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