1 Changes Prior to this release
2 - Support 'mailfrom' line in mdadm.conf so the From: line in alert
3 emails can be explicitly set.
4 - Arrange that SparesMissing (which is similar in import to
5 DegradedArray) generates an Email.
6 - Assume "DEVICE partitions" if no DEVICE line is given.
7 - Support new 'offset' layout for raid10.
8 - When creating a bitmap file, choose a chunksize to limit number
9 of bitmap chunks to 2 million. More than this can cause kmalloc
11 - New 'CREATE' line in mdadm.conf for defaults such as owner, group,
13 - --detail checks if array has been started or not and includes that
15 - When using --update=uuid on an array with a bitmap, update the
17 - Add a copy of /proc/mdstat to the mail message send by mdadm
20 Changes Prior to 2.4.1 release
21 - Honour --write-mostly when adding to an array without persistent
23 - Fix alignment problem in version-1 superblocks.
24 NOTE: This is an incompatable change affecting raid5 reshape.
25 If you want to reshape a raid5 using version-1 superblocks,
26 use 2.6.17-rc2 or later, and mdadm-2.4.1 or later.
28 Changes Prior to 2.4 release
29 - Rewrite 'reshape' support including performing a backup
30 of the critical region for a raid5 growth, and restoring that
32 - Put a 'canary' at each end of the backup so a corruption
33 can be more easily detected.
34 - Remove useless 'ident' arguement from ->getinfo_super method.
35 - Support --backup-file for backing-up critical section during
37 - Erase old superblocks (of different versions) when creating new
39 - Allow --monitor to work with arrays with >28 devices
40 - Report reshape information in --detail
41 - Handle symlinks in /dev better
42 - Fix mess in --detail output which a device is missing.
44 - Support 'bitmap=' in mdadm.conf for auto-assembling arrays with
45 write-intent bitmaps in separate files.
46 - Updates to md.4 man page including section on RESTRIPING and SYSFS
48 Changes Prior to 2.3.1 release
49 - Fixed -O2 compile so I could make and RPM.
50 - Type cast number to be printed %llu so it compiles on 64bit
51 machines. (Thanks Luca).
52 - Stop using asm/byteorder.h - to make Redhat happy :-(
53 - Require bitmap files to have a '/' in their name.
54 - Error-check a few syscalls - code from SuSE package.
56 Changes Prior to 2.3 release
57 - Try /etc/mdadm/mdadm.conf if /etc/mdadm.conf doesn't exist.
58 This provided compatability for Debian.
59 - Fixed for version-1 superblock:
60 report chunksize for raid6 and raid10
61 make sure device size used is a multiple of chunksize
62 - Fix "--assemble --scan" crash.
63 - Fix completely failure to create array on ppc64
64 - Fix memcmp in place of memcpy
65 - A few minor improvements to online help
66 - Clean up usage of 'long long' for used-size of devices, so
67 that it is possible to create a raid1 of 7TB devices!
68 - Make internal bitmaps work on 7TB raid1 arrays.
69 - Provide error message if --examine doesn't find any superblock.
70 - Report 'reshape' status in --examine - this depends on kernel
71 patches that are not yet finalised.
72 - Report bitmap status in --detail and --examine
73 - Default to v1 superblocks instead of v0.90 if the array
74 is too big for 0.90 to handle.
75 - Sort the output of "mdadm --detail --scan" so that it is
76 in a suitable order for assembling arrays. i.e. components come
77 before an array that they are part of.
78 - Print size of large reiserfs array properly went warning of
81 Changes Prior to 2.2 release
83 - Support write-intent-bitmaps on raid10
84 - Support little-endian (Rather than hostendian) bitmaps.
85 - Return correct error code from 'mdadm -S'
86 - Remove extra blank line from 'mdadm -Eb' output.
87 - Improve option parsing so that -a and -b do not have
88 optional arguements: the arg is either required or not
90 - Allow scanning of devices listed in /proc/partitions even
91 if they don't appear in /dev.
92 - Support --assume-clean in --create mode as well as --build
93 - Add support for --monitor to report to syslog: -y or --syslog.
94 Thanks to Ross Vandegrift
95 - --monitor now reports which device failed in a 'Fail' message
97 - Improve chance of array starting properly after a crash.
98 mdadm was insisting the event numbers were identical, but this
99 isn't needed, and is a problem if the crash was while the metadata
101 - Support --update==uuid
102 - Added README.initramfs and mkinitramfs to help people use an
103 initram for starting md arrays at boot.
105 Changes Prior to 2.1 release
106 - Fix assembling of raid10 array when devices are missing.
107 mdadm now correctly detects if a array is workable or not
108 depending on which devices are present, and so will correctly
109 handle "--assemble --force" if multiple devices have failed.
110 - Report raid10 layout in --examine output.
111 - Fix assembling of arrays that use the version-1 superblock and
112 have spares. Previously the spares would be ignored.
113 - Fix bug so that multiple drives can be re-added at once.
114 - Fix problem with hot-adding a bitmap to version-1-superblock
118 - Support assembling from byte-swapped superblocks
119 metadata type "0.swap" and --update=byteorder
120 - write-mostly and write-behind support for raid1.
121 - Support --name= and 'name=' config entry for identifying
123 - RAID10 in man pages.
124 - Lot of minor manpage updates
126 Changes Prior to 2.0-devel-3 release
127 - Assorted fixes for multiple bugs...
130 Changes Prior to 1.12.0 release
131 Several of these are backported from the Debian package
132 - Don't use 'lstat' to check for blockdevices, use stat.
133 - Document --size=max option for --grow
134 - Document SparesMissing event and DeviceDisappeared/WrongLevel
135 - --stop --scan repeatly cycles until no more progress can be made
136 so that stack devices are stopped properly
137 - Minor type rationalisation for ident->uuid - now always 'int[]'
138 - Fix type in online help for --grow (was -F, now -G)
139 - Allow --auto command line option to set default auto=
140 value when running "--assemble --scan". Previously
141 --auto was ignored if --scan was given
142 - Fix a few type casts
143 - Fix parsing of /dev/md/N in is_standard
144 - Fix rounding errors in human_size()
145 - Fix silly example in mdadm.conf-examples
146 - When finding a /dev name for a device, prefer shorter names
147 - Suppress listing of devices= in --brief output of -D or -E,
148 unless -v is given (-vv gives the old behaviour with -Dsv).
149 This is because the device list can change and so is not a
150 stable aspect of the array
151 - Allow --force with --grow so '-Gfn1' works (on raid1)
152 - Replace sprintf calls with snprintf (to quiet diet-libc)
153 - Add man page for mdassemble
154 - Fix compilation with tinyc
156 Changes Prior to 1.11.0 release
157 - Fix embarassing bug which causes --add to always fail.
159 Changes Prior to 1.10.0 release
160 - Fix bug with --config=partitions
161 - Open sub-devices with O_EXCL to detect if already in use
162 - Make sure superblock updates are flushed directly to disk.
164 Changes Prior to 2.0-devel-1 release
165 - Support for version-1 superblock. See --metadata option.
166 - Support for bitmap based intent logging.
169 Changes Prior to 1.9.0 release
170 - Fix rpm build problem (stray %)
171 - Minor manpage updates
172 - Change "dirty" status to "active" as it was confusing people.
173 - --assemble --auto recognises 'standard' name and insists on using
174 the appropriate major/minor number for them.
175 - Remove underscore from partition names, so partitions of
176 "foo" are "foo1", "foo2" etc (unchanged) and partitions of
177 "f00" are "f00p1", "f00p2" etc rather than "f00_p1"...
178 - Use "major", "minor", "makedev" macros instead of
179 "MAJOR", "MINOR", "MKDEV" so that large device numbers work
180 on 2.6 (providing you have glibc 2.3.3 or later).
181 - Add some missing closes of open file descriptors.
182 - Reread /proc/partition for every array assembled when using
183 it to find devices, rather than only once.
184 - Make "mdadm -Ss" stop stacked devices properly, by reversing the
185 order in which arrays are stopped.
186 - Improve some error messages.
187 - Allow device name to appear before first option, so e.g.
188 mdadm /dev/md0 -A /dev/sd[ab]
190 - Assume '-Q' if just a device is given, rather than being silent.
191 - Change "dirty" status to "active" as it was confusing people.
193 Changes Prior to 1.8.0 release
194 - Makefile cleanup from Luca Berra <bluca@comedia.it>
195 - --pid-file (-i) to set a pid file to use with --monitor --daemonise
196 - Fix typo in mdadm man page
197 - Fix coredump when "-s" used with no config file present.
198 - Support new "faulty" personality which can inject synthetic
199 faults. (Not in kernel.org yet at 1Nov2004)
200 - Support raid0/linear on devices > 2 Terabytes
201 - Make sure raid6 resyncs when created with one missing device
203 Changes Prior to 1.7.0 release
204 - Support "--grow --add" to add a device to a linear array, if the
205 kernel supports it. Not documented yet.
206 - Restore support for uclibc which was broken recently.
207 - Several improvements to the output of --detail, including
208 reporting "resyncing" or "recovering" in the state.
209 - Close filedescriptor at end of --detail (exit would have closed it
210 anyway, so this isn't abig deal).
211 - Report "Sync checkpoint" in --examine output if appropriate.
212 - Add --update=resync for --assemble mode to for a resync when the
214 - Add support for "raid10", which is under development in 2.6.
216 - --monitor now reads spare-group and spares info from config file
217 even when names of arrays to scan are given on the command line
219 Changes Prior to 1.6.0 release
220 - Device name given in -Eb is determined by examining /dev rather
221 than assuming /dev/md%d
222 - Fix bug in --monitor where an array could be held open an so
223 could not be stopped without killing mdadm.
224 - Add --grow mode. Currently only --size and --raid-disks can be
225 changed. Both require kernel support which, at the time of
226 writing, is not in a release kernel yet.
227 - Don't print out "errors" or "no-errors" in -D and -E, as the bit
228 is never set or used.
229 - Use md event notification in 2.6.??? to make --monitor mode
230 respond instantly to events.
231 - Add --auto= option and auto= configfile entry to tell mdadm to
232 create device files as needed. This is particularly useful
233 with partitioned arrays where the major device number can change.
234 - When generating --brief listing, if the standard name doesn't
235 exist, search /dev for one rather than using a temp name.
236 - Allow --build to build raid1 and multipath arrays.
237 - Add "--assume-clean" for Create and Build, particularly for raid1
238 Note: this is dangerous. Only use it if you are certain.
239 - Fix bug so that Rebuild status monitoring works again.
240 - Add "degraded" and "recovering" options to the "Status:"
243 Changes Prior to 1.5.0 release
244 - new commands "mdassemble" which is a stripped-down equivalent of
245 "mdadm -As", that can be compiled with dietlibc.
246 Thanks to Luca Berra <bluca@comedia.it>.
247 It can be using in an initramfs or initrd.
248 - Fix compiling error with BLKGETSIZE64 and some signed/unsigned
250 - Add Rebuild Status (% complete) to --detail output.
251 - Support "--monitor --test" which will generate a test alert
252 for each array once, to test notification paths.
253 - Generate RebuildFinished event when rebuild finishes.
254 - Support for raid6 as found in 2.6.2 - thanks to
255 H. Peter Anvin <hpa@zytor.com>
256 - Support partitioned md arrays with a different major number and
257 naming scheme (md_dX in /proc/mdstat, /dev/md/dXpY in /dev).
259 Changes Prior to 1.4.0 release
260 - Document fact that creating a raid5 array really creates a
261 degraded array with a spare.
262 - Add "spares=" tag to config file and generate it wit --detail and
264 - Add "SparesMissing" event when --monitor first sees an array and
265 it doesn't have the enough spare devices.
266 - Add --update=summaries for --assemble to update summary
267 information in superblock, and correct other inconsistancies in
269 - Add --test option to --detail to set a meaningful exit status.
271 Changes Prior to 1.3.0 release
272 - Make 'size' and unsigned long in Create to allow creation of
274 - Explicitly flag spare devices as 'spare' in --detail and --examine
275 output. Previously they simply had no flags lists.
276 - Make MailCmd (for monitor) configurable in Makefile, and default
277 to "/usr/sbin/sendmail -t". Also split out the warning related
278 flags into CWFLAGS for easier build configurability.
279 - Minor bugfix in Manage code.
280 - --monitor now notices and reports degraded arrays at startup using
281 "DegradedArray" event, and also has a --oneshot option to only
282 report DegradedArrays, and then exit.
283 - Small man-page clarification w.r.t. raid levels and raid4 in
285 - Disallow creation of arrays with only one device as this is
286 probably a mistake. --force will override this check.
287 - Correct some misleading documentation in the "mdadm --create --help"
289 - Ignore chunksize if raid1 or multipath.
290 - Explicit statement in man page that raid-disks cannot be changed
291 after array is created.
292 - Improve message when attempting to start an array with
293 insufficient devices. Instead of required the array to be full,
294 we only require it has as many active devices as last time.
296 Changes Prior to 1.2.0 release
297 - Fix bug where --daemonise required an argument.
298 - In --assemble --verbose, print appropriate message if device is
300 - Updated mdadm.conf.5 to reflect fact that device= takes wildcards
301 - Typos: componenet -> component
302 - Reduce size of "--help" message put excess into "--help-options"
303 - Fix bug introduced when MD_SB_DISKS dependancy removed, and which
304 caused spares not be assembled properly.
305 - Print appropriate message if --monitor --scan decides not to
307 Changes Prior to 1.1.0 release
308 - add --deamonise flag for --monitor - forks and prints pid to stdout
309 - Fix bug so we REALLY clear dirty flag with -Af
310 - -Db now prints a 'devices=' word for each array.
311 - "mdadm -A /dev/md0" will get info from configfile, even without scan
312 - When assembling multipath arrays, ignore devices which are flagged
314 - take --super-minor=dev to mean "use the minor number of the mddev
316 - take --config=none to mean "completely ignore config file"
317 - Make --monitor require --scan or a device list.
318 Changes Prior to 1.0.9 release
319 - Documentation updates including kernel parameters documented
321 - --assemble --force for raid4/5 will mark clean, needed for 2.5
322 - --detail prints out the events counter as well
323 - flush device before reading superblock to be sure to get
325 - added mdadm.static target to makefile for static linking
326 - --monitor was ignoring /dev/md0 due to off-by-one error
328 - Fix printing of Gibibytes - calc was wrong.
329 - Fix printing of Array Size in --detail when very big.
330 - --monitor no longer tries to work for raid0 or linear as these
331 have nothing to be monitored.
332 - The word 'partitions' on a DEVICE line will cause all partitions
333 listed in /proc/partitions to be considered
334 - If the config file is called 'partitions' then it will be treated
335 as though it contained exactly 'device partitions' so e.g.
336 mdadm -Ebsc partitions
337 will find all raid partitions easily.
338 - successfully assemble multipath devices by ignoring raid_disk
339 value from superblock (it is always the same).
340 - --assemble not tied to MD_SB_DISKS limit quite so much
341 - Support compiling with tcc
342 - Support compiling with uclibc - just skip scan of /dev
343 - Add --update= option for Assemble mode. Either sparc2.2
344 or super-minor updates are possible. See mdadm.8
346 Changes Prior to 1.0.1 release
347 - Round off MB/GiB etc values instead of round down.
348 - Add --sparc2.2 option to examine to shift superblock around
349 and --sparc2.2update to rewrite the superblock
350 - Fix assorted typos in online help
352 Changes Prior to 1.0.0 release
353 - Allow --config with Misc mode (for --examine --scan)
354 - Add $(CXFLAGS) to end of CFLAGS in makefile
355 - When making an N disk raid5 array, the Nth drive
356 is moved to the end of the array as a spare rather than
357 being shifted up one place. This means that when the
358 kernel builds onto the last spare and inserts it,
359 the devices will be in the expected order.
360 - Man page improvements
361 Changes Prior to 0.8.2 release
362 - Correct spelling of persist[ae]nce/persist[ae]nt.
363 - Change "disk" to "device" in options and config file
364 - convert array size to "long long" *before* shift-left in -D and -Q
366 Changes Prior to 0.8.1 release
367 - Add "INSTALL" file.
368 - Fix some "i" variables that were not being set properly
369 - Initialise minsize and maxsize so that compilers don't complain.
370 - Tidy up Makefile and mdadm.spec installations
371 - Add "multipath" to documentation of valid levels
373 Changes Prior to 0.8 release
374 - Fix another bug in Assemble.c due to confusing 'i' with 'j'
375 - Minimal, untested, support for multipath
376 - re-write of argument parsing to have more coherent modes,
377 - add --query,-Q option
378 - Update mdadm.8 to reflect arg processing change and --query
379 - Change "long" to "unsigned long" for device sizes
380 - Handle "mailaddr" and "program" lines in config file for follow/scan mode.
381 - --follow --scan will exit if no program or mail found
382 - Add MAILADDR and PROGRAM to mdadm.conf-example
383 - Spell check man pages
384 - consistently use "component devices" instead of "subdevices"
385 - Make -Wall -Werror really work and fix lots of errors.
386 - --detail and --stop can have --scan which chooses devices from /proc/mdstat
387 - --monitor detects 20% changes in resync, failed spares,
389 - --monitor --scan will automatically add any devices found in /proc/mdstat
390 - --monitor will move spares between arrays with same spare-group if necessary
391 - Documentation for Monitor Mode
392 - --query notes if the array containing the given device is active or not
393 - Finished md.4 man page.
395 Changes Prior to 0.7.2 release
396 - mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra -- bluca@comedia.it
397 - more mdadm.spec updates from Gregory Leblanc <gleblanc@linuxweasel.com>
398 - make directory for mdadm.conf configurable in Makefile
399 - Finished mdadm.conf.5. Removed details of conf file from
400 mdadm.8 leaving a reference to mdadm.conf.5.
401 - Fix bug in Assemble.c, thanks to Junaid Rizvi <domdev@sat.net.pk>
402 - Get --assemble --force to make sure old major/minor numbers are
403 consistant, as md.c worries about this :-(
406 Changes Prior to 0.7.1 release
408 - use BLKGETSIZE64 if available for array size
409 - give human readable as GiB/MiB and GB and MB, with 2 decimal point precision
410 - Only warn about size variation for raid1/4/5.
411 - Started md.4 man page
412 - Started mdadm.conf.5 man page
414 Changes Prior to 0.7 release
416 - Fix makefile to install binary at /sbin and not /sbin/sbin
417 Also install man page.
418 - Add --zero-superblock based on --destroywithextremeprejudice
419 from Dale Stephenson <steph@snapserver.com>
420 - change name to mdadm. It is palandromic, and much nicer to pronouce.
422 Changes Prior to 0.6 release
424 - Remove the limit on the number of device names that can be
425 given on the command line.
426 - Fix bug in --assemble --force where it would only update a
428 - Fix bogus printing of big numbers not being block devices
429 when given names of devices that don't exist.
430 - When --assemble --force, consider superblocks with an event
431 count that is 1 behind as out-of-date. Normally they are
432 considered up-to-date (as the kernel assumes this too).
433 - When marking drives as not-failed in the superblock,
434 we also mark them as ACTIVE and SYNC.
435 - Don't start arrays for which not all drives are available unless:
436 --scan which implies that all drives were found automatically
437 --run which means the user knows what they want
438 --force which means that we are fixing something broken
439 - Make sure all device numbers passed as 3rd arg of ioctl
440 are passed as unsigned lock, so that it works on SPARC
441 - If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK
442 if we cannot read from the array, i.e. if the array is
445 - Taught Examine to handle --scan. It examines all devices listed
446 on DEVICE lines in the config file.
447 - Added --brief (-b) flag for Examine and Detail to print out
448 and mdctl.conf compatible description with uuid=, level=,
449 disks= and - for Examine - devices=
450 --examine --brief collects all devices the make the one array and
451 list them as one entry.
452 - Added level= and disks= options to ARRAY lines in config files
453 so --brief output could be used as-is.
454 - Make parity style ({left,right}-{,a}symmetric) consistantly use -,
456 - Add "Array Size" to --detail output
457 - Change "Size" to "Device Size" and exclude from Detail of arrays
458 that do not have a consistent device size.
459 - Add Human readable MiB or GiB value on size lines of Detail and Examine
460 - --assemble --scan doesn't complain about active drives
461 - require number of spares given in -x to be listed.
462 - Made --build actually work.
463 Changes Prior to 0.5 release
466 spare drives are handled properly.
468 --force can be used to recover from 2-drive failures on RAID5
469 If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should
470 make a raid5 array, but it has experienced multiple failures and
473 mdctl --assemble --force /dev/md0 /dev/hd[abcd]1
475 Should update the superblock on the newest failed drive and
476 restart the array in degraded mode. You should then remove the
477 remaining failed drive and re-add it (if you are happy that it
480 Ofcourse whenever you have a 2-drive failure, you have a risk
481 of corruption in data that hasn't be changed for a long time. So
482 this doesn't give you your array back all nice and happy, but it
483 does allow you to recover data that might not be corrupt.
485 More flexibility in identifying a RAID array in the mdctl.conf
487 array /dev/md4 super-minor=4
489 assembles /dev/md4 from all devices found that have a raid
490 superblock that says the minor number of the array is 4.
491 If the blocks with the right minor number do not all have the
492 same UUID, an error is flags and no assembly happens.
494 array /dev/md3 devices=/dev/hd[abc]2
496 Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All
497 devices must exist and have raid superblock with the same uuid.
499 If two identity specifiers are used, only devices that match all
500 of them are considered, so
502 array /dev/md2 devices=/dev/hd?2 super-minor=2
504 will assemble /dev/md2 using all /dev/hd?2 devices which have a
505 raid superblock with minor number 2.
508 When listing devices for --create, the word "missing" can be
509 used to indicate that the respective slot does not have a
510 working drive currently. This is similar to the "failed-disk"
511 directive in mkraid/raidtab.
513 mdctl --create --level=5 -raid-disks=4 --spare-disks=2
514 /dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde
516 will create a raid5 array with the third slot empty, and two
519 By default, raid5 arrays are created with the last slot empty
520 and drive listed for the last slot added as a spare. If a
521 "missing" slot is given, or if --force is given, then --create
522 does exactly what you ask and doesn't try to be clever.
525 --follow / --monitor:
527 This is a new mode. I couldn't stop my self from picking a name
528 starting with F (as current modes start A,B,C,D,E) but I
529 relented and provided an alternate name that is somewhat more
532 In this mode, mdctl does not exit, but runs continuously and
533 periodically polls all the md devices to see if they have had
534 any interested state change.
535 The changes that it currently notices are:
536 Fail - an active disc fails
537 FailSpare - a spare, that was presumably being build, fails
538 ActiveSpare - a spare becomes active, presumably after a rebuild.
541 --mail mailaddress - send Email on any Fail* event
542 --program program - run the program on any event.
543 Args are: eventname mddevice subdevice(if-known)
544 --delay seconds - change from the default 60second pause
547 I plan to add functionality to this mode to allow sharing of
548 spare drives. If an array is marks "spare-group=fred", and it
549 has a failed drive and no spares, and if some other array is
550 also "spare-group=fred" and it has no failed drives, but does
551 have a spare drive that is big enough, the spare will be moved
554 I also have the idea of adding a --grow mode which will re-organise
555 the data on an N disk raid0/4/5 array to be on an N+M disk array.
556 I have no concrete plans for this though.
558 I got rid of the "v" in the archive file name, and include the
559 version number in the directory created by the archive.
561 There is now a man page and mdctl.spec (for rpm) thanks to
562 Danilo Godec <danci@agenda.si>.
564 Ofcourse, the man page is now out of date and despite being based on
565 the --help output, is not wholy correct. After I get --follow
566 working properly, I plan to revise the various documentation and/or
567 the code to make sure the two match.