1PVMOVE(8) System Manager's Manual PVMOVE(8)
2
3
4
6 pvmove - Move extents from one physical volume to another
7
9 pvmove position_args
10 [ option_args ]
11 [ position_args ]
12
14 pvmove moves the allocated physical extents (PEs) on a source PV to one
15 or more destination PVs. You can optionally specify a source LV in
16 which case only extents used by that LV will be moved to free (or spec‐
17 ified) extents on the destination PV. If no destination PV is speci‐
18 fied, the normal allocation rules for the VG are used.
19
20 If pvmove is interrupted for any reason (e.g. the machine crashes) then
21 run pvmove again without any PV arguments to restart any operations
22 that were in progress from the last checkpoint. Alternatively, use the
23 abort option at any time to abort the operation. The resulting location
24 of LVs after an abort depends on whether the atomic option was used.
25
26 More than one pvmove can run concurrently if they are moving data from
27 different source PVs, but additional pvmoves will ignore any LVs al‐
28 ready in the process of being changed, so some data might not get
29 moved.
30
32 Move PV extents.
33
34 pvmove PV
35 [ -A|--autobackup y|n ]
36 [ -n|--name LV ]
37 [ --alloc contiguous|cling|cling_by_tags|normal|anywhere|inherit
38 ]
39 [ --atomic ]
40 [ --noudevsync ]
41 [ --reportformat basic|json ]
42 [ COMMON_OPTIONS ]
43 [ PV ... ]
44
45 Continue or abort existing pvmove operations.
46
47 pvmove
48 [ COMMON_OPTIONS ]
49
50 Common options for command:
51 [ -b|--background ]
52 [ -i|--interval Number ]
53 [ --abort ]
54
55 Common options for lvm:
56 [ -d|--debug ]
57 [ -h|--help ]
58 [ -q|--quiet ]
59 [ -t|--test ]
60 [ -v|--verbose ]
61 [ -y|--yes ]
62 [ --commandprofile String ]
63 [ --config String ]
64 [ --driverloaded y|n ]
65 [ --lockopt String ]
66 [ --longhelp ]
67 [ --nolocking ]
68 [ --profile String ]
69 [ --version ]
70
72 --abort
73 Abort any pvmove operations in progress. If a pvmove was started
74 with the --atomic option, then all LVs will remain on the source
75 PV. Otherwise, segments that have been moved will remain on the
76 destination PV, while unmoved segments will remain on the source
77 PV.
78
79 --alloc contiguous|cling|cling_by_tags|normal|anywhere|inherit
80 Determines the allocation policy when a command needs to allo‐
81 cate Physical Extents (PEs) from the VG. Each VG and LV has an
82 allocation policy which can be changed with vgchange/lvchange,
83 or overriden on the command line. normal applies common sense
84 rules such as not placing parallel stripes on the same PV. in‐
85 herit applies the VG policy to an LV. contiguous requires new
86 PEs be placed adjacent to existing PEs. cling places new PEs on
87 the same PV as existing PEs in the same stripe of the LV. If
88 there are sufficient PEs for an allocation, but normal does not
89 use them, anywhere will use them even if it reduces performance,
90 e.g. by placing two stripes on the same PV. Optional positional
91 PV args on the command line can also be used to limit which PVs
92 the command will use for allocation. See lvm(8) for more infor‐
93 mation about allocation.
94
95 --atomic
96 Makes a pvmove operation atomic, ensuring that all affected LVs
97 are moved to the destination PV, or none are if the operation is
98 aborted.
99
100 -A|--autobackup y|n
101 Specifies if metadata should be backed up automatically after a
102 change. Enabling this is strongly advised! See vgcfgbackup(8)
103 for more information.
104
105 -b|--background
106 If the operation requires polling, this option causes the com‐
107 mand to return before the operation is complete, and polling is
108 done in the background.
109
110 --commandprofile String
111 The command profile to use for command configuration. See
112 lvm.conf(5) for more information about profiles.
113
114 --config String
115 Config settings for the command. These override lvm.conf set‐
116 tings. The String arg uses the same format as lvm.conf, or may
117 use section/field syntax. See lvm.conf(5) for more information
118 about config.
119
120 -d|--debug ...
121 Set debug level. Repeat from 1 to 6 times to increase the detail
122 of messages sent to the log file and/or syslog (if configured).
123
124 --driverloaded y|n
125 If set to no, the command will not attempt to use device-mapper.
126 For testing and debugging.
127
128 -h|--help
129 Display help text.
130
131 -i|--interval Number
132 Report progress at regular intervals.
133
134 --lockopt String
135 Used to pass options for special cases to lvmlockd. See lvm‐
136 lockd(8) for more information.
137
138 --longhelp
139 Display long help text.
140
141 -n|--name String
142 Move only the extents belonging to the named LV.
143
144 --nolocking
145 Disable locking.
146
147 --noudevsync
148 Disables udev synchronisation. The process will not wait for no‐
149 tification from udev. It will continue irrespective of any pos‐
150 sible udev processing in the background. Only use this if udev
151 is not running or has rules that ignore the devices LVM creates.
152
153 --profile String
154 An alias for --commandprofile or --metadataprofile, depending on
155 the command.
156
157 -q|--quiet ...
158 Suppress output and log messages. Overrides --debug and --ver‐
159 bose. Repeat once to also suppress any prompts with answer
160 'no'.
161
162 --reportformat basic|json
163 Overrides current output format for reports which is defined
164 globally by the report/output_format setting in lvm.conf. basic
165 is the original format with columns and rows. If there is more
166 than one report per command, each report is prefixed with the
167 report name for identification. json produces report output in
168 JSON format. See lvmreport(7) for more information.
169
170 -t|--test
171 Run in test mode. Commands will not update metadata. This is
172 implemented by disabling all metadata writing but nevertheless
173 returning success to the calling function. This may lead to un‐
174 usual error messages in multi-stage operations if a tool relies
175 on reading back metadata it believes has changed but hasn't.
176
177 -v|--verbose ...
178 Set verbose level. Repeat from 1 to 4 times to increase the de‐
179 tail of messages sent to stdout and stderr.
180
181 --version
182 Display version information.
183
184 -y|--yes
185 Do not prompt for confirmation interactively but always assume
186 the answer yes. Use with extreme caution. (For automatic no,
187 see -qq.)
188
190 PV
191 Physical Volume name, a device path under /dev. For commands
192 managing physical extents, a PV positional arg generally accepts
193 a suffix indicating a range (or multiple ranges) of physical ex‐
194 tents (PEs). When the first PE is omitted, it defaults to the
195 start of the device, and when the last PE is omitted it defaults
196 to end. Start and end range (inclusive): PV[:PE-PE]... Start
197 and length range (counting from 0): PV[:PE+PE]...
198
199 String
200 See the option description for information about the string con‐
201 tent.
202
203 Size[UNIT]
204 Size is an input number that accepts an optional unit. Input
205 units are always treated as base two values, regardless of capi‐
206 talization, e.g. 'k' and 'K' both refer to 1024. The default
207 input unit is specified by letter, followed by |UNIT. UNIT rep‐
208 resents other possible input units: bBsSkKmMgGtTpPeE. b|B is
209 bytes, s|S is sectors of 512 bytes, k|K is KiB, m|M is MiB, g|G
210 is GiB, t|T is TiB, p|P is PiB, e|E is EiB. (This should not be
211 confused with the output control --units, where capital letters
212 mean multiple of 1000.)
213
215 See lvm(8) for information about environment variables used by lvm.
216 For example, LVM_VG_NAME can generally be substituted for a required VG
217 parameter.
218
220 pvmove works as follows:
221
222 1. A temporary 'pvmove' LV is created to store details of all the data
223 movements required.
224
225 2. Every LV in the VG is searched for contiguous data that need moving
226 according to the command line arguments. For each piece of data found,
227 a new segment is added to the end of the pvmove LV. This segment takes
228 the form of a temporary mirror to copy the data from the original loca‐
229 tion to a newly allocated location. The original LV is updated to use
230 the new temporary mirror segment in the pvmove LV instead of accessing
231 the data directly.
232
233 3. The VG metadata is updated on disk.
234
235 4. The first segment of the pvmove LV is activated and starts to mirror
236 the first part of the data. Only one segment is mirrored at once as
237 this is usually more efficient.
238
239 5. A daemon repeatedly checks progress at the specified time interval.
240 When it detects that the first temporary mirror is in sync, it breaks
241 that mirror so that only the new location for that data gets used and
242 writes a checkpoint into the VG metadata on disk. Then it activates
243 the mirror for the next segment of the pvmove LV.
244
245 6. When there are no more segments left to be mirrored, the temporary
246 LV is removed and the VG metadata is updated so that the LVs reflect
247 the new data locations.
248
249 Note that this new process cannot support the original LVM1 type of on-
250 disk metadata. Metadata can be converted using vgconvert(8).
251
252 If the --atomic option is used, a slightly different approach is used
253 for the move. Again, a temporary 'pvmove' LV is created to store the
254 details of all the data movements required. This temporary LV contains
255 all the segments of the various LVs that need to be moved. However, in
256 this case, an identical LV is allocated that contains the same number
257 of segments and a mirror is created to copy the contents from the first
258 temporary LV to the second. After a complete copy is made, the tempo‐
259 rary LVs are removed, leaving behind the segments on the destination
260 PV. If an abort is issued during the move, all LVs being moved will
261 remain on the source PV.
262
263
265 Move all physical extents that are used by simple LVs on the specified
266 PV to free physical extents elsewhere in the VG.
267 pvmove /dev/sdb1
268
269 Use a specific destination PV when moving physical extents.
270 pvmove /dev/sdb1 /dev/sdc1
271
272 Move extents belonging to a single LV.
273 pvmove -n lvol1 /dev/sdb1 /dev/sdc1
274
275 Rather than moving the contents of an entire device, it is possible to
276 move a range of physical extents, for example numbers 1000 to 1999 in‐
277 clusive on the specified PV.
278 pvmove /dev/sdb1:1000-1999
279
280 A range of physical extents to move can be specified as start+length.
281 For example, starting from PE 1000. (Counting starts from 0, so this
282 refers to the 1001st to the 2000th PE inclusive.)
283 pvmove /dev/sdb1:1000+1000
284
285 Move a range of physical extents to a specific PV (which must have suf‐
286 ficient free extents).
287 pvmove /dev/sdb1:1000-1999 /dev/sdc1
288
289 Move a range of physical extents to specific new extents on a new PV.
290 pvmove /dev/sdb1:1000-1999 /dev/sdc1:0-999
291
292 If the source and destination are on the same disk, the anywhere allo‐
293 cation policy is needed.
294 pvmove --alloc anywhere /dev/sdb1:1000-1999 /dev/sdb1:0-999
295
296 The part of a specific LV present within in a range of physical extents
297 can also be picked out and moved.
298 pvmove -n lvol1 /dev/sdb1:1000-1999 /dev/sdc1
299
301 lvm(8) lvm.conf(5) lvmconfig(8)
302
303 pvchange(8) pvck(8) pvcreate(8) pvdisplay(8) pvmove(8) pvremove(8)
304 pvresize(8) pvs(8) pvscan(8)
305
306 vgcfgbackup(8) vgcfgrestore(8) vgchange(8) vgck(8) vgcreate(8) vgcon‐
307 vert(8) vgdisplay(8) vgexport(8) vgextend(8) vgimport(8) vgimport‐
308 clone(8) vgmerge(8) vgmknodes(8) vgreduce(8) vgremove(8) vgrename(8)
309 vgs(8) vgscan(8) vgsplit(8)
310
311 lvcreate(8) lvchange(8) lvconvert(8) lvdisplay(8) lvextend(8) lvre‐
312 duce(8) lvremove(8) lvrename(8) lvresize(8) lvs(8) lvscan(8)
313
314 lvm-fullreport(8) lvm-lvpoll(8) lvm2-activation-generator(8) blkdeacti‐
315 vate(8) lvmdump(8)
316
317 dmeventd(8) lvmpolld(8) lvmlockd(8) lvmlockctl(8) cmirrord(8) lvmd‐
318 busd(8)
319
320 lvmsystemid(7) lvmreport(7) lvmraid(7) lvmthin(7) lvmcache(7)
321
322
323
324Red Hat, Inc. LVM TOOLS 2.03.11(2) (2021-01-08) PVMOVE(8)