bug in --add handling for version-1 superblocks
84 in 2.6.17 (and prior).
85 - Make -assemble a bit more resilient to finding strange
86 information in superblocks.
87 - Don't claim newly added spares are InSync!! (don't know why that
88 code was ever in there)
89 - Work better when no 'ftw' is available, and check to see
90 if current uclibc provides ftw.
91 - Never use /etc/mdadm.conf if --config file is given (previously
92 some code used one, some used the other).
94 Changes Prior to 2.5.1 release
95 - Various fixes for gcc warnings
97 - Makefile improvements for static linking/intalling
98 - Makefile uninstall target
99 - Really fix return status of --examine
101 - Byteorder stuff (again)
102 - Don't try to create devices with --manage or --grow
103 - allow default metadata (superblock) type to be specified
105 - Get --stop to list devices stopped but honour --quiet
106 - remove libssl dependency
107 - Avoid some misdetection of overlapping partitions
108 - Fix memory leak in --monitor mode
110 Changes Prior to 2.5 release
111 - Support 'mailfrom' line in mdadm.conf so the From: line in alert
112 emails can be explicitly set.
113 - Arrange that SparesMissing (which is similar in import to
114 DegradedArray) generates an Email.
115 - Assume "DEVICE partitions" if no DEVICE line is given.
116 - Support new 'offset' layout for raid10.
117 - When creating a bitmap file, choose a chunksize to limit number
118 of bitmap chunks to 2 million. More than this can cause kmalloc
120 - New 'CREATE' line in mdadm.conf for defaults such as owner, group,
122 - --detail checks if array has been started or not and includes that
124 - When using --update=uuid on an array with a bitmap, update the
126 - Add a copy of /proc/mdstat to the mail message sent by mdadm
128 - New flag --no-degraded to avoid starting arrays if there are
129 fewer devices available than last time the array was started.
130 This is only needed with --scan, as with --scan, that behaviour
132 - Support for 'homehost' concept. This is a fairly major update.
133 It includes a configfile option and a command line option for
134 specifying a homehost, records that host in the superblock,
135 and reports the homehost where possible.
136 - Support for Auto Assembly. "mdadm -As" will, if provided with
137 the name of a homehost, try to assemble all arrays it can find
138 that were created for that homehost. See man pages for more details.
140 Changes Prior to 2.4.1 release
141 - Honour --write-mostly when adding to an array without persistent
143 - Fix alignment problem in version-1 superblocks.
144 NOTE: This is an incompatable change affecting raid5 reshape.
145 If you want to reshape a raid5 using version-1 superblocks,
146 use 2.6.17-rc2 or later, and mdadm-2.4.1 or later.
148 Changes Prior to 2.4 release
149 - Rewrite 'reshape' support including performing a backup
150 of the critical region for a raid5 growth, and restoring that
151 backup after a crash.
152 - Put a 'canary' at each end of the backup so a corruption
153 can be more easily detected.
154 - Remove useless 'ident' arguement from ->getinfo_super method.
155 - Support --backup-file for backing-up critical section during
157 - Erase old superblocks (of different versions) when creating new
159 - Allow --monitor to work with arrays with >28 devices
160 - Report reshape information in --detail
161 - Handle symlinks in /dev better
162 - Fix mess in --detail output which a device is missing.
164 - Support 'bitmap=' in mdadm.conf for auto-assembling arrays with
165 write-intent bitmaps in separate files.
166 - Updates to md.4 man page including section on RESTRIPING and SYSFS
168 Changes Prior to 2.3.1 release
169 - Fixed -O2 compile so I could make and RPM.
170 - Type cast number to be printed %llu so it compiles on 64bit
171 machines. (Thanks Luca).
172 - Stop using asm/byteorder.h - to make Redhat happy :-(
173 - Require bitmap files to have a '/' in their name.
174 - Error-check a few syscalls - code from SuSE package.
176 Changes Prior to 2.3 release
177 - Try /etc/mdadm/mdadm.conf if /etc/mdadm.conf doesn't exist.
178 This provided compatability for Debian.
179 - Fixed for version-1 superblock:
180 report chunksize for raid6 and raid10
181 make sure device size used is a multiple of chunksize
182 - Fix "--assemble --scan" crash.
183 - Fix completely failure to create array on ppc64
184 - Fix memcmp in place of memcpy
185 - A few minor improvements to online help
186 - Clean up usage of 'long long' for used-size of devices, so
187 that it is possible to create a raid1 of 7TB devices!
188 - Make internal bitmaps work on 7TB raid1 arrays.
189 - Provide error message if --examine doesn't find any superblock.
190 - Report 'reshape' status in --examine - this depends on kernel
191 patches that are not yet finalised.
192 - Report bitmap status in --detail and --examine
193 - Default to v1 superblocks instead of v0.90 if the array
194 is too big for 0.90 to handle.
195 - Sort the output of "mdadm --detail --scan" so that it is
196 in a suitable order for assembling arrays. i.e. components come
197 before an array that they are part of.
198 - Print size of large reiserfs array properly went warning of
201 Changes Prior to 2.2 release
203 - Support write-intent-bitmaps on raid10
204 - Support little-endian (Rather than hostendian) bitmaps.
205 - Return correct error code from 'mdadm -S'
206 - Remove extra blank line from 'mdadm -Eb' output.
207 - Improve option parsing so that -a and -b do not have
208 optional arguements: the arg is either required or not
209 depending on context.
210 - Allow scanning of devices listed in /proc/partitions even
211 if they don't appear in /dev.
212 - Support --assume-clean in --create mode as well as --build
213 - Add support for --monitor to report to syslog: -y or --syslog.
214 Thanks to Ross Vandegrift
215 - --monitor now reports which device failed in a 'Fail' message
217 - Improve chance of array starting properly after a crash.
218 mdadm was insisting the event numbers were identical, but this
219 isn't needed, and is a problem if the crash was while the metadata
221 - Support --update==uuid
222 - Added README.initramfs and mkinitramfs to help people use an
223 initram for starting md arrays at boot.
225 Changes Prior to 2.1 release
226 - Fix assembling of raid10 array when devices are missing.
227 mdadm now correctly detects if a array is workable or not
228 depending on which devices are present, and so will correctly
229 handle "--assemble --force" if multiple devices have failed.
230 - Report raid10 layout in --examine output.
231 - Fix assembling of arrays that use the version-1 superblock and
232 have spares. Previously the spares would be ignored.
233 - Fix bug so that multiple drives can be re-added at once.
234 - Fix problem with hot-adding a bitmap to version-1-superblock
238 - Support assembling from byte-swapped superblocks
239 metadata type "0.swap" and --update=byteorder
240 - write-mostly and write-behind support for raid1.
241 - Support --name= and 'name=' config entry for identifying
243 - RAID10 in man pages.
244 - Lot of minor manpage updates
246 Changes Prior to 2.0-devel-3 release
247 - Assorted fixes for multiple bugs...
250 Changes Prior to 1.12.0 release
251 Several of these are backported from the Debian package
252 - Don't use 'lstat' to check for blockdevices, use stat.
253 - Document --size=max option for --grow
254 - Document SparesMissing event and DeviceDisappeared/WrongLevel
255 - --stop --scan repeatly cycles until no more progress can be made
256 so that stack devices are stopped properly
257 - Minor type rationalisation for ident->uuid - now always 'int[]'
258 - Fix type in online help for --grow (was -F, now -G)
259 - Allow --auto command line option to set default auto=
260 value when running "--assemble --scan". Previously
261 --auto was ignored if --scan was given
262 - Fix a few type casts
263 - Fix parsing of /dev/md/N in is_standard
264 - Fix rounding errors in human_size()
265 - Fix silly example in mdadm.conf-examples
266 - When finding a /dev name for a device, prefer shorter names
267 - Suppress listing of devices= in --brief output of -D or -E,
268 unless -v is given (-vv gives the old behaviour with -Dsv).
269 This is because the device list can change and so is not a
270 stable aspect of the array
271 - Allow --force with --grow so '-Gfn1' works (on raid1)
272 - Replace sprintf calls with snprintf (to quiet diet-libc)
273 - Add man page for mdassemble
274 - Fix compilation with tinyc
276 Changes Prior to 1.11.0 release
277 - Fix embarassing bug which causes --add to always fail.
279 Changes Prior to 1.10.0 release
280 - Fix bug with --config=partitions
281 - Open sub-devices with O_EXCL to detect if already in use
282 - Make sure superblock updates are flushed directly to disk.
284 Changes Prior to 2.0-devel-1 release
285 - Support for version-1 superblock. See --metadata option.
286 - Support for bitmap based intent logging.
289 Changes Prior to 1.9.0 release
290 - Fix rpm build problem (stray %)
291 - Minor manpage updates
292 - Change "dirty" status to "active" as it was confusing people.
293 - --assemble --auto recognises 'standard' name and insists on using
294 the appropriate major/minor number for them.
295 - Remove underscore from partition names, so partitions of
296 "foo" are "foo1", "foo2" etc (unchanged) and partitions of
297 "f00" are "f00p1", "f00p2" etc rather than "f00_p1"...
298 - Use "major", "minor", "makedev" macros instead of
299 "MAJOR", "MINOR", "MKDEV" so that large device numbers work
300 on 2.6 (providing you have glibc 2.3.3 or later).
301 - Add some missing closes of open file descriptors.
302 - Reread /proc/partition for every array assembled when using
303 it to find devices, rather than only once.
304 - Make "mdadm -Ss" stop stacked devices properly, by reversing the
305 order in which arrays are stopped.
306 - Improve some error messages.
307 - Allow device name to appear before first option, so e.g.
308 mdadm /dev/md0 -A /dev/sd[ab]
310 - Assume '-Q' if just a device is given, rather than being silent.
311 - Change "dirty" status to "active" as it was confusing people.
313 Changes Prior to 1.8.0 release
314 - Makefile cleanup from Luca Berra <bluca@comedia.it>
315 - --pid-file (-i) to set a pid file to use with --monitor --daemonise
316 - Fix typo in mdadm man page
317 - Fix coredump when "-s" used with no config file present.
318 - Support new "faulty" personality which can inject synthetic
319 faults. (Not in kernel.org yet at 1Nov2004)
320 - Support raid0/linear on devices > 2 Terabytes
321 - Make sure raid6 resyncs when created with one missing device
323 Changes Prior to 1.7.0 release
324 - Support "--grow --add" to add a device to a linear array, if the
325 kernel supports it. Not documented yet.
326 - Restore support for uclibc which was broken recently.
327 - Several improvements to the output of --detail, including
328 reporting "resyncing" or "recovering" in the state.
329 - Close filedescriptor at end of --detail (exit would have closed it
330 anyway, so this isn't abig deal).
331 - Report "Sync checkpoint" in --examine output if appropriate.
332 - Add --update=resync for --assemble mode to for a resync when the
334 - Add support for "raid10", which is under development in 2.6.
336 - --monitor now reads spare-group and spares info from config file
337 even when names of arrays to scan are given on the command line
339 Changes Prior to 1.6.0 release
340 - Device name given in -Eb is determined by examining /dev rather
341 than assuming /dev/md%d
342 - Fix bug in --monitor where an array could be held open an so
343 could not be stopped without killing mdadm.
344 - Add --grow mode. Currently only --size and --raid-disks can be
345 changed. Both require kernel support which, at the time of
346 writing, is not in a release kernel yet.
347 - Don't print out "errors" or "no-errors" in -D and -E, as the bit
348 is never set or used.
349 - Use md event notification in 2.6.??? to make --monitor mode
350 respond instantly to events.
351 - Add --auto= option and auto= configfile entry to tell mdadm to
352 create device files as needed. This is particularly useful
353 with partitioned arrays where the major device number can change.
354 - When generating --brief listing, if the standard name doesn't
355 exist, search /dev for one rather than using a temp name.
356 - Allow --build to build raid1 and multipath arrays.
357 - Add "--assume-clean" for Create and Build, particularly for raid1
358 Note: this is dangerous. Only use it if you are certain.
359 - Fix bug so that Rebuild status monitoring works again.
360 - Add "degraded" and "recovering" options to the "Status:"
363 Changes Prior to 1.5.0 release
364 - new commands "mdassemble" which is a stripped-down equivalent of
365 "mdadm -As", that can be compiled with dietlibc.
366 Thanks to Luca Berra <bluca@comedia.it>.
367 It can be using in an initramfs or initrd.
368 - Fix compiling error with BLKGETSIZE64 and some signed/unsigned
370 - Add Rebuild Status (% complete) to --detail output.
371 - Support "--monitor --test" which will generate a test alert
372 for each array once, to test notification paths.
373 - Generate RebuildFinished event when rebuild finishes.
374 - Support for raid6 as found in 2.6.2 - thanks to
375 H. Peter Anvin <hpa@zytor.com>
376 - Support partitioned md arrays with a different major number and
377 naming scheme (md_dX in /proc/mdstat, /dev/md/dXpY in /dev).
379 Changes Prior to 1.4.0 release
380 - Document fact that creating a raid5 array really creates a
381 degraded array with a spare.
382 - Add "spares=" tag to config file and generate it wit --detail and
384 - Add "SparesMissing" event when --monitor first sees an array and
385 it doesn't have the enough spare devices.
386 - Add --update=summaries for --assemble to update summary
387 information in superblock, and correct other inconsistancies in
389 - Add --test option to --detail to set a meaningful exit status.
391 Changes Prior to 1.3.0 release
392 - Make 'size' and unsigned long in Create to allow creation of
394 - Explicitly flag spare devices as 'spare' in --detail and --examine
395 output. Previously they simply had no flags lists.
396 - Make MailCmd (for monitor) configurable in Makefile, and default
397 to "/usr/sbin/sendmail -t". Also split out the warning related
398 flags into CWFLAGS for easier build configurability.
399 - Minor bugfix in Manage code.
400 - --monitor now notices and reports degraded arrays at startup using
401 "DegradedArray" event, and also has a --oneshot option to only
402 report DegradedArrays, and then exit.
403 - Small man-page clarification w.r.t. raid levels and raid4 in
405 - Disallow creation of arrays with only one device as this is
406 probably a mistake. --force will override this check.
407 - Correct some misleading documentation in the "mdadm --create --help"
409 - Ignore chunksize if raid1 or multipath.
410 - Explicit statement in man page that raid-disks cannot be changed
411 after array is created.
412 - Improve message when attempting to start an array with
413 insufficient devices. Instead of required the array to be full,
414 we only require it has as many active devices as last time.
416 Changes Prior to 1.2.0 release
417 - Fix bug where --daemonise required an argument.
418 - In --assemble --verbose, print appropriate message if device is
420 - Updated mdadm.conf.5 to reflect fact that device= takes wildcards
421 - Typos: componenet -> component
422 - Reduce size of "--help" message put excess into "--help-options"
423 - Fix bug introduced when MD_SB_DISKS dependancy removed, and which
424 caused spares not be assembled properly.
425 - Print appropriate message if --monitor --scan decides not to
427 Changes Prior to 1.1.0 release
428 - add --deamonise flag for --monitor - forks and prints pid to stdout
429 - Fix bug so we REALLY clear dirty flag with -Af
430 - -Db now prints a 'devices=' word for each array.
431 - "mdadm -A /dev/md0" will get info from configfile, even without scan
432 - When assembling multipath arrays, ignore devices which are flagged
434 - take --super-minor=dev to mean "use the minor number of the mddev
436 - take --config=none to mean "completely ignore config file"
437 - Make --monitor require --scan or a device list.
438 Changes Prior to 1.0.9 release
439 - Documentation updates including kernel parameters documented
441 - --assemble --force for raid4/5 will mark clean, needed for 2.5
442 - --detail prints out the events counter as well
443 - flush device before reading superblock to be sure to get
445 - added mdadm.static target to makefile for static linking
446 - --monitor was ignoring /dev/md0 due to off-by-one error
448 - Fix printing of Gibibytes - calc was wrong.
449 - Fix printing of Array Size in --detail when very big.
450 - --monitor no longer tries to work for raid0 or linear as these
451 have nothing to be monitored.
452 - The word 'partitions' on a DEVICE line will cause all partitions
453 listed in /proc/partitions to be considered
454 - If the config file is called 'partitions' then it will be treated
455 as though it contained exactly 'device partitions' so e.g.
456 mdadm -Ebsc partitions
457 will find all raid partitions easily.
458 - successfully assemble multipath devices by ignoring raid_disk
459 value from superblock (it is always the same).
460 - --assemble not tied to MD_SB_DISKS limit quite so much
461 - Support compiling with tcc
462 - Support compiling with uclibc - just skip scan of /dev
463 - Add --update= option for Assemble mode. Either sparc2.2
464 or super-minor updates are possible. See mdadm.8
466 Changes Prior to 1.0.1 release
467 - Round off MB/GiB etc values instead of round down.
468 - Add --sparc2.2 option to examine to shift superblock around
469 and --sparc2.2update to rewrite the superblock
470 - Fix assorted typos in online help
472 Changes Prior to 1.0.0 release
473 - Allow --config with Misc mode (for --examine --scan)
474 - Add $(CXFLAGS) to end of CFLAGS in makefile
475 - When making an N disk raid5 array, the Nth drive
476 is moved to the end of the array as a spare rather than
477 being shifted up one place. This means that when the
478 kernel builds onto the last spare and inserts it,
479 the devices will be in the expected order.
480 - Man page improvements
481 Changes Prior to 0.8.2 release
482 - Correct spelling of persist[ae]nce/persist[ae]nt.
483 - Change "disk" to "device" in options and config file
484 - convert array size to "long long" *before* shift-left in -D and -Q
486 Changes Prior to 0.8.1 release
487 - Add "INSTALL" file.
488 - Fix some "i" variables that were not being set properly
489 - Initialise minsize and maxsize so that compilers don't complain.
490 - Tidy up Makefile and mdadm.spec installations
491 - Add "multipath" to documentation of valid levels
493 Changes Prior to 0.8 release
494 - Fix another bug in Assemble.c due to confusing 'i' with 'j'
495 - Minimal, untested, support for multipath
496 - re-write of argument parsing to have more coherent modes,
497 - add --query,-Q option
498 - Update mdadm.8 to reflect arg processing change and --query
499 - Change "long" to "unsigned long" for device sizes
500 - Handle "mailaddr" and "program" lines in config file for follow/scan mode.
501 - --follow --scan will exit if no program or mail found
502 - Add MAILADDR and PROGRAM to mdadm.conf-example
503 - Spell check man pages
504 - consistently use "component devices" instead of "subdevices"
505 - Make -Wall -Werror really work and fix lots of errors.
506 - --detail and --stop can have --scan which chooses devices from /proc/mdstat
507 - --monitor detects 20% changes in resync, failed spares,
509 - --monitor --scan will automatically add any devices found in /proc/mdstat
510 - --monitor will move spares between arrays with same spare-group if necessary
511 - Documentation for Monitor Mode
512 - --query notes if the array containing the given device is active or not
513 - Finished md.4 man page.
515 Changes Prior to 0.7.2 release
516 - mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra -- bluca@comedia.it
517 - more mdadm.spec updates from Gregory Leblanc <gleblanc@linuxweasel.com>
518 - make directory for mdadm.conf configurable in Makefile
519 - Finished mdadm.conf.5. Removed details of conf file from
520 mdadm.8 leaving a reference to mdadm.conf.5.
521 - Fix bug in Assemble.c, thanks to Junaid Rizvi <domdev@sat.net.pk>
522 - Get --assemble --force to make sure old major/minor numbers are
523 consistant, as md.c worries about this :-(
526 Changes Prior to 0.7.1 release
528 - use BLKGETSIZE64 if available for array size
529 - give human readable as GiB/MiB and GB and MB, with 2 decimal point precision
530 - Only warn about size variation for raid1/4/5.
531 - Started md.4 man page
532 - Started mdadm.conf.5 man page
534 Changes Prior to 0.7 release
536 - Fix makefile to install binary at /sbin and not /sbin/sbin
537 Also install man page.
538 - Add --zero-superblock based on --destroywithextremeprejudice
539 from Dale Stephenson <steph@snapserver.com>
540 - change name to mdadm. It is palandromic, and much nicer to pronouce.
542 Changes Prior to 0.6 release
544 - Remove the limit on the number of device names that can be
545 given on the command line.
546 - Fix bug in --assemble --force where it would only update a
548 - Fix bogus printing of big numbers not being block devices
549 when given names of devices that don't exist.
550 - When --assemble --force, consider superblocks with an event
551 count that is 1 behind as out-of-date. Normally they are
552 considered up-to-date (as the kernel assumes this too).
553 - When marking drives as not-failed in the superblock,
554 we also mark them as ACTIVE and SYNC.
555 - Don't start arrays for which not all drives are available unless:
556 --scan which implies that all drives were found automatically
557 --run which means the user knows what they want
558 --force which means that we are fixing something broken
559 - Make sure all device numbers passed as 3rd arg of ioctl
560 are passed as unsigned lock, so that it works on SPARC
561 - If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK
562 if we cannot read from the array, i.e. if the array is
565 - Taught Examine to handle --scan. It examines all devices listed
566 on DEVICE lines in the config file.
567 - Added --brief (-b) flag for Examine and Detail to print out
568 and mdctl.conf compatible description with uuid=, level=,
569 disks= and - for Examine - devices=
570 --examine --brief collects all devices the make the one array and
571 list them as one entry.
572 - Added level= and disks= options to ARRAY lines in config files
573 so --brief output could be used as-is.
574 - Make parity style ({left,right}-{,a}symmetric) consistantly use -,
576 - Add "Array Size" to --detail output
577 - Change "Size" to "Device Size" and exclude from Detail of arrays
578 that do not have a consistent device size.
579 - Add Human readable MiB or GiB value on size lines of Detail and Examine
580 - --assemble --scan doesn't complain about active drives
581 - require number of spares given in -x to be listed.
582 - Made --build actually work.
583 Changes Prior to 0.5 release
586 spare drives are handled properly.
588 --force can be used to recover from 2-drive failures on RAID5
589 If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should
590 make a raid5 array, but it has experienced multiple failures and
593 mdctl --assemble --force /dev/md0 /dev/hd[abcd]1
595 Should update the superblock on the newest failed drive and
596 restart the array in degraded mode. You should then remove the
597 remaining failed drive and re-add it (if you are happy that it
600 Ofcourse whenever you have a 2-drive failure, you have a risk
601 of corruption in data that hasn't be changed for a long time. So
602 this doesn't give you your array back all nice and happy, but it
603 does allow you to recover data that might not be corrupt.
605 More flexibility in identifying a RAID array in the mdctl.conf
607 array /dev/md4 super-minor=4
609 assembles /dev/md4 from all devices found that have a raid
610 superblock that says the minor number of the array is 4.
611 If the blocks with the right minor number do not all have the
612 same UUID, an error is flags and no assembly happens.
614 array /dev/md3 devices=/dev/hd[abc]2
616 Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All
617 devices must exist and have raid superblock with the same uuid.
619 If two identity specifiers are used, only devices that match all
620 of them are considered, so
622 array /dev/md2 devices=/dev/hd?2 super-minor=2
624 will assemble /dev/md2 using all /dev/hd?2 devices which have a
625 raid superblock with minor number 2.
628 When listing devices for --create, the word "missing" can be
629 used to indicate that the respective slot does not have a
630 working drive currently. This is similar to the "failed-disk"
631 directive in mkraid/raidtab.
633 mdctl --create --level=5 -raid-disks=4 --spare-disks=2
634 /dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde
636 will create a raid5 array with the third slot empty, and two
639 By default, raid5 arrays are created with the last slot empty
640 and drive listed for the last slot added as a spare. If a
641 "missing" slot is given, or if --force is given, then --create
642 does exactly what you ask and doesn't try to be clever.
645 --follow / --monitor:
647 This is a new mode. I couldn't stop my self from picking a name
648 starting with F (as current modes start A,B,C,D,E) but I
649 relented and provided an alternate name that is somewhat more
652 In this mode, mdctl does not exit, but runs continuously and
653 periodically polls all the md devices to see if they have had
654 any interested state change.
655 The changes that it currently notices are:
656 Fail - an active disc fails
657 FailSpare - a spare, that was presumably being build, fails
658 ActiveSpare - a spare becomes active, presumably after a rebuild.
661 --mail mailaddress - send Email on any Fail* event
662 --program program - run the program on any event.
663 Args are: eventname mddevice subdevice(if-known)
664 --delay seconds - change from the default 60second pause
667 I plan to add functionality to this mode to allow sharing of
668 spare drives. If an array is marks "spare-group=fred", and it
669 has a failed drive and no spares, and if some other array is
670 also "spare-group=fred" and it has no failed drives, but does
671 have a spare drive that is big enough, the spare will be moved
674 I also have the idea of adding a --grow mode which will re-organise
675 the data on an N disk raid0/4/5 array to be on an N+M disk array.
676 I have no concrete plans for this though.
678 I got rid of the "v" in the archive file name, and include the
679 version number in the directory created by the archive.
681 There is now a man page and mdctl.spec (for rpm) thanks to
682 Danilo Godec <danci@agenda.si>.
684 Ofcourse, the man page is now out of date and despite being based on
685 the --help output, is not wholy correct. After I get --follow
686 working properly, I plan to revise the various documentation and/or
687 the code to make sure the two match.