1
Changes Prior to 1.12.0 release
2
Several of these are backported from the Debian package
3
- Don't use 'lstat' to check for blockdevices, use stat.
4
- Get size calculation right when in Create when BLKGESIZE64 not available
5
- Document --size=max option for --grow
6
- Document SparesMissing event and DeviceDisappeared/WrongLevel
7
- --stop --scan repeatly cycles until no more progress can be made
8
so that stack devices are stopped properly
9
- Minor type rationalisation for ident->uuid - now always 'int[]'
10
- Fix type in online help for --grow (was -F, now -G)
11
- Allow --auto command line option to set default auto=
12
value when running "--assemble --scan". Previously
13
--auto was ignored if --scan was given
14
- Fix a few type casts
15
- Fix parsing of /dev/md/N in is_standard
16
- Fix rounding errors in human_size()
17
- Fix silly example in mdadm.conf-examples
18
- When finding a /dev name for a device, prefer shorter names
19
- Suppress listing of devices= in --brief output of -D or -E,
20
unless -v is given (-vv gives the old behaviour with -Dsv).
21
This is because the device list can change and so is not a
22
stable aspect of the array
23
- Allow --force with --grow so '-Gfn1' works (on raid1)
24
- Make --examine --brief report spares= properly.
25
- When device files are made with "--assemble --auto",
26
add them to the list of known device name, so that
27
assembling stacked devices has a better chance of working.
28
- Replace sprintf calls with snprintf (to quiet diet-libc)
29
- Add man page for mdassemble
30
- Fix compilation with tinyc
31
- Lot of minor manpage updates
33
Changes Prior to 1.11.0 release
34
- Fix embarassing bug which causes --add to always fail.
36
Changes Prior to 1.10.0 release
37
- Fix bug with --config=partitions
38
- Open sub-devices with O_EXCL to detect if already in use
39
- Make sure superblock updates are flushed directly to disk.
41
Changes Prior to 1.9.0 release
42
- Fix rpm build problem (stray %)
43
- Minor manpage updates
44
- Change "dirty" status to "active" as it was confusing people.
45
- --assemble --auto recognises 'standard' name and insists on using
46
the appropriate major/minor number for them.
47
- Remove underscore from partition names, so partitions of
48
"foo" are "foo1", "foo2" etc (unchanged) and partitions of
49
"f00" are "f00p1", "f00p2" etc rather than "f00_p1"...
50
- Use "major", "minor", "makedev" macros instead of
51
"MAJOR", "MINOR", "MKDEV" so that large device numbers work
52
on 2.6 (providing you have glibc 2.3.3 or later).
53
- Add some missing closes of open file descriptors.
54
- Reread /proc/partition for every array assembled when using
55
it to find devices, rather than only once.
56
- Make "mdadm -Ss" stop stacked devices properly, by reversing the
57
order in which arrays are stopped.
58
- Improve some error messages.
59
- Allow device name to appear before first option, so e.g.
60
mdadm /dev/md0 -A /dev/sd[ab]
62
- Assume '-Q' if just a device is given, rather than being silent.
64
Changes Prior to 1.8.0 release
65
- Makefile cleanup from Luca Berra <bluca@comedia.it>
66
- --pid-file (-i) to set a pid file to use with --monitor --daemonise
67
- Fix typo in mdadm man page
68
- Fix coredump when "-s" used with no config file present.
69
- Support new "faulty" personality which can inject synthetic
70
faults. (Not in kernel.org yet at 1Nov2004)
71
- Support raid0/linear on devices > 2 Terabytes
72
- Make sure raid6 resyncs when created with one missing device
74
Changes Prior to 1.7.0 release
75
- Support "--grow --add" to add a device to a linear array, if the
76
kernel supports it. Not documented yet.
77
- Restore support for uclibc which was broken recently.
78
- Several improvements to the output of --detail, including
79
reporting "resyncing" or "recovering" in the state.
80
- Close filedescriptor at end of --detail (exit would have closed it
81
anyway, so this isn't abig deal).
82
- Report "Sync checkpoint" in --examine output if appropriate.
83
- Add --update=resync for --assemble mode to for a resync when the
85
- Add support for "raid10", which is under development in 2.6.
87
- --monitor now reads spare-group and spares info from config file
88
even when names of arrays to scan are given on the command line
90
Changes Prior to 1.6.0 release
91
- Device name given in -Eb is determined by examining /dev rather
92
than assuming /dev/md%d
93
- Fix bug in --monitor where an array could be held open an so
94
could not be stopped without killing mdadm.
95
- Add --grow mode. Currently only --size and --raid-disks can be
96
changed. Both require kernel support which, at the time of
97
writing, is not in a release kernel yet.
98
- Don't print out "errors" or "no-errors" in -D and -E, as the bit
100
- Use md event notification in 2.6.??? to make --monitor mode
101
respond instantly to events.
102
- Add --auto= option and auto= configfile entry to tell mdadm to
103
create device files as needed. This is particularly useful
104
with partitioned arrays where the major device number can change.
105
- When generating --brief listing, if the standard name doesn't
106
exist, search /dev for one rather than using a temp name.
107
- Allow --build to build raid1 and multipath arrays.
108
- Add "--assume-clean" for Create and Build, particularly for raid1
109
Note: this is dangerous. Only use it if you are certain.
110
- Fix bug so that Rebuild status monitoring works again.
111
- Add "degraded" and "recovering" options to the "Status:"
114
Changes Prior to 1.5.0 release
115
- new commands "mdassemble" which is a stripped-down equivalent of
116
"mdadm -As", that can be compiled with dietlibc.
117
Thanks to Luca Berra <bluca@comedia.it>.
118
It can be using in an initramfs or initrd.
119
- Fix compiling error with BLKGETSIZE64 and some signed/unsigned
121
- Add Rebuild Status (% complete) to --detail output.
122
- Support "--monitor --test" which will generate a test alert
123
for each array once, to test notification paths.
124
- Generate RebuildFinished event when rebuild finishes.
125
- Support for raid6 as found in 2.6.2 - thanks to
126
H. Peter Anvin <hpa@zytor.com>
127
- Support partitioned md arrays with a different major number and
128
naming scheme (md_dX in /proc/mdstat, /dev/md/dXpY in /dev).
130
Changes Prior to 1.4.0 release
131
- Document fact that creating a raid5 array really creates a
132
degraded array with a spare.
133
- Add "spares=" tag to config file and generate it wit --detail and
135
- Add "SparesMissing" event when --monitor first sees an array and
136
it doesn't have the enough spare devices.
137
- Add --update=summaries for --assemble to update summary
138
information in superblock, and correct other inconsistancies in
140
- Add --test option to --detail to set a meaningful exit status.
142
Changes Prior to 1.3.0 release
143
- Make 'size' and unsigned long in Create to allow creation of
145
- Explicitly flag spare devices as 'spare' in --detail and --examine
146
output. Previously they simply had no flags lists.
147
- Make MailCmd (for monitor) configurable in Makefile, and default
148
to "/usr/sbin/sendmail -t". Also split out the warning related
149
flags into CWFLAGS for easier build configurability.
150
- Minor bugfix in Manage code.
151
- --monitor now notices and reports degraded arrays at startup using
152
"DegradedArray" event, and also has a --oneshot option to only
153
report DegradedArrays, and then exit.
154
- Small man-page clarification w.r.t. raid levels and raid4 in
156
- Disallow creation of arrays with only one device as this is
157
probably a mistake. --force will override this check.
158
- Correct some misleading documentation in the "mdadm --create --help"
160
- Ignore chunksize if raid1 or multipath.
161
- Explicit statement in man page that raid-disks cannot be changed
162
after array is created.
163
- Improve message when attempting to start an array with
164
insufficient devices. Instead of required the array to be full,
165
we only require it has as many active devices as last time.
167
Changes Prior to 1.2.0 release
168
- Fix bug where --daemonise required an argument.
169
- In --assemble --verbose, print appropriate message if device is
171
- Updated mdadm.conf.5 to reflect fact that device= takes wildcards
172
- Typos: componenet -> component
173
- Reduce size of "--help" message put excess into "--help-options"
174
- Fix bug introduced when MD_SB_DISKS dependancy removed, and which
175
caused spares not be assembled properly.
176
- Print appropriate message if --monitor --scan decides not to
178
Changes Prior to 1.1.0 release
179
- add --deamonise flag for --monitor - forks and prints pid to stdout
180
- Fix bug so we REALLY clear dirty flag with -Af
181
- -Db now prints a 'devices=' word for each array.
182
- "mdadm -A /dev/md0" will get info from configfile, even without scan
183
- When assembling multipath arrays, ignore devices which are flagged
185
- take --super-minor=dev to mean "use the minor number of the mddev
187
- take --config=none to mean "completely ignore config file"
188
- Make --monitor require --scan or a device list.
189
Changes Prior to 1.0.9 release
190
- Documentation updates including kernel parameters documented
192
- --assemble --force for raid4/5 will mark clean, needed for 2.5
193
- --detail prints out the events counter as well
194
- flush device before reading superblock to be sure to get
196
- added mdadm.static target to makefile for static linking
197
- --monitor was ignoring /dev/md0 due to off-by-one error
199
- Fix printing of Gibibytes - calc was wrong.
200
- Fix printing of Array Size in --detail when very big.
201
- --monitor no longer tries to work for raid0 or linear as these
202
have nothing to be monitored.
203
- The word 'partitions' on a DEVICE line will cause all partitions
204
listed in /proc/partitions to be considered
205
- If the config file is called 'partitions' then it will be treated
206
as though it contained exactly 'device partitions' so e.g.
207
mdadm -Ebsc partitions
208
will find all raid partitions easily.
209
- successfully assemble multipath devices by ignoring raid_disk
210
value from superblock (it is always the same).
211
- --assemble not tied to MD_SB_DISKS limit quite so much
212
- Support compiling with tcc
213
- Support compiling with uclibc - just skip scan of /dev
214
- Add --update= option for Assemble mode. Either sparc2.2
215
or super-minor updates are possible. See mdadm.8
217
Changes Prior to 1.0.1 release
218
- Round off MB/GiB etc values instead of round down.
219
- Add --sparc2.2 option to examine to shift superblock around
220
and --sparc2.2update to rewrite the superblock
221
- Fix assorted typos in online help
223
Changes Prior to 1.0.0 release
224
- Allow --config with Misc mode (for --examine --scan)
225
- Add $(CXFLAGS) to end of CFLAGS in makefile
226
- When making an N disk raid5 array, the Nth drive
227
is moved to the end of the array as a spare rather than
228
being shifted up one place. This means that when the
229
kernel builds onto the last spare and inserts it,
230
the devices will be in the expected order.
231
- Man page improvements
232
Changes Prior to 0.8.2 release
233
- Correct spelling of persist[ae]nce/persist[ae]nt.
234
- Change "disk" to "device" in options and config file
235
- convert array size to "long long" *before* shift-left in -D and -Q
237
Changes Prior to 0.8.1 release
238
- Add "INSTALL" file.
239
- Fix some "i" variables that were not being set properly
240
- Initialise minsize and maxsize so that compilers don't complain.
241
- Tidy up Makefile and mdadm.spec installations
242
- Add "multipath" to documentation of valid levels
244
Changes Prior to 0.8 release
245
- Fix another bug in Assemble.c due to confusing 'i' with 'j'
246
- Minimal, untested, support for multipath
247
- re-write of argument parsing to have more coherent modes,
248
- add --query,-Q option
249
- Update mdadm.8 to reflect arg processing change and --query
250
- Change "long" to "unsigned long" for device sizes
251
- Handle "mailaddr" and "program" lines in config file for follow/scan mode.
252
- --follow --scan will exit if no program or mail found
253
- Add MAILADDR and PROGRAM to mdadm.conf-example
254
- Spell check man pages
255
- consistently use "component devices" instead of "subdevices"
256
- Make -Wall -Werror really work and fix lots of errors.
257
- --detail and --stop can have --scan which chooses devices from /proc/mdstat
258
- --monitor detects 20% changes in resync, failed spares,
260
- --monitor --scan will automatically add any devices found in /proc/mdstat
261
- --monitor will move spares between arrays with same spare-group if necessary
262
- Documentation for Monitor Mode
263
- --query notes if the array containing the given device is active or not
264
- Finished md.4 man page.
266
Changes Prior to 0.7.2 release
267
- mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra -- bluca@comedia.it
268
- more mdadm.spec updates from Gregory Leblanc <gleblanc@linuxweasel.com>
269
- make directory for mdadm.conf configurable in Makefile
270
- Finished mdadm.conf.5. Removed details of conf file from
271
mdadm.8 leaving a reference to mdadm.conf.5.
272
- Fix bug in Assemble.c, thanks to Junaid Rizvi <domdev@sat.net.pk>
273
- Get --assemble --force to make sure old major/minor numbers are
274
consistant, as md.c worries about this :-(
277
Changes Prior to 0.7.1 release
279
- use BLKGETSIZE64 if available for array size
280
- give human readable as GiB/MiB and GB and MB, with 2 decimal point precision
281
- Only warn about size variation for raid1/4/5.
282
- Started md.4 man page
283
- Started mdadm.conf.5 man page
285
Changes Prior to 0.7 release
287
- Fix makefile to install binary at /sbin and not /sbin/sbin
288
Also install man page.
289
- Add --zero-superblock based on --destroywithextremeprejudice
290
from Dale Stephenson <steph@snapserver.com>
291
- change name to mdadm. It is palandromic, and much nicer to pronouce.
293
Changes Prior to 0.6 release
295
- Remove the limit on the number of device names that can be
296
given on the command line.
297
- Fix bug in --assemble --force where it would only update a
299
- Fix bogus printing of big numbers not being block devices
300
when given names of devices that don't exist.
301
- When --assemble --force, consider superblocks with an event
302
count that is 1 behind as out-of-date. Normally they are
303
considered up-to-date (as the kernel assumes this too).
304
- When marking drives as not-failed in the superblock,
305
we also mark them as ACTIVE and SYNC.
306
- Don't start arrays for which not all drives are available unless:
307
--scan which implies that all drives were found automatically
308
--run which means the user knows what they want
309
--force which means that we are fixing something broken
310
- Make sure all device numbers passed as 3rd arg of ioctl
311
are passed as unsigned lock, so that it works on SPARC
312
- If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK
313
if we cannot read from the array, i.e. if the array is
316
- Taught Examine to handle --scan. It examines all devices listed
317
on DEVICE lines in the config file.
318
- Added --brief (-b) flag for Examine and Detail to print out
319
and mdctl.conf compatible description with uuid=, level=,
320
disks= and - for Examine - devices=
321
--examine --brief collects all devices the make the one array and
322
list them as one entry.
323
- Added level= and disks= options to ARRAY lines in config files
324
so --brief output could be used as-is.
325
- Make parity style ({left,right}-{,a}symmetric) consistantly use -,
327
- Add "Array Size" to --detail output
328
- Change "Size" to "Device Size" and exclude from Detail of arrays
329
that do not have a consistent device size.
330
- Add Human readable MiB or GiB value on size lines of Detail and Examine
331
- --assemble --scan doesn't complain about active drives
332
- require number of spares given in -x to be listed.
333
- Made --build actually work.
334
Changes Prior to 0.5 release
337
spare drives are handled properly.
339
--force can be used to recover from 2-drive failures on RAID5
340
If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should
341
make a raid5 array, but it has experienced multiple failures and
344
mdctl --assemble --force /dev/md0 /dev/hd[abcd]1
346
Should update the superblock on the newest failed drive and
347
restart the array in degraded mode. You should then remove the
348
remaining failed drive and re-add it (if you are happy that it
351
Ofcourse whenever you have a 2-drive failure, you have a risk
352
of corruption in data that hasn't be changed for a long time. So
353
this doesn't give you your array back all nice and happy, but it
354
does allow you to recover data that might not be corrupt.
356
More flexibility in identifying a RAID array in the mdctl.conf
358
array /dev/md4 super-minor=4
360
assembles /dev/md4 from all devices found that have a raid
361
superblock that says the minor number of the array is 4.
362
If the blocks with the right minor number do not all have the
363
same UUID, an error is flags and no assembly happens.
365
array /dev/md3 devices=/dev/hd[abc]2
367
Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All
368
devices must exist and have raid superblock with the same uuid.
370
If two identity specifiers are used, only devices that match all
371
of them are considered, so
373
array /dev/md2 devices=/dev/hd?2 super-minor=2
375
will assemble /dev/md2 using all /dev/hd?2 devices which have a
376
raid superblock with minor number 2.
379
When listing devices for --create, the word "missing" can be
380
used to indicate that the respective slot does not have a
381
working drive currently. This is similar to the "failed-disk"
382
directive in mkraid/raidtab.
384
mdctl --create --level=5 -raid-disks=4 --spare-disks=2
385
/dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde
387
will create a raid5 array with the third slot empty, and two
390
By default, raid5 arrays are created with the last slot empty
391
and drive listed for the last slot added as a spare. If a
392
"missing" slot is given, or if --force is given, then --create
393
does exactly what you ask and doesn't try to be clever.
396
--follow / --monitor:
398
This is a new mode. I couldn't stop my self from picking a name
399
starting with F (as current modes start A,B,C,D,E) but I
400
relented and provided an alternate name that is somewhat more
403
In this mode, mdctl does not exit, but runs continuously and
404
periodically polls all the md devices to see if they have had
405
any interested state change.
406
The changes that it currently notices are:
407
Fail - an active disc fails
408
FailSpare - a spare, that was presumably being build, fails
409
ActiveSpare - a spare becomes active, presumably after a rebuild.
412
--mail mailaddress - send Email on any Fail* event
413
--program program - run the program on any event.
414
Args are: eventname mddevice subdevice(if-known)
415
--delay seconds - change from the default 60second pause
418
I plan to add functionality to this mode to allow sharing of
419
spare drives. If an array is marks "spare-group=fred", and it
420
has a failed drive and no spares, and if some other array is
421
also "spare-group=fred" and it has no failed drives, but does
422
have a spare drive that is big enough, the spare will be moved
425
I also have the idea of adding a --grow mode which will re-organise
426
the data on an N disk raid0/4/5 array to be on an N+M disk array.
427
I have no concrete plans for this though.
429
I got rid of the "v" in the archive file name, and include the
430
version number in the directory created by the archive.
432
There is now a man page and mdctl.spec (for rpm) thanks to
433
Danilo Godec <danci@agenda.si>.
435
Ofcourse, the man page is now out of date and despite being based on
436
the --help output, is not wholy correct. After I get --follow
437
working properly, I plan to revise the various documentation and/or
438
the code to make sure the two match.