]>
Commit | Line | Data |
---|---|---|
1 | Changes Prior to this release | |
2 | - allow --write-behind to be set for --grow. | |
3 | - When adding new disk to an array, don't reserve so much bitmap | |
4 | space that the disk cannot store the required data. (Needed when | |
5 | 1.x array was created with older mdadm). | |
6 | ||
7 | Changes Prior to 2.6.2 release | |
8 | - --fail detached and --remove faulty can be used to fail and | |
9 | remove devices that are no longer physically present. | |
10 | - --export option for --detail or present information in a format | |
11 | that can be processed by udev. | |
12 | - fix internal bitmap allocation problems with v1.1, v1.2 metadata. | |
13 | - --help now goes to stdout so you can direct it to a pager. | |
14 | - Various manpage updates. | |
15 | - Make "--grow --add" for linear arrays really work. | |
16 | - --auto-detect to trigger in-kernel autodetect. | |
17 | - Make return code for "--detail --test" more reliable. Missing | |
18 | devices as well as failed devices cause an error. | |
19 | ||
20 | Changes Prior to 2.6.1 release | |
21 | - --monitor was producing some meaningless warnings due to a bug. | |
22 | - Fix some compiler warnings. | |
23 | - Fully support --grow for raid6. If a reshape crashed during the | |
24 | critical period, mdadm wouldn't restore the Q information | |
25 | properly. | |
26 | - Update documentation for --grow. | |
27 | ||
28 | Changes Prior to 2.6 release | |
29 | - Fixed UUID printing in "--detail --brief" for version1 metadata. | |
30 | - --update=resync did exactly the wrong thing for version1 metadata. | |
31 | It caused a resync to not happen, rather than to happen. | |
32 | - Allow --assemble --force to mark a raid6 clean when it has two | |
33 | missing devices (which is needed else if won't assemble. | |
34 | Without this fix it would only assemble if one or zero | |
35 | missing devices. | |
36 | - Support --update=devicesize for cases where the underlying device | |
37 | can change size. | |
38 | - Default to --auto=yes so the array devices with 'standard' names | |
39 | get created automatically, as this is almost always what is wanted. | |
40 | - Give useful message if raid4/5/6 cannot be started because it is | |
41 | not clean and is also degraded. | |
42 | - Increase raid456 stripe cache size if needed to --grow the array. | |
43 | The setting used unfortunately requires intimate knowledge of the | |
44 | kernel, and it not reset when the reshape finishes. | |
45 | - Change 'Device Size' to 'Used Dev Size' because it only shows how | |
46 | much of each device is actually used, not how big they are. | |
47 | - --wait or -W will wait for resync activity to finish on the given | |
48 | devices. | |
49 | - Fix some problems with --update=uuid and add a test. | |
50 | - If two drives in a raid5 disappear at the same time, then "-Af" | |
51 | will add them both in rather than just one and forcing the array | |
52 | to 'clean'. This is slightly safer in some cases. | |
53 | - Check device is large enough before hot-add: this improves quality | |
54 | of error message. | |
55 | - Don't hold md device open for so long in --monitor mode - map_dev | |
56 | can be slow and interferes with trying to stop the array. | |
57 | - Support --uuid= with --create to choose your own UUID. | |
58 | - New major more "--incremental" for incremental assemble of arrays, | |
59 | intended for use with udev. | |
60 | ||
61 | Changes Prior to 2.5.6 release | |
62 | - Fix bug which meant "bitmap=xxx" in mdadm.conf was not handled | |
63 | properly. | |
64 | - Documentation updates. | |
65 | - Fix bug that caused infinite loop when doing auto-assembly, | |
66 | in certain cases where arrays couldn't be assembled. | |
67 | ||
68 | Changes Prior to 2.5.5 release | |
69 | - Don't #include linux/blkpg.h as that isn't safe. Just | |
70 | include the content literally. | |
71 | - Reduce maximum bitmap usage when working with bitmap files, | |
72 | so that a only single-page allocations are made, even on | |
73 | 64bit hosts with 4K pages. | |
74 | - Close stray fd in mdassemble so that it can assemble stacked | |
75 | devices | |
76 | - If mdassemble finds an array already assembled, it marks it | |
77 | read-write. | |
78 | - Remove error in md_open if array is already active. This isn't | |
79 | needed and gets in the ways if an array was created e.g. in | |
80 | initramfs, but device doesn't yet exist in /dev. | |
81 | - When --assemble --scan is run, if all arrays that could be found | |
82 | have already been started, don't report an error. | |
83 | - Fix a couple of bugs related to raid10 and the new 'offset' layout. | |
84 | - Improve error message when a wrong '--update' option is given. | |
85 | ||
86 | Changes Prior to 2.5.4 release | |
87 | - When creating devices in /dev/md/ create matching symlinks | |
88 | from /dev. e.g. /dev/md0 -> /dev/md/0. | |
89 | Allow this to be disabled in mdadm.conf or on command line. | |
90 | - Fix some endian-ness issues with version-1 superblocks (affects | |
91 | bigendian only). | |
92 | - Fix endian problem with 'bitmap' metadata | |
93 | - Allow a number (of partitions) after the 'yes' option to --auto= | |
94 | This is particularly useful in the 'create' line in mdadm.conf. | |
95 | - Remove partitions from any whole device that is made part of | |
96 | an md array. This is a work-around for annoying messages | |
97 | when the first block on some drive accidentally looks like a | |
98 | partition table. | |
99 | ||
100 | Changes Prior to 2.5.3 release | |
101 | - Document v0.91 superblocks in md.4 | |
102 | - Make GPL explicit in man pages. | |
103 | - Fix recent breakage of starting degraded arrays. | |
104 | - Tidyup automatic name choice for v-1 arrays: | |
105 | /dev/md_d0 now becomes '0', not '_d0'. | |
106 | ||
107 | Changes Prior to 2.5.2 release | |
108 | - Fix problem with compiling with gcc-2 compilers | |
109 | - Fix compile problem of post-incrmenting a variable in a macro arg. | |
110 | - Stop map_dev from returning [0:0], as that breaks things. | |
111 | - Add 'Array Slot' line to --examine for version-1 superblocks | |
112 | to make it a bit easier to see what is happening. | |
113 | - Work around bug in --add handling for version-1 superblocks | |
114 | in 2.6.17 (and prior). | |
115 | - Make -assemble a bit more resilient to finding strange | |
116 | information in superblocks. | |
117 | - Don't claim newly added spares are InSync!! (don't know why that | |
118 | code was ever in there) | |
119 | - Work better when no 'ftw' is available, and check to see | |
120 | if current uclibc provides ftw. | |
121 | - Never use /etc/mdadm.conf if --config file is given (previously | |
122 | some code used one, some used the other). | |
123 | ||
124 | Changes Prior to 2.5.1 release | |
125 | - Various fixes for gcc warnings | |
126 | - uclibc warnings | |
127 | - Makefile improvements for static linking/intalling | |
128 | - Makefile uninstall target | |
129 | - Really fix return status of --examine | |
130 | - Typos | |
131 | - Byteorder stuff (again) | |
132 | - Don't try to create devices with --manage or --grow | |
133 | - allow default metadata (superblock) type to be specified | |
134 | in mdadm.conf | |
135 | - Get --stop to list devices stopped but honour --quiet | |
136 | - remove libssl dependency | |
137 | - Avoid some misdetection of overlapping partitions | |
138 | - Fix memory leak in --monitor mode | |
139 | ||
140 | Changes Prior to 2.5 release | |
141 | - Support 'mailfrom' line in mdadm.conf so the From: line in alert | |
142 | emails can be explicitly set. | |
143 | - Arrange that SparesMissing (which is similar in import to | |
144 | DegradedArray) generates an Email. | |
145 | - Assume "DEVICE partitions" if no DEVICE line is given. | |
146 | - Support new 'offset' layout for raid10. | |
147 | - When creating a bitmap file, choose a chunksize to limit number | |
148 | of bitmap chunks to 2 million. More than this can cause kmalloc | |
149 | failure. | |
150 | - New 'CREATE' line in mdadm.conf for defaults such as owner, group, | |
151 | mode and auto-flag | |
152 | - --detail checks if array has been started or not and includes that | |
153 | in report. | |
154 | - When using --update=uuid on an array with a bitmap, update the | |
155 | bitmap's uuid too. | |
156 | - Add a copy of /proc/mdstat to the mail message sent by mdadm | |
157 | --monitor. | |
158 | - New flag --no-degraded to avoid starting arrays if there are | |
159 | fewer devices available than last time the array was started. | |
160 | This is only needed with --scan, as with --scan, that behaviour | |
161 | is the default. | |
162 | - Support for 'homehost' concept. This is a fairly major update. | |
163 | It includes a configfile option and a command line option for | |
164 | specifying a homehost, records that host in the superblock, | |
165 | and reports the homehost where possible. | |
166 | - Support for Auto Assembly. "mdadm -As" will, if provided with | |
167 | the name of a homehost, try to assemble all arrays it can find | |
168 | that were created for that homehost. See man pages for more details. | |
169 | ||
170 | Changes Prior to 2.4.1 release | |
171 | - Honour --write-mostly when adding to an array without persistent | |
172 | superblocks. | |
173 | - Fix alignment problem in version-1 superblocks. | |
174 | NOTE: This is an incompatable change affecting raid5 reshape. | |
175 | If you want to reshape a raid5 using version-1 superblocks, | |
176 | use 2.6.17-rc2 or later, and mdadm-2.4.1 or later. | |
177 | ||
178 | Changes Prior to 2.4 release | |
179 | - Rewrite 'reshape' support including performing a backup | |
180 | of the critical region for a raid5 growth, and restoring that | |
181 | backup after a crash. | |
182 | - Put a 'canary' at each end of the backup so a corruption | |
183 | can be more easily detected. | |
184 | - Remove useless 'ident' arguement from ->getinfo_super method. | |
185 | - Support --backup-file for backing-up critical section during | |
186 | growth. | |
187 | - Erase old superblocks (of different versions) when creating new | |
188 | array. | |
189 | - Allow --monitor to work with arrays with >28 devices | |
190 | - Report reshape information in --detail | |
191 | - Handle symlinks in /dev better | |
192 | - Fix mess in --detail output which a device is missing. | |
193 | - Manpage tidyup | |
194 | - Support 'bitmap=' in mdadm.conf for auto-assembling arrays with | |
195 | write-intent bitmaps in separate files. | |
196 | - Updates to md.4 man page including section on RESTRIPING and SYSFS | |
197 | ||
198 | Changes Prior to 2.3.1 release | |
199 | - Fixed -O2 compile so I could make and RPM. | |
200 | - Type cast number to be printed %llu so it compiles on 64bit | |
201 | machines. (Thanks Luca). | |
202 | - Stop using asm/byteorder.h - to make Redhat happy :-( | |
203 | - Require bitmap files to have a '/' in their name. | |
204 | - Error-check a few syscalls - code from SuSE package. | |
205 | ||
206 | Changes Prior to 2.3 release | |
207 | - Try /etc/mdadm/mdadm.conf if /etc/mdadm.conf doesn't exist. | |
208 | This provided compatability for Debian. | |
209 | - Fixed for version-1 superblock: | |
210 | report chunksize for raid6 and raid10 | |
211 | make sure device size used is a multiple of chunksize | |
212 | - Fix "--assemble --scan" crash. | |
213 | - Fix completely failure to create array on ppc64 | |
214 | - Fix memcmp in place of memcpy | |
215 | - A few minor improvements to online help | |
216 | - Clean up usage of 'long long' for used-size of devices, so | |
217 | that it is possible to create a raid1 of 7TB devices! | |
218 | - Make internal bitmaps work on 7TB raid1 arrays. | |
219 | - Provide error message if --examine doesn't find any superblock. | |
220 | - Report 'reshape' status in --examine - this depends on kernel | |
221 | patches that are not yet finalised. | |
222 | - Report bitmap status in --detail and --examine | |
223 | - Default to v1 superblocks instead of v0.90 if the array | |
224 | is too big for 0.90 to handle. | |
225 | - Sort the output of "mdadm --detail --scan" so that it is | |
226 | in a suitable order for assembling arrays. i.e. components come | |
227 | before an array that they are part of. | |
228 | - Print size of large reiserfs array properly went warning of | |
229 | possible confilcts. | |
230 | ||
231 | Changes Prior to 2.2 release | |
232 | - Assorted bug fixes | |
233 | - Support write-intent-bitmaps on raid10 | |
234 | - Support little-endian (Rather than hostendian) bitmaps. | |
235 | - Return correct error code from 'mdadm -S' | |
236 | - Remove extra blank line from 'mdadm -Eb' output. | |
237 | - Improve option parsing so that -a and -b do not have | |
238 | optional arguements: the arg is either required or not | |
239 | depending on context. | |
240 | - Allow scanning of devices listed in /proc/partitions even | |
241 | if they don't appear in /dev. | |
242 | - Support --assume-clean in --create mode as well as --build | |
243 | - Add support for --monitor to report to syslog: -y or --syslog. | |
244 | Thanks to Ross Vandegrift | |
245 | - --monitor now reports which device failed in a 'Fail' message | |
246 | This broke with 2.6 | |
247 | - Improve chance of array starting properly after a crash. | |
248 | mdadm was insisting the event numbers were identical, but this | |
249 | isn't needed, and is a problem if the crash was while the metadata | |
250 | was being updated. | |
251 | - Support --update==uuid | |
252 | - Added README.initramfs and mkinitramfs to help people use an | |
253 | initram for starting md arrays at boot. | |
254 | ||
255 | Changes Prior to 2.1 release | |
256 | - Fix assembling of raid10 array when devices are missing. | |
257 | mdadm now correctly detects if a array is workable or not | |
258 | depending on which devices are present, and so will correctly | |
259 | handle "--assemble --force" if multiple devices have failed. | |
260 | - Report raid10 layout in --examine output. | |
261 | - Fix assembling of arrays that use the version-1 superblock and | |
262 | have spares. Previously the spares would be ignored. | |
263 | - Fix bug so that multiple drives can be re-added at once. | |
264 | - Fix problem with hot-adding a bitmap to version-1-superblock | |
265 | arrays. | |
266 | ||
267 | Changes Prior to 2.0 | |
268 | - Support assembling from byte-swapped superblocks | |
269 | metadata type "0.swap" and --update=byteorder | |
270 | - write-mostly and write-behind support for raid1. | |
271 | - Support --name= and 'name=' config entry for identifying | |
272 | arrays be name. | |
273 | - RAID10 in man pages. | |
274 | - Lot of minor manpage updates | |
275 | ||
276 | Changes Prior to 2.0-devel-3 release | |
277 | - Assorted fixes for multiple bugs... | |
278 | - Add test suite | |
279 | ||
280 | Changes Prior to 1.12.0 release | |
281 | Several of these are backported from the Debian package | |
282 | - Don't use 'lstat' to check for blockdevices, use stat. | |
283 | - Document --size=max option for --grow | |
284 | - Document SparesMissing event and DeviceDisappeared/WrongLevel | |
285 | - --stop --scan repeatly cycles until no more progress can be made | |
286 | so that stack devices are stopped properly | |
287 | - Minor type rationalisation for ident->uuid - now always 'int[]' | |
288 | - Fix type in online help for --grow (was -F, now -G) | |
289 | - Allow --auto command line option to set default auto= | |
290 | value when running "--assemble --scan". Previously | |
291 | --auto was ignored if --scan was given | |
292 | - Fix a few type casts | |
293 | - Fix parsing of /dev/md/N in is_standard | |
294 | - Fix rounding errors in human_size() | |
295 | - Fix silly example in mdadm.conf-examples | |
296 | - When finding a /dev name for a device, prefer shorter names | |
297 | - Suppress listing of devices= in --brief output of -D or -E, | |
298 | unless -v is given (-vv gives the old behaviour with -Dsv). | |
299 | This is because the device list can change and so is not a | |
300 | stable aspect of the array | |
301 | - Allow --force with --grow so '-Gfn1' works (on raid1) | |
302 | - Replace sprintf calls with snprintf (to quiet diet-libc) | |
303 | - Add man page for mdassemble | |
304 | - Fix compilation with tinyc | |
305 | ||
306 | Changes Prior to 1.11.0 release | |
307 | - Fix embarassing bug which causes --add to always fail. | |
308 | ||
309 | Changes Prior to 1.10.0 release | |
310 | - Fix bug with --config=partitions | |
311 | - Open sub-devices with O_EXCL to detect if already in use | |
312 | - Make sure superblock updates are flushed directly to disk. | |
313 | ||
314 | Changes Prior to 2.0-devel-1 release | |
315 | - Support for version-1 superblock. See --metadata option. | |
316 | - Support for bitmap based intent logging. | |
317 | - Minor fixes. | |
318 | ||
319 | Changes Prior to 1.9.0 release | |
320 | - Fix rpm build problem (stray %) | |
321 | - Minor manpage updates | |
322 | - Change "dirty" status to "active" as it was confusing people. | |
323 | - --assemble --auto recognises 'standard' name and insists on using | |
324 | the appropriate major/minor number for them. | |
325 | - Remove underscore from partition names, so partitions of | |
326 | "foo" are "foo1", "foo2" etc (unchanged) and partitions of | |
327 | "f00" are "f00p1", "f00p2" etc rather than "f00_p1"... | |
328 | - Use "major", "minor", "makedev" macros instead of | |
329 | "MAJOR", "MINOR", "MKDEV" so that large device numbers work | |
330 | on 2.6 (providing you have glibc 2.3.3 or later). | |
331 | - Add some missing closes of open file descriptors. | |
332 | - Reread /proc/partition for every array assembled when using | |
333 | it to find devices, rather than only once. | |
334 | - Make "mdadm -Ss" stop stacked devices properly, by reversing the | |
335 | order in which arrays are stopped. | |
336 | - Improve some error messages. | |
337 | - Allow device name to appear before first option, so e.g. | |
338 | mdadm /dev/md0 -A /dev/sd[ab] | |
339 | works. | |
340 | - Assume '-Q' if just a device is given, rather than being silent. | |
341 | - Change "dirty" status to "active" as it was confusing people. | |
342 | ||
343 | Changes Prior to 1.8.0 release | |
344 | - Makefile cleanup from Luca Berra <bluca@comedia.it> | |
345 | - --pid-file (-i) to set a pid file to use with --monitor --daemonise | |
346 | - Fix typo in mdadm man page | |
347 | - Fix coredump when "-s" used with no config file present. | |
348 | - Support new "faulty" personality which can inject synthetic | |
349 | faults. (Not in kernel.org yet at 1Nov2004) | |
350 | - Support raid0/linear on devices > 2 Terabytes | |
351 | - Make sure raid6 resyncs when created with one missing device | |
352 | ||
353 | Changes Prior to 1.7.0 release | |
354 | - Support "--grow --add" to add a device to a linear array, if the | |
355 | kernel supports it. Not documented yet. | |
356 | - Restore support for uclibc which was broken recently. | |
357 | - Several improvements to the output of --detail, including | |
358 | reporting "resyncing" or "recovering" in the state. | |
359 | - Close filedescriptor at end of --detail (exit would have closed it | |
360 | anyway, so this isn't abig deal). | |
361 | - Report "Sync checkpoint" in --examine output if appropriate. | |
362 | - Add --update=resync for --assemble mode to for a resync when the | |
363 | array is assembled. | |
364 | - Add support for "raid10", which is under development in 2.6. | |
365 | Not documented yet. | |
366 | - --monitor now reads spare-group and spares info from config file | |
367 | even when names of arrays to scan are given on the command line | |
368 | ||
369 | Changes Prior to 1.6.0 release | |
370 | - Device name given in -Eb is determined by examining /dev rather | |
371 | than assuming /dev/md%d | |
372 | - Fix bug in --monitor where an array could be held open an so | |
373 | could not be stopped without killing mdadm. | |
374 | - Add --grow mode. Currently only --size and --raid-disks can be | |
375 | changed. Both require kernel support which, at the time of | |
376 | writing, is not in a release kernel yet. | |
377 | - Don't print out "errors" or "no-errors" in -D and -E, as the bit | |
378 | is never set or used. | |
379 | - Use md event notification in 2.6.??? to make --monitor mode | |
380 | respond instantly to events. | |
381 | - Add --auto= option and auto= configfile entry to tell mdadm to | |
382 | create device files as needed. This is particularly useful | |
383 | with partitioned arrays where the major device number can change. | |
384 | - When generating --brief listing, if the standard name doesn't | |
385 | exist, search /dev for one rather than using a temp name. | |
386 | - Allow --build to build raid1 and multipath arrays. | |
387 | - Add "--assume-clean" for Create and Build, particularly for raid1 | |
388 | Note: this is dangerous. Only use it if you are certain. | |
389 | - Fix bug so that Rebuild status monitoring works again. | |
390 | - Add "degraded" and "recovering" options to the "Status:" | |
391 | entry for --detail | |
392 | ||
393 | Changes Prior to 1.5.0 release | |
394 | - new commands "mdassemble" which is a stripped-down equivalent of | |
395 | "mdadm -As", that can be compiled with dietlibc. | |
396 | Thanks to Luca Berra <bluca@comedia.it>. | |
397 | It can be using in an initramfs or initrd. | |
398 | - Fix compiling error with BLKGETSIZE64 and some signed/unsigned | |
399 | comparison warnings. | |
400 | - Add Rebuild Status (% complete) to --detail output. | |
401 | - Support "--monitor --test" which will generate a test alert | |
402 | for each array once, to test notification paths. | |
403 | - Generate RebuildFinished event when rebuild finishes. | |
404 | - Support for raid6 as found in 2.6.2 - thanks to | |
405 | H. Peter Anvin <hpa@zytor.com> | |
406 | - Support partitioned md arrays with a different major number and | |
407 | naming scheme (md_dX in /proc/mdstat, /dev/md/dXpY in /dev). | |
408 | ||
409 | Changes Prior to 1.4.0 release | |
410 | - Document fact that creating a raid5 array really creates a | |
411 | degraded array with a spare. | |
412 | - Add "spares=" tag to config file and generate it wit --detail and | |
413 | --examine | |
414 | - Add "SparesMissing" event when --monitor first sees an array and | |
415 | it doesn't have the enough spare devices. | |
416 | - Add --update=summaries for --assemble to update summary | |
417 | information in superblock, and correct other inconsistancies in | |
418 | the superblock. | |
419 | - Add --test option to --detail to set a meaningful exit status. | |
420 | ||
421 | Changes Prior to 1.3.0 release | |
422 | - Make 'size' and unsigned long in Create to allow creation of | |
423 | larger arrays. | |
424 | - Explicitly flag spare devices as 'spare' in --detail and --examine | |
425 | output. Previously they simply had no flags lists. | |
426 | - Make MailCmd (for monitor) configurable in Makefile, and default | |
427 | to "/usr/sbin/sendmail -t". Also split out the warning related | |
428 | flags into CWFLAGS for easier build configurability. | |
429 | - Minor bugfix in Manage code. | |
430 | - --monitor now notices and reports degraded arrays at startup using | |
431 | "DegradedArray" event, and also has a --oneshot option to only | |
432 | report DegradedArrays, and then exit. | |
433 | - Small man-page clarification w.r.t. raid levels and raid4 in | |
434 | particular. | |
435 | - Disallow creation of arrays with only one device as this is | |
436 | probably a mistake. --force will override this check. | |
437 | - Correct some misleading documentation in the "mdadm --create --help" | |
438 | message. | |
439 | - Ignore chunksize if raid1 or multipath. | |
440 | - Explicit statement in man page that raid-disks cannot be changed | |
441 | after array is created. | |
442 | - Improve message when attempting to start an array with | |
443 | insufficient devices. Instead of required the array to be full, | |
444 | we only require it has as many active devices as last time. | |
445 | ||
446 | Changes Prior to 1.2.0 release | |
447 | - Fix bug where --daemonise required an argument. | |
448 | - In --assemble --verbose, print appropriate message if device is | |
449 | not in devices= list | |
450 | - Updated mdadm.conf.5 to reflect fact that device= takes wildcards | |
451 | - Typos: componenet -> component | |
452 | - Reduce size of "--help" message put excess into "--help-options" | |
453 | - Fix bug introduced when MD_SB_DISKS dependancy removed, and which | |
454 | caused spares not be assembled properly. | |
455 | - Print appropriate message if --monitor --scan decides not to | |
456 | monitor anything. | |
457 | Changes Prior to 1.1.0 release | |
458 | - add --deamonise flag for --monitor - forks and prints pid to stdout | |
459 | - Fix bug so we REALLY clear dirty flag with -Af | |
460 | - -Db now prints a 'devices=' word for each array. | |
461 | - "mdadm -A /dev/md0" will get info from configfile, even without scan | |
462 | - When assembling multipath arrays, ignore devices which are flagged | |
463 | as having errors. | |
464 | - take --super-minor=dev to mean "use the minor number of the mddev | |
465 | being assembled. | |
466 | - take --config=none to mean "completely ignore config file" | |
467 | - Make --monitor require --scan or a device list. | |
468 | Changes Prior to 1.0.9 release | |
469 | - Documentation updates including kernel parameters documented | |
470 | in md.4 | |
471 | - --assemble --force for raid4/5 will mark clean, needed for 2.5 | |
472 | - --detail prints out the events counter as well | |
473 | - flush device before reading superblock to be sure to get | |
474 | current data | |
475 | - added mdadm.static target to makefile for static linking | |
476 | - --monitor was ignoring /dev/md0 due to off-by-one error | |
477 | - Fix assorted typos | |
478 | - Fix printing of Gibibytes - calc was wrong. | |
479 | - Fix printing of Array Size in --detail when very big. | |
480 | - --monitor no longer tries to work for raid0 or linear as these | |
481 | have nothing to be monitored. | |
482 | - The word 'partitions' on a DEVICE line will cause all partitions | |
483 | listed in /proc/partitions to be considered | |
484 | - If the config file is called 'partitions' then it will be treated | |
485 | as though it contained exactly 'device partitions' so e.g. | |
486 | mdadm -Ebsc partitions | |
487 | will find all raid partitions easily. | |
488 | - successfully assemble multipath devices by ignoring raid_disk | |
489 | value from superblock (it is always the same). | |
490 | - --assemble not tied to MD_SB_DISKS limit quite so much | |
491 | - Support compiling with tcc | |
492 | - Support compiling with uclibc - just skip scan of /dev | |
493 | - Add --update= option for Assemble mode. Either sparc2.2 | |
494 | or super-minor updates are possible. See mdadm.8 | |
495 | ||
496 | Changes Prior to 1.0.1 release | |
497 | - Round off MB/GiB etc values instead of round down. | |
498 | - Add --sparc2.2 option to examine to shift superblock around | |
499 | and --sparc2.2update to rewrite the superblock | |
500 | - Fix assorted typos in online help | |
501 | ||
502 | Changes Prior to 1.0.0 release | |
503 | - Allow --config with Misc mode (for --examine --scan) | |
504 | - Add $(CXFLAGS) to end of CFLAGS in makefile | |
505 | - When making an N disk raid5 array, the Nth drive | |
506 | is moved to the end of the array as a spare rather than | |
507 | being shifted up one place. This means that when the | |
508 | kernel builds onto the last spare and inserts it, | |
509 | the devices will be in the expected order. | |
510 | - Man page improvements | |
511 | Changes Prior to 0.8.2 release | |
512 | - Correct spelling of persist[ae]nce/persist[ae]nt. | |
513 | - Change "disk" to "device" in options and config file | |
514 | - convert array size to "long long" *before* shift-left in -D and -Q | |
515 | ||
516 | Changes Prior to 0.8.1 release | |
517 | - Add "INSTALL" file. | |
518 | - Fix some "i" variables that were not being set properly | |
519 | - Initialise minsize and maxsize so that compilers don't complain. | |
520 | - Tidy up Makefile and mdadm.spec installations | |
521 | - Add "multipath" to documentation of valid levels | |
522 | ||
523 | Changes Prior to 0.8 release | |
524 | - Fix another bug in Assemble.c due to confusing 'i' with 'j' | |
525 | - Minimal, untested, support for multipath | |
526 | - re-write of argument parsing to have more coherent modes, | |
527 | - add --query,-Q option | |
528 | - Update mdadm.8 to reflect arg processing change and --query | |
529 | - Change "long" to "unsigned long" for device sizes | |
530 | - Handle "mailaddr" and "program" lines in config file for follow/scan mode. | |
531 | - --follow --scan will exit if no program or mail found | |
532 | - Add MAILADDR and PROGRAM to mdadm.conf-example | |
533 | - Spell check man pages | |
534 | - consistently use "component devices" instead of "subdevices" | |
535 | - Make -Wall -Werror really work and fix lots of errors. | |
536 | - --detail and --stop can have --scan which chooses devices from /proc/mdstat | |
537 | - --monitor detects 20% changes in resync, failed spares, | |
538 | disappearing arrays, | |
539 | - --monitor --scan will automatically add any devices found in /proc/mdstat | |
540 | - --monitor will move spares between arrays with same spare-group if necessary | |
541 | - Documentation for Monitor Mode | |
542 | - --query notes if the array containing the given device is active or not | |
543 | - Finished md.4 man page. | |
544 | ||
545 | Changes Prior to 0.7.2 release | |
546 | - mdadm.spec updates and ifdef BLKGETSIZE64 from Luca Berra -- bluca@comedia.it | |
547 | - more mdadm.spec updates from Gregory Leblanc <gleblanc@linuxweasel.com> | |
548 | - make directory for mdadm.conf configurable in Makefile | |
549 | - Finished mdadm.conf.5. Removed details of conf file from | |
550 | mdadm.8 leaving a reference to mdadm.conf.5. | |
551 | - Fix bug in Assemble.c, thanks to Junaid Rizvi <domdev@sat.net.pk> | |
552 | - Get --assemble --force to make sure old major/minor numbers are | |
553 | consistant, as md.c worries about this :-( | |
554 | ||
555 | ||
556 | Changes Prior to 0.7.1 release | |
557 | - update mdadm.spec | |
558 | - use BLKGETSIZE64 if available for array size | |
559 | - give human readable as GiB/MiB and GB and MB, with 2 decimal point precision | |
560 | - Only warn about size variation for raid1/4/5. | |
561 | - Started md.4 man page | |
562 | - Started mdadm.conf.5 man page | |
563 | ||
564 | Changes Prior to 0.7 release | |
565 | ||
566 | - Fix makefile to install binary at /sbin and not /sbin/sbin | |
567 | Also install man page. | |
568 | - Add --zero-superblock based on --destroywithextremeprejudice | |
569 | from Dale Stephenson <steph@snapserver.com> | |
570 | - change name to mdadm. It is palandromic, and much nicer to pronouce. | |
571 | ||
572 | Changes Prior to 0.6 release | |
573 | ||
574 | - Remove the limit on the number of device names that can be | |
575 | given on the command line. | |
576 | - Fix bug in --assemble --force where it would only update a | |
577 | single superblock. | |
578 | - Fix bogus printing of big numbers not being block devices | |
579 | when given names of devices that don't exist. | |
580 | - When --assemble --force, consider superblocks with an event | |
581 | count that is 1 behind as out-of-date. Normally they are | |
582 | considered up-to-date (as the kernel assumes this too). | |
583 | - When marking drives as not-failed in the superblock, | |
584 | we also mark them as ACTIVE and SYNC. | |
585 | - Don't start arrays for which not all drives are available unless: | |
586 | --scan which implies that all drives were found automatically | |
587 | --run which means the user knows what they want | |
588 | --force which means that we are fixing something broken | |
589 | - Make sure all device numbers passed as 3rd arg of ioctl | |
590 | are passed as unsigned lock, so that it works on SPARC | |
591 | - If HOT_ADD_DISK failes for -a, then only try ADD_NEW_DISK | |
592 | if we cannot read from the array, i.e. if the array is | |
593 | not started yet. | |
594 | - man page update | |
595 | - Taught Examine to handle --scan. It examines all devices listed | |
596 | on DEVICE lines in the config file. | |
597 | - Added --brief (-b) flag for Examine and Detail to print out | |
598 | and mdctl.conf compatible description with uuid=, level=, | |
599 | disks= and - for Examine - devices= | |
600 | --examine --brief collects all devices the make the one array and | |
601 | list them as one entry. | |
602 | - Added level= and disks= options to ARRAY lines in config files | |
603 | so --brief output could be used as-is. | |
604 | - Make parity style ({left,right}-{,a}symmetric) consistantly use -, | |
605 | never _. | |
606 | - Add "Array Size" to --detail output | |
607 | - Change "Size" to "Device Size" and exclude from Detail of arrays | |
608 | that do not have a consistent device size. | |
609 | - Add Human readable MiB or GiB value on size lines of Detail and Examine | |
610 | - --assemble --scan doesn't complain about active drives | |
611 | - require number of spares given in -x to be listed. | |
612 | - Made --build actually work. | |
613 | Changes Prior to 0.5 release | |
614 | ||
615 | --assemble: | |
616 | spare drives are handled properly. | |
617 | ||
618 | --force can be used to recover from 2-drive failures on RAID5 | |
619 | If you belive that /dev/hda1 /dev/hdb1 /dev/hdc1 /dev/hdd1 should | |
620 | make a raid5 array, but it has experienced multiple failures and | |
621 | wont start, then | |
622 | ||
623 | mdctl --assemble --force /dev/md0 /dev/hd[abcd]1 | |
624 | ||
625 | Should update the superblock on the newest failed drive and | |
626 | restart the array in degraded mode. You should then remove the | |
627 | remaining failed drive and re-add it (if you are happy that it | |
628 | might work). | |
629 | ||
630 | Ofcourse whenever you have a 2-drive failure, you have a risk | |
631 | of corruption in data that hasn't be changed for a long time. So | |
632 | this doesn't give you your array back all nice and happy, but it | |
633 | does allow you to recover data that might not be corrupt. | |
634 | ||
635 | More flexibility in identifying a RAID array in the mdctl.conf | |
636 | e.g. | |
637 | array /dev/md4 super-minor=4 | |
638 | ||
639 | assembles /dev/md4 from all devices found that have a raid | |
640 | superblock that says the minor number of the array is 4. | |
641 | If the blocks with the right minor number do not all have the | |
642 | same UUID, an error is flags and no assembly happens. | |
643 | ||
644 | array /dev/md3 devices=/dev/hd[abc]2 | |
645 | ||
646 | Assembles /dev/md3 drom /dev/hda2 /dev/hdb2 and/dev/hdc2. All | |
647 | devices must exist and have raid superblock with the same uuid. | |
648 | ||
649 | If two identity specifiers are used, only devices that match all | |
650 | of them are considered, so | |
651 | ||
652 | array /dev/md2 devices=/dev/hd?2 super-minor=2 | |
653 | ||
654 | will assemble /dev/md2 using all /dev/hd?2 devices which have a | |
655 | raid superblock with minor number 2. | |
656 | ||
657 | --create: | |
658 | When listing devices for --create, the word "missing" can be | |
659 | used to indicate that the respective slot does not have a | |
660 | working drive currently. This is similar to the "failed-disk" | |
661 | directive in mkraid/raidtab. | |
662 | e.g. | |
663 | mdctl --create --level=5 -raid-disks=4 --spare-disks=2 | |
664 | /dev/md0 /dev/sda /dev/sdb missing /dev/sdc /dev/sdd /dev/sde | |
665 | ||
666 | will create a raid5 array with the third slot empty, and two | |
667 | spares. | |
668 | ||
669 | By default, raid5 arrays are created with the last slot empty | |
670 | and drive listed for the last slot added as a spare. If a | |
671 | "missing" slot is given, or if --force is given, then --create | |
672 | does exactly what you ask and doesn't try to be clever. | |
673 | ||
674 | ||
675 | --follow / --monitor: | |
676 | ||
677 | This is a new mode. I couldn't stop my self from picking a name | |
678 | starting with F (as current modes start A,B,C,D,E) but I | |
679 | relented and provided an alternate name that is somewhat more | |
680 | meaningful. | |
681 | ||
682 | In this mode, mdctl does not exit, but runs continuously and | |
683 | periodically polls all the md devices to see if they have had | |
684 | any interested state change. | |
685 | The changes that it currently notices are: | |
686 | Fail - an active disc fails | |
687 | FailSpare - a spare, that was presumably being build, fails | |
688 | ActiveSpare - a spare becomes active, presumably after a rebuild. | |
689 | ||
690 | Options: | |
691 | --mail mailaddress - send Email on any Fail* event | |
692 | --program program - run the program on any event. | |
693 | Args are: eventname mddevice subdevice(if-known) | |
694 | --delay seconds - change from the default 60second pause | |
695 | between polls. | |
696 | ||
697 | I plan to add functionality to this mode to allow sharing of | |
698 | spare drives. If an array is marks "spare-group=fred", and it | |
699 | has a failed drive and no spares, and if some other array is | |
700 | also "spare-group=fred" and it has no failed drives, but does | |
701 | have a spare drive that is big enough, the spare will be moved | |
702 | to the first array. | |
703 | ||
704 | I also have the idea of adding a --grow mode which will re-organise | |
705 | the data on an N disk raid0/4/5 array to be on an N+M disk array. | |
706 | I have no concrete plans for this though. | |
707 | ||
708 | I got rid of the "v" in the archive file name, and include the | |
709 | version number in the directory created by the archive. | |
710 | ||
711 | There is now a man page and mdctl.spec (for rpm) thanks to | |
712 | Danilo Godec <danci@agenda.si>. | |
713 | ||
714 | Ofcourse, the man page is now out of date and despite being based on | |
715 | the --help output, is not wholy correct. After I get --follow | |
716 | working properly, I plan to revise the various documentation and/or | |
717 | the code to make sure the two match. |