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