]> git.ipfire.org Git - thirdparty/mdadm.git/blob - ChangeLog
Don't list device= in --examine --scan output.
[thirdparty/mdadm.git] / ChangeLog
1 Changes Prior to this release
2 Several of these are backported from the Debian package
3 - Don't use 'lstat' to check for blockdevices, use stat.
4 - Document --size=max option for --grow
5 - Document SparesMissing event and DeviceDisappeared/WrongLevel
6 - --stop --scan repeatly cycles until no more progress can be made
7 so that stack devices are stopped properly
8 - Minor type rationalisation for ident->uuid - now always 'int[]'
9 - Fix type in online help for --grow (was -F, now -G)
10 - Allow --auto command line option to set default auto=
11 value when running "--assemble --scan". Previously
12 --auto was ignored if --scan was given
13 - Fix a few type casts
14 - Fix parsing of /dev/md/N in is_standard
15 - Fix rounding errors in human_size()
16 - Fix silly example in mdadm.conf-examples
17 - When finding a /dev name for a device, prefer shorter names
18 - Suppress listing of devices= in --brief output of -D or -E,
19 unless -v is given (-vv gives the old behaviour with -Dsv).
20 This is because the device list can change and so is not a
21 stable aspect of the array
22
23 Changes Prior to 1.11.0 release
24 - Fix embarassing bug which causes --add to always fail.
25
26 Changes Prior to 1.10.0 release
27 - Fix bug with --config=partitions
28 - Open sub-devices with O_EXCL to detect if already in use
29 - Make sure superblock updates are flushed directly to disk.
30
31 Changes Prior to 2.0-deve-1 release
32 - Support for version-1 superblock. See --metadata option.
33 - Support for bitmap based intent logging.
34 - Minor fixes.
35
36 Changes Prior to 1.9.0 release
37 - Fix rpm build problem (stray %)
38 - Minor manpage updates
39 - Change "dirty" status to "active" as it was confusing people.
40 - --assemble --auto recognises 'standard' name and insists on using
41 the appropriate major/minor number for them.
42 - Remove underscore from partition names, so partitions of
43 "foo" are "foo1", "foo2" etc (unchanged) and partitions of
44 "f00" are "f00p1", "f00p2" etc rather than "f00_p1"...
45 - Use "major", "minor", "makedev" macros instead of
46 "MAJOR", "MINOR", "MKDEV" so that large device numbers work
47 on 2.6 (providing you have glibc 2.3.3 or later).
48 - Add some missing closes of open file descriptors.
49 - Reread /proc/partition for every array assembled when using
50 it to find devices, rather than only once.
51 - Make "mdadm -Ss" stop stacked devices properly, by reversing the
52 order in which arrays are stopped.
53 - Improve some error messages.
54 - Allow device name to appear before first option, so e.g.
55 mdadm /dev/md0 -A /dev/sd[ab]
56 works.
57 - Assume '-Q' if just a device is given, rather than being silent.
58 - Change "dirty" status to "active" as it was confusing people.
59
60 Changes Prior to 1.8.0 release
61 - Makefile cleanup from Luca Berra <bluca@comedia.it>
62 - --pid-file (-i) to set a pid file to use with --monitor --daemonise
63 - Fix typo in mdadm man page
64 - Fix coredump when "-s" used with no config file present.
65 - Support new "faulty" personality which can inject synthetic
66 faults. (Not in kernel.org yet at 1Nov2004)
67 - Support raid0/linear on devices > 2 Terabytes
68 - Make sure raid6 resyncs when created with one missing device
69
70 Changes Prior to 1.7.0 release
71 - Support "--grow --add" to add a device to a linear array, if the
72 kernel supports it. Not documented yet.
73 - Restore support for uclibc which was broken recently.
74 - Several improvements to the output of --detail, including
75 reporting "resyncing" or "recovering" in the state.
76 - Close filedescriptor at end of --detail (exit would have closed it
77 anyway, so this isn't abig deal).
78 - Report "Sync checkpoint" in --examine output if appropriate.
79 - Add --update=resync for --assemble mode to for a resync when the
80 array is assembled.
81 - Add support for "raid10", which is under development in 2.6.
82 Not documented yet.
83 - --monitor now reads spare-group and spares info from config file
84 even when names of arrays to scan are given on the command line
85
86 Changes Prior to 1.6.0 release
87 - Device name given in -Eb is determined by examining /dev rather
88 than assuming /dev/md%d
89 - Fix bug in --monitor where an array could be held open an so
90 could not be stopped without killing mdadm.
91 - Add --grow mode. Currently only --size and --raid-disks can be
92 changed. Both require kernel support which, at the time of
93 writing, is not in a release kernel yet.
94 - Don't print out "errors" or "no-errors" in -D and -E, as the bit
95 is never set or used.
96 - Use md event notification in 2.6.??? to make --monitor mode
97 respond instantly to events.
98 - Add --auto= option and auto= configfile entry to tell mdadm to
99 create device files as needed. This is particularly useful
100 with partitioned arrays where the major device number can change.
101 - When generating --brief listing, if the standard name doesn't
102 exist, search /dev for one rather than using a temp name.
103 - Allow --build to build raid1 and multipath arrays.
104 - Add "--assume-clean" for Create and Build, particularly for raid1
105 Note: this is dangerous. Only use it if you are certain.
106 - Fix bug so that Rebuild status monitoring works again.
107 - Add "degraded" and "recovering" options to the "Status:"
108 entry for --detail
109
110 Changes Prior to 1.5.0 release
111 - new commands "mdassemble" which is a stripped-down equivalent of
112 "mdadm -As", that can be compiled with dietlibc.
113 Thanks to Luca Berra <bluca@comedia.it>.
114 It can be using in an initramfs or initrd.
115 - Fix compiling error with BLKGETSIZE64 and some signed/unsigned
116 comparison warnings.
117 - Add Rebuild Status (% complete) to --detail output.
118 - Support "--monitor --test" which will generate a test alert
119 for each array once, to test notification paths.
120 - Generate RebuildFinished event when rebuild finishes.
121 - Support for raid6 as found in 2.6.2 - thanks to
122 H. Peter Anvin <hpa@zytor.com>
123 - Support partitioned md arrays with a different major number and
124 naming scheme (md_dX in /proc/mdstat, /dev/md/dXpY in /dev).
125
126 Changes Prior to 1.4.0 release
127 - Document fact that creating a raid5 array really creates a
128 degraded array with a spare.
129 - Add "spares=" tag to config file and generate it wit --detail and
130 --examine
131 - Add "SparesMissing" event when --monitor first sees an array and
132 it doesn't have the enough spare devices.
133 - Add --update=summaries for --assemble to update summary
134 information in superblock, and correct other inconsistancies in
135 the superblock.
136 - Add --test option to --detail to set a meaningful exit status.
137
138 Changes Prior to 1.3.0 release
139 - Make 'size' and unsigned long in Create to allow creation of
140 larger arrays.
141 - Explicitly flag spare devices as 'spare' in --detail and --examine
142 output. Previously they simply had no flags lists.
143 - Make MailCmd (for monitor) configurable in Makefile, and default
144 to "/usr/sbin/sendmail -t". Also split out the warning related
145 flags into CWFLAGS for easier build configurability.
146 - Minor bugfix in Manage code.
147 - --monitor now notices and reports degraded arrays at startup using
148 "DegradedArray" event, and also has a --oneshot option to only
149 report DegradedArrays, and then exit.
150 - Small man-page clarification w.r.t. raid levels and raid4 in
151 particular.
152 - Disallow creation of arrays with only one device as this is
153 probably a mistake. --force will override this check.
154 - Correct some misleading documentation in the "mdadm --create --help"
155 message.
156 - Ignore chunksize if raid1 or multipath.
157 - Explicit statement in man page that raid-disks cannot be changed
158 after array is created.
159 - Improve message when attempting to start an array with
160 insufficient devices. Instead of required the array to be full,
161 we only require it has as many active devices as last time.
162
163 Changes Prior to 1.2.0 release
164 - Fix bug where --daemonise required an argument.
165 - In --assemble --verbose, print appropriate message if device is
166 not in devices= list
167 - Updated mdadm.conf.5 to reflect fact that device= takes wildcards
168 - Typos: componenet -> component
169 - Reduce size of "--help" message put excess into "--help-options"
170 - Fix bug introduced when MD_SB_DISKS dependancy removed, and which
171 caused spares not be assembled properly.
172 - Print appropriate message if --monitor --scan decides not to
173 monitor anything.
174 Changes Prior to 1.1.0 release
175 - add --deamonise flag for --monitor - forks and prints pid to stdout
176 - Fix bug so we REALLY clear dirty flag with -Af
177 - -Db now prints a 'devices=' word for each array.
178 - "mdadm -A /dev/md0" will get info from configfile, even without scan
179 - When assembling multipath arrays, ignore devices which are flagged
180 as having errors.
181 - take --super-minor=dev to mean "use the minor number of the mddev
182 being assembled.
183 - take --config=none to mean "completely ignore config file"
184 - Make --monitor require --scan or a device list.
185 Changes Prior to 1.0.9 release
186 - Documentation updates including kernel parameters documented
187 in md.4
188 - --assemble --force for raid4/5 will mark clean, needed for 2.5
189 - --detail prints out the events counter as well
190 - flush device before reading superblock to be sure to get
191 current data
192 - added mdadm.static target to makefile for static linking
193 - --monitor was ignoring /dev/md0 due to off-by-one error
194 - Fix assorted typos
195 - Fix printing of Gibibytes - calc was wrong.
196 - Fix printing of Array Size in --detail when very big.
197 - --monitor no longer tries to work for raid0 or linear as these
198 have nothing to be monitored.
199 - The word 'partitions' on a DEVICE line will cause all partitions
200 listed in /proc/partitions to be considered
201 - If the config file is called 'partitions' then it will be treated
202 as though it contained exactly 'device partitions' so e.g.
203 mdadm -Ebsc partitions
204 will find all raid partitions easily.
205 - successfully assemble multipath devices by ignoring raid_disk
206 value from superblock (it is always the same).
207 - --assemble not tied to MD_SB_DISKS limit quite so much
208 - Support compiling with tcc
209 - Support compiling with uclibc - just skip scan of /dev
210 - Add --update= option for Assemble mode. Either sparc2.2
211 or super-minor updates are possible. See mdadm.8
212
213 Changes Prior to 1.0.1 release
214 - Round off MB/GiB etc values instead of round down.
215 - Add --sparc2.2 option to examine to shift superblock around
216 and --sparc2.2update to rewrite the superblock
217 - Fix assorted typos in online help
218
219 Changes Prior to 1.0.0 release
220 - Allow --config with Misc mode (for --examine --scan)
221 - Add $(CXFLAGS) to end of CFLAGS in makefile
222 - When making an N disk raid5 array, the Nth drive
223 is moved to the end of the array as a spare rather than
224 being shifted up one place. This means that when the
225 kernel builds onto the last spare and inserts it,
226 the devices will be in the expected order.
227 - Man page improvements
228 Changes Prior to 0.8.2 release
229 - Correct spelling of persist[ae]nce/persist[ae]nt.
230 - Change "disk" to "device" in options and config file
231 - convert array size to "long long" *before* shift-left in -D and -Q
232
233 Changes Prior to 0.8.1 release
234 - Add "INSTALL" file.
235 - Fix some "i" variables that were not being set properly
236 - Initialise minsize and maxsize so that compilers don't complain.
237 - Tidy up Makefile and mdadm.spec installations
238 - Add "multipath" to documentation of valid levels
239
240 Changes Prior to 0.8 release
241 - Fix another bug in Assemble.c due to confusing 'i' with 'j'
242 - Minimal, untested, support for multipath
243 - re-write of argument parsing to have more coherent modes,
244 - add --query,-Q option
245 - Update mdadm.8 to reflect arg processing change and --query
246 - Change "long" to "unsigned long" for device sizes
247 - Handle "mailaddr" and "program" lines in config file for follow/scan mode.
248 - --follow --scan will exit if no program or mail found
249 - Add MAILADDR and PROGRAM to mdadm.conf-example
250 - Spell check man pages
251 - consistently use "component devices" instead of "subdevices"
252 - Make -Wall -Werror really work and fix lots of errors.
253 - --detail and --stop can have --scan which chooses devices from /proc/mdstat
254 - --monitor detects 20% changes in resync, failed spares,
255 disappearing arrays,
256 - --monitor --scan will automatically add any devices found in /proc/mdstat
257 - --monitor will move spares between arrays with same spare-group if necessary
258 - Documentation for Monitor Mode
259 - --query notes if the array containing the given device is active or not
260 - Finished md.4 man page.
261
262 Changes Prior to 0.7.2 release
263 - mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra -- bluca@comedia.it
264 - more mdadm.spec updates from Gregory Leblanc <gleblanc@linuxweasel.com>
265 - make directory for mdadm.conf configurable in Makefile
266 - Finished mdadm.conf.5. Removed details of conf file from
267 mdadm.8 leaving a reference to mdadm.conf.5.
268 - Fix bug in Assemble.c, thanks to Junaid Rizvi <domdev@sat.net.pk>
269 - Get --assemble --force to make sure old major/minor numbers are
270 consistant, as md.c worries about this :-(
271
272
273 Changes Prior to 0.7.1 release
274 - update mdadm.spec
275 - use BLKGETSIZE64 if available for array size
276 - give human readable as GiB/MiB and GB and MB, with 2 decimal point precision
277 - Only warn about size variation for raid1/4/5.
278 - Started md.4 man page
279 - Started mdadm.conf.5 man page
280
281 Changes Prior to 0.7 release
282
283 - Fix makefile to install binary at /sbin and not /sbin/sbin
284 Also install man page.
285 - Add --zero-superblock based on --destroywithextremeprejudice
286 from Dale Stephenson <steph@snapserver.com>
287 - change name to mdadm. It is palandromic, and much nicer to pronouce.
288
289 Changes Prior to 0.6 release
290
291 - Remove the limit on the number of device names that can be
292 given on the command line.
293 - Fix bug in --assemble --force where it would only update a
294 single superblock.
295 - Fix bogus printing of big numbers not being block devices
296 when given names of devices that don't exist.
297 - When --assemble --force, consider superblocks with an event
298 count that is 1 behind as out-of-date. Normally they are
299 considered up-to-date (as the kernel assumes this too).
300 - When marking drives as not-failed in the superblock,
301 we also mark them as ACTIVE and SYNC.
302 - Don't start arrays for which not all drives are available unless:
303 --scan which implies that all drives were found automatically
304 --run which means the user knows what they want
305 --force which means that we are fixing something broken
306 - Make sure all device numbers passed as 3rd arg of ioctl
307 are passed as unsigned lock, so that it works on SPARC
308 - If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK
309 if we cannot read from the array, i.e. if the array is
310 not started yet.
311 - man page update
312 - Taught Examine to handle --scan. It examines all devices listed
313 on DEVICE lines in the config file.
314 - Added --brief (-b) flag for Examine and Detail to print out
315 and mdctl.conf compatible description with uuid=, level=,
316 disks= and - for Examine - devices=
317 --examine --brief collects all devices the make the one array and
318 list them as one entry.
319 - Added level= and disks= options to ARRAY lines in config files
320 so --brief output could be used as-is.
321 - Make parity style ({left,right}-{,a}symmetric) consistantly use -,
322 never _.
323 - Add "Array Size" to --detail output
324 - Change "Size" to "Device Size" and exclude from Detail of arrays
325 that do not have a consistent device size.
326 - Add Human readable MiB or GiB value on size lines of Detail and Examine
327 - --assemble --scan doesn't complain about active drives
328 - require number of spares given in -x to be listed.
329 - Made --build actually work.
330 Changes Prior to 0.5 release
331
332 --assemble:
333 spare drives are handled properly.
334
335 --force can be used to recover from 2-drive failures on RAID5
336 If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should
337 make a raid5 array, but it has experienced multiple failures and
338 wont start, then
339
340 mdctl --assemble --force /dev/md0 /dev/hd[abcd]1
341
342 Should update the superblock on the newest failed drive and
343 restart the array in degraded mode. You should then remove the
344 remaining failed drive and re-add it (if you are happy that it
345 might work).
346
347 Ofcourse whenever you have a 2-drive failure, you have a risk
348 of corruption in data that hasn't be changed for a long time. So
349 this doesn't give you your array back all nice and happy, but it
350 does allow you to recover data that might not be corrupt.
351
352 More flexibility in identifying a RAID array in the mdctl.conf
353 e.g.
354 array /dev/md4 super-minor=4
355
356 assembles /dev/md4 from all devices found that have a raid
357 superblock that says the minor number of the array is 4.
358 If the blocks with the right minor number do not all have the
359 same UUID, an error is flags and no assembly happens.
360
361 array /dev/md3 devices=/dev/hd[abc]2
362
363 Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All
364 devices must exist and have raid superblock with the same uuid.
365
366 If two identity specifiers are used, only devices that match all
367 of them are considered, so
368
369 array /dev/md2 devices=/dev/hd?2 super-minor=2
370
371 will assemble /dev/md2 using all /dev/hd?2 devices which have a
372 raid superblock with minor number 2.
373
374 --create:
375 When listing devices for --create, the word "missing" can be
376 used to indicate that the respective slot does not have a
377 working drive currently. This is similar to the "failed-disk"
378 directive in mkraid/raidtab.
379 e.g.
380 mdctl --create --level=5 -raid-disks=4 --spare-disks=2
381 /dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde
382
383 will create a raid5 array with the third slot empty, and two
384 spares.
385
386 By default, raid5 arrays are created with the last slot empty
387 and drive listed for the last slot added as a spare. If a
388 "missing" slot is given, or if --force is given, then --create
389 does exactly what you ask and doesn't try to be clever.
390
391
392 --follow / --monitor:
393
394 This is a new mode. I couldn't stop my self from picking a name
395 starting with F (as current modes start A,B,C,D,E) but I
396 relented and provided an alternate name that is somewhat more
397 meaningful.
398
399 In this mode, mdctl does not exit, but runs continuously and
400 periodically polls all the md devices to see if they have had
401 any interested state change.
402 The changes that it currently notices are:
403 Fail - an active disc fails
404 FailSpare - a spare, that was presumably being build, fails
405 ActiveSpare - a spare becomes active, presumably after a rebuild.
406
407 Options:
408 --mail mailaddress - send Email on any Fail* event
409 --program program - run the program on any event.
410 Args are: eventname mddevice subdevice(if-known)
411 --delay seconds - change from the default 60second pause
412 between polls.
413
414 I plan to add functionality to this mode to allow sharing of
415 spare drives. If an array is marks "spare-group=fred", and it
416 has a failed drive and no spares, and if some other array is
417 also "spare-group=fred" and it has no failed drives, but does
418 have a spare drive that is big enough, the spare will be moved
419 to the first array.
420
421 I also have the idea of adding a --grow mode which will re-organise
422 the data on an N disk raid0/4/5 array to be on an N+M disk array.
423 I have no concrete plans for this though.
424
425 I got rid of the "v" in the archive file name, and include the
426 version number in the directory created by the archive.
427
428 There is now a man page and mdctl.spec (for rpm) thanks to
429 Danilo Godec <danci@agenda.si>.
430
431 Ofcourse, the man page is now out of date and despite being based on
432 the --help output, is not wholy correct. After I get --follow
433 working properly, I plan to revise the various documentation and/or
434 the code to make sure the two match.