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