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