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