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