Change MAJOR() etc to major() etc
[thirdparty/mdadm.git] / ChangeLog
1Changes Prior to 1.9.0 release
2 - Fix rpm build problem (stray %)
19f8b8fc 3 - Minor manpage updates
182661f4 4 - Change "dirty" status to "active" as it was confusing people.
5 - --assemble --auto recognises 'standard' name and insists on using
6 the appropriate major/minor number for them.
7 - Remove underscore from partition names, so partitions of
8 "foo" are "foo1", "foo2" etc (unchanged) and partitions of
9 "f00" are "f00p1", "f00p2" etc rather than "f00_p1"...
10 - Use "major", "minor", "makedev" macros instead of
11 "MAJOR", "MINOR", "MKDEV" so that large device numbers work
12 on 2.6 (providing you have glibc 2.3.3 or later).
9a4e08d2 13
14Changes Prior to 1.8.0 release
15 - Makefile cleanup from Luca Berra <>
16 - --pid-file (-i) to set a pid file to use with --monitor --daemonise
17 - Fix typo in mdadm man page
18 - Fix coredump when "-s" used with no config file present.
19 - Support new "faulty" personality which can inject synthetic
20 faults. (Not in yet at 1Nov2004)
21 - Support raid0/linear on devices > 2 Terabytes
22 - Make sure raid6 resyncs when created with one missing device
24Changes Prior to 1.7.0 release
25 - Support "--grow --add" to add a device to a linear array, if the
26 kernel supports it. Not documented yet.
27 - Restore support for uclibc which was broken recently.
28 - Several improvements to the output of --detail, including
29 reporting "resyncing" or "recovering" in the state.
30 - Close filedescriptor at end of --detail (exit would have closed it
31 anyway, so this isn't abig deal).
32 - Report "Sync checkpoint" in --examine output if appropriate.
33 - Add --update=resync for --assemble mode to for a resync when the
34 array is assembled.
35 - Add support for "raid10", which is under development in 2.6.
36 Not documented yet.
37 - --monitor now reads spare-group and spares info from config file
38 even when names of arrays to scan are given on the command line
40Changes Prior to 1.6.0 release
41 - Device name given in -Eb is determined by examining /dev rather
42 than assuming /dev/md%d
43 - Fix bug in --monitor where an array could be held open an so
44 could not be stopped without killing mdadm.
45 - Add --grow mode. Currently only --size and --raid-disks can be
46 changed. Both require kernel support which, at the time of
47 writing, is not in a release kernel yet.
48 - Don't print out "errors" or "no-errors" in -D and -E, as the bit
49 is never set or used.
50 - Use md event notification in 2.6.??? to make --monitor mode
51 respond instantly to events.
52 - Add --auto= option and auto= configfile entry to tell mdadm to
53 create device files as needed. This is particularly useful
54 with partitioned arrays where the major device number can change.
55 - When generating --brief listing, if the standard name doesn't
56 exist, search /dev for one rather than using a temp name.
57 - Allow --build to build raid1 and multipath arrays.
58 - Add "--assume-clean" for Create and Build, particularly for raid1
59 Note: this is dangerous. Only use it if you are certain.
60 - Fix bug so that Rebuild status monitoring works again.
61 - Add "degraded" and "recovering" options to the "Status:"
62 entry for --detail
64Changes Prior to 1.5.0 release
65 - new commands "mdassemble" which is a stripped-down equivalent of
66 "mdadm -As", that can be compiled with dietlibc.
67 Thanks to Luca Berra <>.
68 It can be using in an initramfs or initrd.
69 - Fix compiling error with BLKGETSIZE64 and some signed/unsigned
70 comparison warnings.
71 - Add Rebuild Status (% complete) to --detail output.
72 - Support "--monitor --test" which will generate a test alert
73 for each array once, to test notification paths.
74 - Generate RebuildFinished event when rebuild finishes.
75 - Support for raid6 as found in 2.6.2 - thanks to
76 H. Peter Anvin <>
77 - Support partitioned md arrays with a different major number and
78 naming scheme (md_dX in /proc/mdstat, /dev/md/dXpY in /dev).
80Changes Prior to 1.4.0 release
81 - Document fact that creating a raid5 array really creates a
82 degraded array with a spare.
83 - Add "spares=" tag to config file and generate it wit --detail and
84 --examine
85 - Add "SparesMissing" event when --monitor first sees an array and
86 it doesn't have the enough spare devices.
87 - Add --update=summaries for --assemble to update summary
88 information in superblock, and correct other inconsistancies in
89 the superblock.
90 - Add --test option to --detail to set a meaningful exit status.
92Changes Prior to 1.3.0 release
93 - Make 'size' and unsigned long in Create to allow creation of
94 larger arrays.
95 - Explicitly flag spare devices as 'spare' in --detail and --examine
96 output. Previously they simply had no flags lists.
97 - Make MailCmd (for monitor) configurable in Makefile, and default
98 to "/usr/sbin/sendmail -t". Also split out the warning related
99 flags into CWFLAGS for easier build configurability.
100 - Minor bugfix in Manage code.
101 - --monitor now notices and reports degraded arrays at startup using
102 "DegradedArray" event, and also has a --oneshot option to only
103 report DegradedArrays, and then exit.
104 - Small man-page clarification w.r.t. raid levels and raid4 in
105 particular.
106 - Disallow creation of arrays with only one device as this is
107 probably a mistake. --force will override this check.
108 - Correct some misleading documentation in the "mdadm --create --help"
109 message.
110 - Ignore chunksize if raid1 or multipath.
111 - Explicit statement in man page that raid-disks cannot be changed
112 after array is created.
113 - Improve message when attempting to start an array with
114 insufficient devices. Instead of required the array to be full,
115 we only require it has as many active devices as last time.
117Changes Prior to 1.2.0 release
118 - Fix bug where --daemonise required an argument.
119 - In --assemble --verbose, print appropriate message if device is
120 not in devices= list
121 - Updated mdadm.conf.5 to reflect fact that device= takes wildcards
122 - Typos: componenet -> component
123 - Reduce size of "--help" message put excess into "--help-options"
124 - Fix bug introduced when MD_SB_DISKS dependancy removed, and which
125 caused spares not be assembled properly.
126 - Print appropriate message if --monitor --scan decides not to
127 monitor anything.
128Changes Prior to 1.1.0 release
129 - add --deamonise flag for --monitor - forks and prints pid to stdout
130 - Fix bug so we REALLY clear dirty flag with -Af
131 - -Db now prints a 'devices=' word for each array.
132 - "mdadm -A /dev/md0" will get info from configfile, even without scan
133 - When assembling multipath arrays, ignore devices which are flagged
134 as having errors.
135 - take --super-minor=dev to mean "use the minor number of the mddev
136 being assembled.
137 - take --config=none to mean "completely ignore config file"
138 - Make --monitor require --scan or a device list.
139Changes Prior to 1.0.9 release
140 - Documentation updates including kernel parameters documented
141 in md.4
142 - --assemble --force for raid4/5 will mark clean, needed for 2.5
143 - --detail prints out the events counter as well
144 - flush device before reading superblock to be sure to get
145 current data
146 - added mdadm.static target to makefile for static linking
147 - --monitor was ignoring /dev/md0 due to off-by-one error
148 - Fix assorted typos
149 - Fix printing of Gibibytes - calc was wrong.
150 - Fix printing of Array Size in --detail when very big.
151 - --monitor no longer tries to work for raid0 or linear as these
d013a55e 152 have nothing to be monitored.
153 - The word 'partitions' on a DEVICE line will cause all partitions
154 listed in /proc/partitions to be considered
155 - If the config file is called 'partitions' then it will be treated
156 as though it contained exactly 'device partitions' so e.g.
157 mdadm -Ebsc partitions
d013a55e 158 will find all raid partitions easily.
159 - successfully assemble multipath devices by ignoring raid_disk
160 value from superblock (it is always the same).
161 - --assemble not tied to MD_SB_DISKS limit quite so much
162 - Support compiling with tcc
163 - Support compiling with uclibc - just skip scan of /dev
164 - Add --update= option for Assemble mode. Either sparc2.2
165 or super-minor updates are possible. See mdadm.8
167Changes Prior to 1.0.1 release
168 - Round off MB/GiB etc values instead of round down.
169 - Add --sparc2.2 option to examine to shift superblock around
170 and --sparc2.2update to rewrite the superblock
5787fa49 171 - Fix assorted typos in online help
173Changes Prior to 1.0.0 release
174 - Allow --config with Misc mode (for --examine --scan)
175 - Add $(CXFLAGS) to end of CFLAGS in makefile
176 - When making an N disk raid5 array, the Nth drive
177 is moved to the end of the array as a spare rather than
178 being shifted up one place. This means that when the
179 kernel builds onto the last spare and inserts it,
180 the devices will be in the expected order.
181 - Man page improvements
182Changes Prior to 0.8.2 release
183 - Correct spelling of persist[ae]nce/persist[ae]nt.
184 - Change "disk" to "device" in options and config file
185 - convert array size to "long long" *before* shift-left in -D and -Q
187Changes Prior to 0.8.1 release
188 - Add "INSTALL" file.
189 - Fix some "i" variables that were not being set properly
190 - Initialise minsize and maxsize so that compilers don't complain.
191 - Tidy up Makefile and mdadm.spec installations
192 - Add "multipath" to documentation of valid levels
194Changes Prior to 0.8 release
195 - Fix another bug in Assemble.c due to confusing 'i' with 'j'
196 - Minimal, untested, support for multipath
197 - re-write of argument parsing to have more coherent modes,
198 - add --query,-Q option
199 - Update mdadm.8 to reflect arg processing change and --query
200 - Change "long" to "unsigned long" for device sizes
201 - Handle "mailaddr" and "program" lines in config file for follow/scan mode.
202 - --follow --scan will exit if no program or mail found
203 - Add MAILADDR and PROGRAM to mdadm.conf-example
204 - Spell check man pages
205 - consistently use "component devices" instead of "subdevices"
206 - Make -Wall -Werror really work and fix lots of errors.
207 - --detail and --stop can have --scan which chooses devices from /proc/mdstat
208 - --monitor detects 20% changes in resync, failed spares,
209 disappearing arrays,
210 - --monitor --scan will automatically add any devices found in /proc/mdstat
211 - --monitor will move spares between arrays with same spare-group if necessary
212 - Documentation for Monitor Mode
213 - --query notes if the array containing the given device is active or not
214 - Finished md.4 man page.
216Changes Prior to 0.7.2 release
217 - mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra --
218 - more mdadm.spec updates from Gregory Leblanc <>
219 - make directory for mdadm.conf configurable in Makefile
220 - Finished mdadm.conf.5. Removed details of conf file from
221 mdadm.8 leaving a reference to mdadm.conf.5.
222 - Fix bug in Assemble.c, thanks to Junaid Rizvi <>
223 - Get --assemble --force to make sure old major/minor numbers are
224 consistant, as md.c worries about this :-(
227Changes Prior to 0.7.1 release
228 - update mdadm.spec
229 - use BLKGETSIZE64 if available for array size
230 - give human readable as GiB/MiB and GB and MB, with 2 decimal point precision
231 - Only warn about size variation for raid1/4/5.
232 - Started md.4 man page
233 - Started mdadm.conf.5 man page
235Changes Prior to 0.7 release
237 - Fix makefile to install binary at /sbin and not /sbin/sbin
238 Also install man page.
239 - Add --zero-superblock based on --destroywithextremeprejudice
240 from Dale Stephenson <>
241 - change name to mdadm. It is palandromic, and much nicer to pronouce.
243Changes Prior to 0.6 release
9a9dab36 245 - Remove the limit on the number of device names that can be
246 given on the command line.
247 - Fix bug in --assemble --force where it would only update a
248 single superblock.
249 - Fix bogus printing of big numbers not being block devices
250 when given names of devices that don't exist.
251 - When --assemble --force, consider superblocks with an event
252 count that is 1 behind as out-of-date. Normally they are
253 considered up-to-date (as the kernel assumes this too).
254 - When marking drives as not-failed in the superblock,
255 we also mark them as ACTIVE and SYNC.
256 - Don't start arrays for which not all drives are available unless:
257 --scan which implies that all drives were found automatically
258 --run which means the user knows what they want
259 --force which means that we are fixing something broken
260 - Make sure all device numbers passed as 3rd arg of ioctl
261 are passed as unsigned lock, so that it works on SPARC
262 - If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK
263 if we cannot read from the array, i.e. if the array is
264 not started yet.
265 - man page update
266 - Taught Examine to handle --scan. It examines all devices listed
267 on DEVICE lines in the config file.
268 - Added --brief (-b) flag for Examine and Detail to print out
269 and mdctl.conf compatible description with uuid=, level=,
270 disks= and - for Examine - devices=
271 --examine --brief collects all devices the make the one array and
272 list them as one entry.
273 - Added level= and disks= options to ARRAY lines in config files
274 so --brief output could be used as-is.
275 - Make parity style ({left,right}-{,a}symmetric) consistantly use -,
276 never _.
277 - Add "Array Size" to --detail output
278 - Change "Size" to "Device Size" and exclude from Detail of arrays
279 that do not have a consistent device size.
280 - Add Human readable MiB or GiB value on size lines of Detail and Examine
281 - --assemble --scan doesn't complain about active drives
282 - require number of spares given in -x to be listed.
283 - Made --build actually work.
284Changes Prior to 0.5 release
286 --assemble:
287 spare drives are handled properly.
289 --force can be used to recover from 2-drive failures on RAID5
290 If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should
291 make a raid5 array, but it has experienced multiple failures and
292 wont start, then
294 mdctl --assemble --force /dev/md0 /dev/hd[abcd]1
296 Should update the superblock on the newest failed drive and
297 restart the array in degraded mode. You should then remove the
298 remaining failed drive and re-add it (if you are happy that it
299 might work).
301 Ofcourse whenever you have a 2-drive failure, you have a risk
302 of corruption in data that hasn't be changed for a long time. So
303 this doesn't give you your array back all nice and happy, but it
304 does allow you to recover data that might not be corrupt.
306 More flexibility in identifying a RAID array in the mdctl.conf
307 e.g.
308 array /dev/md4 super-minor=4
310 assembles /dev/md4 from all devices found that have a raid
311 superblock that says the minor number of the array is 4.
312 If the blocks with the right minor number do not all have the
313 same UUID, an error is flags and no assembly happens.
315 array /dev/md3 devices=/dev/hd[abc]2
317 Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All
318 devices must exist and have raid superblock with the same uuid.
320 If two identity specifiers are used, only devices that match all
321 of them are considered, so
323 array /dev/md2 devices=/dev/hd?2 super-minor=2
325 will assemble /dev/md2 using all /dev/hd?2 devices which have a
326 raid superblock with minor number 2.
328 --create:
329 When listing devices for --create, the word "missing" can be
330 used to indicate that the respective slot does not have a
331 working drive currently. This is similar to the "failed-disk"
332 directive in mkraid/raidtab.
333 e.g.
334 mdctl --create --level=5 -raid-disks=4 --spare-disks=2
335 /dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde
337 will create a raid5 array with the third slot empty, and two
338 spares.
340 By default, raid5 arrays are created with the last slot empty
341 and drive listed for the last slot added as a spare. If a
342 "missing" slot is given, or if --force is given, then --create
343 does exactly what you ask and doesn't try to be clever.
346 --follow / --monitor:
348 This is a new mode. I couldn't stop my self from picking a name
349 starting with F (as current modes start A,B,C,D,E) but I
350 relented and provided an alternate name that is somewhat more
351 meaningful.
353 In this mode, mdctl does not exit, but runs continuously and
354 periodically polls all the md devices to see if they have had
355 any interested state change.
356 The changes that it currently notices are:
357 Fail - an active disc fails
358 FailSpare - a spare, that was presumably being build, fails
359 ActiveSpare - a spare becomes active, presumably after a rebuild.
361 Options:
362 --mail mailaddress - send Email on any Fail* event
363 --program program - run the program on any event.
364 Args are: eventname mddevice subdevice(if-known)
365 --delay seconds - change from the default 60second pause
366 between polls.
368 I plan to add functionality to this mode to allow sharing of
369 spare drives. If an array is marks "spare-group=fred", and it
370 has a failed drive and no spares, and if some other array is
371 also "spare-group=fred" and it has no failed drives, but does
372 have a spare drive that is big enough, the spare will be moved
373 to the first array.
375 I also have the idea of adding a --grow mode which will re-organise
376 the data on an N disk raid0/4/5 array to be on an N+M disk array.
377 I have no concrete plans for this though.
379 I got rid of the "v" in the archive file name, and include the
380 version number in the directory created by the archive.
382 There is now a man page and mdctl.spec (for rpm) thanks to
383 Danilo Godec <>.
385 Ofcourse, the man page is now out of date and despite being based on
386 the --help output, is not wholy correct. After I get --follow
387 working properly, I plan to revise the various documentation and/or
388 the code to make sure the two match.