1DMRAID(8) System Manager's Manual DMRAID(8)
2
6 dmraid - discover, configure and activate software (ATA)RAID
7
9 dmraid
10 {-a|--activate} {y|n|yes|no}
11 [-d|--debug]... [-v|--verbose]... [-i|--ignorelocking]
12 [-f|--format FORMAT[,FORMAT...]]
13 [{-P|--partchar} CHAR]
14 [-p|--no_partitions]
15 [-Z|--rm_partitions]
16 [--separator SEPARATOR]
17 [-t|--test]
18 [-u|--update_defer]
19 [RAID-set...]
20 dmraid
22 {-b|--block_devices}
23 [-c|--display_columns][FIELD[,FIELD...]]...
24 [-d|--debug]... [-v|--verbose]...
25 [--separator SEPARATOR]
26 [device-path...]
27 dmraid
29 {-h|--help}
30 dmraid
32 {-l|--list_formats}
33 [-d|--debug]... [-v|--verbose]...
34 dmraid
36 {-n|--native_log}
37 [-d|--debug]... [-v|--verbose]... [-i|--ignorelocking]
38 [-f|--format FORMAT[,FORMAT...]]
39 [--separator SEPARATOR]
40 [device-path...]
41 dmraid
43 {-R| --rebuild}
44 RAID-set
45 [device-path]
46 [-u|--update_defer]
47 dmraid
49 {-x| --remove}
50 [RAID-set]
51 dmraid
53 -f FORMAT-handler {-C| --create} set
54 --type raidlevel
55 [--size=setsize --strip stridesize]
56 --disk "device-path, device-path [, device-path ...]"
57 dmraid [ -f|--format FORMAT-handler] -S|--spare [RAID-set] -M|--media
59 "device-path"
60 dmraid
62 {-r|--raid_devices}
63 [-c|--display_columns][FIELD[,FIELD...]]...
64 [-d|--debug]... [-v|--verbose]... [-i|--ignorelocking]
65 [-D|--dump_metadata]
66 [-f|--format FORMAT[,FORMAT...]]
67 [--separator SEPARATOR]
68 [device-path...]
69 dmraid
71 {-r|--raid_devices}
72 [-d|--debug]... [-v|--verbose]... [-i|--ignorelocking]
73 [-E|--erase_metadata]
74 [-f|--format FORMAT[,FORMAT...]]
75 [--separator SEPARATOR]
76 [device-path...]
77 dmraid
79 {-s|--sets}...[a|i|active|inactive]
80 [-c|--display_columns][FIELD[,FIELD...]]...
81 [-d|--debug]... [-v|--verbose]... [-i|--ignorelocking]
82 [-f|--format FORMAT[,FORMAT...]]
83 [-g|--display_group]
84 [--separator SEPARATOR]
85 [RAID-set...]
86 dmraid
88 {-V/--version}
89
92 dmraid discovers block and software RAID devices (eg, ATARAID) by using
93 multiple different metadata format handlers which support various for‐
94 mats (eg, Highpoint 37x series). It offers activating RAID sets made
95 up by 2 or more discovered RAID devices, display properties of devices
96 and sets (see option -l for supported metadata formats). Block device
97 access to activated RAID sets occurs via device-mapper nodes /dev/map‐
98 per/RaidSetName. RaidSetName starts with the format name (see -l
99 option) which can be used to access all RAID sets of a specific format
100 easily with certain options (eg, -a below).
101 OPTIONS
104 -a, --activate {y|n} [RAID set...]
105 Activates or deactivates all or particular software RAID set.
106 In case metadata format handlers are chosen with -f , only RAID
107 sets with such format(s) can be activated or deactivated. Use‐
108 ful if devices have multiple metadata signatures. When activat‐
109 ing RAID sets, -p disables the activation of partitions on them,
110 and -Z will make dmraid tell the kernel to remove the partitions
111 from the disks underlying the set, ie if sda is part of the set,
112 remove sda1, sda2, etc. This prevents applications from
113 directly accessiong the disks bypassing dmraid. RAID set names
114 given on command line don't need to be fully specified (eg,
115 "dmraid -ay sil" would activate all discovered Silicon Image
116 Medley RAID sets). Option -u defers metadata update in case of
117 rebuild is triggered parallelly with activation. Awoids meta‐
118 data update to "OK" state if volume is not registered to the
119 event monitoring. Useful if volume is activating at early stage
120 of booting process when registration to the event monitoring is
121 impossible.
122 {-b|--block_devices} [device-path...]
126 List all or particular discovered block devices with their prop‐
127 erties (size, serial number). Add -c to display block device
128 names only and -cc for CSV column output of block device proper‐
129 ties. See description of -c below for FIELD identifiers.
130 [-d|--debug]...
133 Enable debugging output. Opion can be given multiple times
134 increasing the debug output level.
135 [-c|--display_columns][FIELD[,FIELD...]]...
138 Display properties of block devices, RAID sets and devices in
139 column(s). Optional list specifying which FIELDs to display.
140 For -b:
141 d[evpath]|p[ath], sec[tors]|si[ze], ser[ialnumber].
142 For -r:
143 de[vpath]|p[ath], f[ormat], r[aidname], t[ype], st[atus],
144 se[ctors]|si[ze], da[taoffset]|o[ffset].
145 For -s:
146 f[ormat], r[aidname], t[ype], sta[tus], str[ide],
147 se[ctors]|si[ze], su[bsets], d[evices], sp[ares].
148 [-f|--format FORMAT[,FORMAT...]]
150 Use metadata format handler(s) to discover RAID devices. See -l
151 for a list of supported format handler names. This is useful to
152 select particular formats in case multiple metadata signatures
153 are found on a device. A comma seperated list of format names
154 can be specified which may not contain white space.
155 {-h|--help}
158 Display help text.
159 {-i|--ignorelocking}
162 Don't take out any locks. Useful in early boot where no
163 read/write access to /var is available.
164 {-l|--list_formats}
167 List all available metadata format handlers with their names and
168 descriptions. Supported RAID levels are listed in parenthesis:
169 S: Span (concatination)
171 0: RAID0 (stripe)
172 1: RAID1 (mirror)
173 10: RAID10 (mirror on top of stripes)
174 01: RAID10 (stripe on top of mirrors) Note: Intel OROM displays
175 this as RAID10
176 {-n|--native_log} [device-path...]
179 Display metadata in native, vendor-specific format. In case a
180 metadata format handler is chosen with -f only RAID devices with
181 such format will be displayed in native format. If device-
182 path(s) is/are given on the command line, native metadata output
183 is restricted to those listed.
184 [{-P|--partchar} CHAR]
186 Use CHAR as the separator between the device name and the parti‐
187 tion number.
188 {-R| --rebuild} RAID-set [device-path]
190 Rebuild raid array after a drive has failed and a new drive is
191 added. For Intel chipset based systems, there are two methods
192 in which a new drive is added to the system. Option -u defers
193 metadata update in case of rebuild is triggered. Awoids meta‐
194 data update to "OK" state if volume is not registered to the
195 event monitoring.
196 1. Using OROM to identify a new drive
198 During system reboot, enter OROM and mark the new drive as
199 the rebuild drive.
200 After booting to the OS, use the dmraid command to rebuild.
201 Example: dmraid -R raid_set
203 2. Using dmraid to identify a new drive
205 Boot to the OS and use the dmraid command with the new drive
206 as the second parameter.
207 Example: dmraid -R raid_set /dev/sdc
209 3. Using hot spare drive
211 Mark a drive as hot spare using the "dmraid -f isw -S" com‐
212 mand. Then use the dmraid command to start the rebuild.
213 Example: dmraid -R raid_set
215 {-x|--remove} [RAID-set]
218 Delete one or all existing software RAID devices from the meta‐
219 data.
220 -f FORMAT-handler {-C|--create} --type raidlevel [--size=setsize
223 --strip stripsize] --disk device-path, device-path [,device-path]
224 Delete one or all existing Configure a software RAID device and
225 store the configuration data in a group of hard drive devices
226 consisting of this array. This command requires the following
227 options:
228 -f FORMAT-handler
230 metadata format (see "dmraid -l")
231 --type digit[digit...]
232 specify the raid level of the software RAID set.
233 0: raid0
234 1: raid1
235 5: raid5
236 01: raid01 (isw raid10)
237 --size: [digits[k|K|m|M|g|G][b|B]]
238 specify the size of the RAID set.The number is an integer
239 followed by [kKmMgG] and/or [bB].
240 b: byte (default)
241 B: block (512 bytes)
242 K or K: on the base of 1024
243 m or M: on the base of 1024*1024
244 g or G: on the base of 1024*1024*1024
245 If this option is missing, it's set to the default value pre-
246 configured by the vendor. Note that different vendors may apply
247 different constraints on the granularity of the size or the min‐
248 imal value.
249 --strip: [digits[k|K|m|M|g|G][b|B]]
250 specify the strip size of a RAID1, RAID5, and RAID10 RAID
251 set (as above)
252 --disk: device-path[{,| }device-path...]
253 specify the array of the hard drives, e.g. /dev/sda.
254 -f FORMAT-handler -S -M device-path
257 -S -M device-path
258 This command adds hot spare support for one or more RAID sets.
260 1. When used with a format handler, which supports hot spare
262 sets (e.g. isw), a hot spare is marked to be used when rebuild‐
263 ing any RAID set of that format. 2. When used when specifying a
264 RAID set, the drive is added to that RAID set and will be used
265 only to rebuild that set. Note: If the specified name does not
266 match an existing RAID-set, a set with the new name will be cre‐
267 ated.
268 {-r|--raid_devices} [device-path...]
271 List all discovered RAID devices with format, RAID level, sec‐
272 tors used and data offset into the device. In case a metadata
273 format handler is chosen with -f , only RAID devices with such
274 format can be discovered. Useful if devices have multiple meta‐
275 data signatures. If -D is added to -r the RAID metadata gets
276 dumped into a subdirectory named dmraid.format_name (eg. for‐
277 mat_name = isw) in files named devicename.dat. The byte offset
278 where the metadata is located on the device is written into
279 files named devicename.offset and the size of the device in sec‐
280 tors into files named devicename.size.
281 If -E is added to -r the RAID metadata on the devices gets con‐
283 ditionally erased. Useful to erase old metadata after new one
284 of different type has been stored on a device in order to avoid
285 discovering both. If you enter -E option -D will be enforced in
286 order to have a fallback in case the wrong metadata got erased.
287 Manual copying back onto the device is needed to recover from
288 erasing the wrong metadata using the dumped files device‐
289 name_formatname.dat and devicename_formatname.offset. Eg, to
290 restore all *.dat files in the working directory to the respec‐
291 tive devices:
292 for f in *.dat
294 do
295 dd if=$f of=/dev/${f%%.dat} \
296 seek=`cat ${f%%dat}offset` bs=1
297 done
298 If device-path(s) is/are given on the command line, the above
300 actions are restricted to those listed. Add -c to display RAID
301 device names only and -cc for CSV column output of RAID device
302 properties. See description of -c above for FIELD identifiers.
303 --separator SEPARATOR
306 Use SEPARATOR as a delimiter for all options taking or display‐
307 ing lists.
308 -s... [a|i] [RAID-set...]
311 Display properties of RAID sets. Multiple RAID set names can be
312 given on the command line which don't need to be fully specified
313 (eg, "dmraid -s hpt" would display all discovered Highpoint RAID
314 sets). Enter -s twice to display RAID subsets too. Add -c to
315 display names of RAID sets only, -cc for CSV column output of
316 RAID set properties and -ccc for inclusion of block devices in
317 the listing. Doesn't imply -s -s to show RAID subsets (implied
318 for group sets, e.g. isw). Add -g to include information about
319 group RAID sets (as with Intel Software RAID) in the listing.
320 See description of -c above for FIELD identifiers. Note: Size
321 is given in sectors (not bytes).
322 [-v|--verbose]...
325 Enable verbose runtime information output. Opion can be given
326 multiple times increasing the verbosity level.
327
330 "dmraid -l" lists all supported metadata formats with their names along
331 with some descriptive information, eg:
332 hpt37x : (+) Highpoint HPT37X
333 hpt45x : (+) Highpoint HPT45X
334 isw : (+) Intel Software RAID
335 lsi : (0) LSI Logic MegaRAID
336 nvidia : (+) NVidia RAID
337 pdc : (+) Promise FastTrack
338 sil : (+) Silicon Image(tm) Medley(tm)
339 via : (+) VIA Software RAID
340 dos : (+) DOS partitions on SW RAIDs
341 (0): Discover, (+): Discover+Activate
342 "dmraid -ay" activates all software RAID sets discovered.
344 "dmraid -an" deactivates all active software RAID sets which are not
346 open (eg, mounted filesystem on them).
347 "dmraid -ay -f pdc" (pdc looked up from "dmraid -l") activates all
349 software RAID sets with Promise format discovered and ignores all other
350 supported formats.
351 "dmraid -r" discovers all software RAID devices supported on your sys‐
353 tem, eg:
354 /dev/dm-46: hpt45x, "hpt45x_chidjhaiaa-0", striped, ok, 320172928 sec‐
355 tors, data@ 0
356 /dev/dm-50: hpt45x, "hpt45x_chidjhaiaa-0", striped, ok, 320172928 sec‐
357 tors, data@ 0
358 /dev/dm-54: hpt45x, "hpt45x_chidjhaiaa-1", striped, ok, 320172928 sec‐
359 tors, data@ 0
360 /dev/dm-58: hpt45x, "hpt45x_chidjhaiaa-1", striped, ok, 320172928 sec‐
361 tors, data@ 0
362 "dmraid -s -s hpt45x_chidjhaiaa" displays properties of set
365 "hpt45x_chidjhaiaa", eg:
366 *** Superset
367 name : hpt45x_chidjhaiaa
368 size : 640345856
369 stride : 128
370 type : raid10
371 status : ok
372 subsets: 2
373 dev : 4
374 spare : 0
375 ---> Subset
376 name : hpt45x_chidjhaiaa-0
377 size : 640345856
378 stride : 128
379 type : stripe
380 status : ok
381 subsets: 0
382 dev : 2
383 spare : 0
384 ---> Subset
385 name : hpt45x_chidjhaiaa-1
386 size : 640345856
387 stride : 128
388 type : stripe
389 status : ok
390 subsets: 0
391 dev : 2
392 spare : 0
393 "dmraid -s -ccs hpt45" displays properties in column format of all sets
395 and subsets with hpt45* format, eg:
396 hpt45x_chidjhaiaa,640345856,128,raid10,ok,4,0
397 hpt45x_chidjhaiaa-a,640345856,128,stripe,ok,2,0
398 hpt45x_chidjhaiaa-b,640345856,128,stripe,ok,2,0
399 "dmraid -r --sep : -cpath:size" display paths and sizes in sectors for
401 RAID devices in column format using ':' as a delimiter, eg:
402 /dev/dm-8:320173055
403 /dev/dm-12:320173055
404 /dev/dm-22:320173055
405 /dev/dm-26:320173055
406 /dev/dm-30:586114703
407 /dev/dm-34:586114703
408 /dev/dm-38:586114703
409 /dev/dm-42:586114703
410 /dev/dm-46:156301487
411 /dev/dm-50:156301487
412 /dev/dm-54:390624896
413 /dev/dm-58:390624896
414 /dev/dm-62:390624896
415 /dev/dm-66:390624896
416 "dmraid -f isw -C Raid0 --type 0 --strip 8k --size 20g --disk "/dev/sdb
418 /dev/sdc"" creates an ISW volume with a name of "Raid0", 20Gig bytes in
419 total, and 8kilo bytes strip size on two disks.
420 "dmraid -f isw -C Test0 --type 0 --disk "/dev/sdd /dev/sde"" creates an
422 ISW volume with the default size and strip size.
423 "dmraid -f isw -C Test10 --type 01 --strip 128B --disk "/dev/sda
425 /dev/sdb /dev/sdc /dev/sdd" creates a stacked RAID device, RAID10 (isw
426 format), with a name of "Test10", 128 blocks (512bytes) strip size ,
427 and the default volume size on 4 disks.
428 "dmraid -f isw -S -M /dev/sde" marks the device /dev/sde as a hot spare
430 for rebuild
431 "dmraid -R isw_djaggchdde_RAID1 /dev/sde" starts rebuild of the RAID
433 volume on device /dev/sde
434
437 dmraid returns an exit code of 0 for success or 1 for error.
438
441 Heinz Mauelshagen <Mauelshagen@RedHat.com>
442Heinz Mauelshagen DMRAID TOOL DMRAID(8)