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