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