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