[thirdparty/mdadm.git] / ChangeLog
1Changes Prior to 1.4.0 release
2 - Document fact that creating a raid5 array really creates a
3 degraded array with a spare.
4 - Add "spares=" tag to config file and generate it wit --detail and
5 --examine
6 - Add "SparesMissing" event when --monitor first sees an array and
7 it doesn't have the enough spare devices.
8 - Add --update=summaries for --assemble to update summary
9 information in superblock, and correct other inconsistancies in
10 the superblock.
11 - Add --test option to --detail to set a meaningful exit status.
13Changes Prior to 1.3.0 release
14 - Make 'size' and unsigned long in Create to allow creation of
15 larger arrays.
16 - Explicitly flag spare devices as 'spare' in --detail and --examine
17 output. Previously they simply had no flags lists.
18 - Make MailCmd (for monitor) configurable in Makefile, and default
19 to "/usr/sbin/sendmail -t". Also split out the warning related
20 flags into CWFLAGS for easier build configurability.
21 - Minor bugfix in Manage code.
22 - --monitor now notices and reports degraded arrays at startup using
23 "DegradedArray" event, and also has a --oneshot option to only
24 report DegradedArrays, and then exit.
25 - Small man-page clarification w.r.t. raid levels and raid4 in
26 particular.
27 - Disallow creation of arrays with only one device as this is
28 probably a mistake. --force will override this check.
29 - Correct some misleading documentation in the "mdadm --create --help"
30 message.
31 - Ignore chunksize if raid1 or multipath.
32 - Explicit statement in man page that raid-disks cannot be changed
33 after array is created.
34 - Improve message when attempting to start an array with
35 insufficient devices. Instead of required the array to be full,
36 we only require it has as many active devices as last time.
38Changes Prior to 1.2.0 release
39 - Fix bug where --daemonise required an argument.
40 - In --assemble --verbose, print appropriate message if device is
41 not in devices= list
42 - Updated mdadm.conf.5 to reflect fact that device= takes wildcards
43 - Typos: componenet -> component
44 - Reduce size of "--help" message put excess into "--help-options"
45 - Fix bug introduced when MD_SB_DISKS dependancy removed, and which
46 caused spares not be assembled properly.
47 - Print appropriate message if --monitor --scan decides not to
48 monitor anything.
49Changes Prior to 1.1.0 release
50 - add --deamonise flag for --monitor - forks and prints pid to stdout
51 - Fix bug so we REALLY clear dirty flag with -Af
52 - -Db now prints a 'devices=' word for each array.
53 - "mdadm -A /dev/md0" will get info from configfile, even without scan
54 - When assembling multipath arrays, ignore devices which are flagged
55 as having errors.
56 - take --super-minor=dev to mean "use the minor number of the mddev
57 being assembled.
58 - take --config=none to mean "completely ignore config file"
59 - Make --monitor require --scan or a device list.
60Changes Prior to 1.0.9 release
61 - Documentation updates including kernel parameters documented
62 in md.4
63 - --assemble --force for raid4/5 will mark clean, needed for 2.5
64 - --detail prints out the events counter as well
65 - flush device before reading superblock to be sure to get
66 current data
67 - added mdadm.static target to makefile for static linking
68 - --monitor was ignoring /dev/md0 due to off-by-one error
69 - Fix assorted typos
70 - Fix printing of Gibibytes - calc was wrong.
71 - Fix printing of Array Size in --detail when very big.
72 - --monitor no longer tries to work for raid0 or linear as these
d013a55e 73 have nothing to be monitored.
74 - The word 'partitions' on a DEVICE line will cause all partitions
75 listed in /proc/partitions to be considered
76 - If the config file is called 'partitions' then it will be treated
77 as though it contained exactly 'device partitions' so e.g.
78 mdadm -Ebsc partitions
d013a55e 79 will find all raid partitions easily.
80 - successfully assemble multipath devices by ignoring raid_disk
81 value from superblock (it is always the same).
82 - --assemble not tied to MD_SB_DISKS limit quite so much
83 - Support compiling with tcc
84 - Support compiling with uclibc - just skip scan of /dev
85 - Add --update= option for Assemble mode. Either sparc2.2
86 or super-minor updates are possible. See mdadm.8
88Changes Prior to 1.0.1 release
89 - Round off MB/GiB etc values instead of round down.
90 - Add --sparc2.2 option to examine to shift superblock around
91 and --sparc2.2update to rewrite the superblock
5787fa49 92 - Fix assorted typos in online help
94Changes Prior to 1.0.0 release
95 - Allow --config with Misc mode (for --examine --scan)
96 - Add $(CXFLAGS) to end of CFLAGS in makefile
97 - When making an N disk raid5 array, the Nth drive
98 is moved to the end of the array as a spare rather than
99 being shifted up one place. This means that when the
100 kernel builds onto the last spare and inserts it,
101 the devices will be in the expected order.
102 - Man page improvements
103Changes Prior to 0.8.2 release
104 - Correct spelling of persist[ae]nce/persist[ae]nt.
105 - Change "disk" to "device" in options and config file
106 - convert array size to "long long" *before* shift-left in -D and -Q
108Changes Prior to 0.8.1 release
109 - Add "INSTALL" file.
110 - Fix some "i" variables that were not being set properly
111 - Initialise minsize and maxsize so that compilers don't complain.
112 - Tidy up Makefile and mdadm.spec installations
113 - Add "multipath" to documentation of valid levels
115Changes Prior to 0.8 release
116 - Fix another bug in Assemble.c due to confusing 'i' with 'j'
117 - Minimal, untested, support for multipath
118 - re-write of argument parsing to have more coherent modes,
119 - add --query,-Q option
120 - Update mdadm.8 to reflect arg processing change and --query
121 - Change "long" to "unsigned long" for device sizes
122 - Handle "mailaddr" and "program" lines in config file for follow/scan mode.
123 - --follow --scan will exit if no program or mail found
124 - Add MAILADDR and PROGRAM to mdadm.conf-example
125 - Spell check man pages
126 - consistently use "component devices" instead of "subdevices"
127 - Make -Wall -Werror really work and fix lots of errors.
128 - --detail and --stop can have --scan which chooses devices from /proc/mdstat
129 - --monitor detects 20% changes in resync, failed spares,
130 disappearing arrays,
131 - --monitor --scan will automatically add any devices found in /proc/mdstat
132 - --monitor will move spares between arrays with same spare-group if necessary
133 - Documentation for Monitor Mode
134 - --query notes if the array containing the given device is active or not
135 - Finished md.4 man page.
137Changes Prior to 0.7.2 release
138 - mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra --
139 - more mdadm.spec updates from Gregory Leblanc <>
140 - make directory for mdadm.conf configurable in Makefile
141 - Finished mdadm.conf.5. Removed details of conf file from
142 mdadm.8 leaving a reference to mdadm.conf.5.
143 - Fix bug in Assemble.c, thanks to Junaid Rizvi <>
144 - Get --assemble --force to make sure old major/minor numbers are
145 consistant, as md.c worries about this :-(
148Changes Prior to 0.7.1 release
149 - update mdadm.spec
150 - use BLKGETSIZE64 if available for array size
151 - give human readable as GiB/MiB and GB and MB, with 2 decimal point precision
152 - Only warn about size variation for raid1/4/5.
153 - Started md.4 man page
154 - Started mdadm.conf.5 man page
156Changes Prior to 0.7 release
158 - Fix makefile to install binary at /sbin and not /sbin/sbin
159 Also install man page.
160 - Add --zero-superblock based on --destroywithextremeprejudice
161 from Dale Stephenson <>
162 - change name to mdadm. It is palandromic, and much nicer to pronouce.
164Changes Prior to 0.6 release
9a9dab36 166 - Remove the limit on the number of device names that can be
167 given on the command line.
168 - Fix bug in --assemble --force where it would only update a
169 single superblock.
170 - Fix bogus printing of big numbers not being block devices
171 when given names of devices that don't exist.
172 - When --assemble --force, consider superblocks with an event
173 count that is 1 behind as out-of-date. Normally they are
174 considered up-to-date (as the kernel assumes this too).
175 - When marking drives as not-failed in the superblock,
176 we also mark them as ACTIVE and SYNC.
177 - Don't start arrays for which not all drives are available unless:
178 --scan which implies that all drives were found automatically
179 --run which means the user knows what they want
180 --force which means that we are fixing something broken
181 - Make sure all device numbers passed as 3rd arg of ioctl
182 are passed as unsigned lock, so that it works on SPARC
183 - If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK
184 if we cannot read from the array, i.e. if the array is
185 not started yet.
186 - man page update
187 - Taught Examine to handle --scan. It examines all devices listed
188 on DEVICE lines in the config file.
189 - Added --brief (-b) flag for Examine and Detail to print out
190 and mdctl.conf compatible description with uuid=, level=,
191 disks= and - for Examine - devices=
192 --examine --brief collects all devices the make the one array and
193 list them as one entry.
194 - Added level= and disks= options to ARRAY lines in config files
195 so --brief output could be used as-is.
196 - Make parity style ({left,right}-{,a}symmetric) consistantly use -,
197 never _.
198 - Add "Array Size" to --detail output
199 - Change "Size" to "Device Size" and exclude from Detail of arrays
200 that do not have a consistent device size.
201 - Add Human readable MiB or GiB value on size lines of Detail and Examine
202 - --assemble --scan doesn't complain about active drives
203 - require number of spares given in -x to be listed.
204 - Made --build actually work.
205Changes Prior to 0.5 release
207 --assemble:
208 spare drives are handled properly.
210 --force can be used to recover from 2-drive failures on RAID5
211 If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should
212 make a raid5 array, but it has experienced multiple failures and
213 wont start, then
215 mdctl --assemble --force /dev/md0 /dev/hd[abcd]1
217 Should update the superblock on the newest failed drive and
218 restart the array in degraded mode. You should then remove the
219 remaining failed drive and re-add it (if you are happy that it
220 might work).
222 Ofcourse whenever you have a 2-drive failure, you have a risk
223 of corruption in data that hasn't be changed for a long time. So
224 this doesn't give you your array back all nice and happy, but it
225 does allow you to recover data that might not be corrupt.
227 More flexibility in identifying a RAID array in the mdctl.conf
228 e.g.
229 array /dev/md4 super-minor=4
231 assembles /dev/md4 from all devices found that have a raid
232 superblock that says the minor number of the array is 4.
233 If the blocks with the right minor number do not all have the
234 same UUID, an error is flags and no assembly happens.
236 array /dev/md3 devices=/dev/hd[abc]2
238 Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All
239 devices must exist and have raid superblock with the same uuid.
241 If two identity specifiers are used, only devices that match all
242 of them are considered, so
244 array /dev/md2 devices=/dev/hd?2 super-minor=2
246 will assemble /dev/md2 using all /dev/hd?2 devices which have a
247 raid superblock with minor number 2.
249 --create:
250 When listing devices for --create, the word "missing" can be
251 used to indicate that the respective slot does not have a
252 working drive currently. This is similar to the "failed-disk"
253 directive in mkraid/raidtab.
254 e.g.
255 mdctl --create --level=5 -raid-disks=4 --spare-disks=2
256 /dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde
258 will create a raid5 array with the third slot empty, and two
259 spares.
261 By default, raid5 arrays are created with the last slot empty
262 and drive listed for the last slot added as a spare. If a
263 "missing" slot is given, or if --force is given, then --create
264 does exactly what you ask and doesn't try to be clever.
267 --follow / --monitor:
269 This is a new mode. I couldn't stop my self from picking a name
270 starting with F (as current modes start A,B,C,D,E) but I
271 relented and provided an alternate name that is somewhat more
272 meaningful.
274 In this mode, mdctl does not exit, but runs continuously and
275 periodically polls all the md devices to see if they have had
276 any interested state change.
277 The changes that it currently notices are:
278 Fail - an active disc fails
279 FailSpare - a spare, that was presumably being build, fails
280 ActiveSpare - a spare becomes active, presumably after a rebuild.
282 Options:
283 --mail mailaddress - send Email on any Fail* event
284 --program program - run the program on any event.
285 Args are: eventname mddevice subdevice(if-known)
286 --delay seconds - change from the default 60second pause
287 between polls.
289 I plan to add functionality to this mode to allow sharing of
290 spare drives. If an array is marks "spare-group=fred", and it
291 has a failed drive and no spares, and if some other array is
292 also "spare-group=fred" and it has no failed drives, but does
293 have a spare drive that is big enough, the spare will be moved
294 to the first array.
296 I also have the idea of adding a --grow mode which will re-organise
297 the data on an N disk raid0/4/5 array to be on an N+M disk array.
298 I have no concrete plans for this though.
300 I got rid of the "v" in the archive file name, and include the
301 version number in the directory created by the archive.
303 There is now a man page and mdctl.spec (for rpm) thanks to
304 Danilo Godec <>.
306 Ofcourse, the man page is now out of date and despite being based on
307 the --help output, is not wholy correct. After I get --follow
308 working properly, I plan to revise the various documentation and/or
309 the code to make sure the two match.