Root/
1 | Device-mapper snapshot support |
2 | ============================== |
3 | |
4 | Device-mapper allows you, without massive data copying: |
5 | |
6 | *) To create snapshots of any block device i.e. mountable, saved states of |
7 | the block device which are also writable without interfering with the |
8 | original content; |
9 | *) To create device "forks", i.e. multiple different versions of the |
10 | same data stream. |
11 | *) To merge a snapshot of a block device back into the snapshot's origin |
12 | device. |
13 | |
14 | In the first two cases, dm copies only the chunks of data that get |
15 | changed and uses a separate copy-on-write (COW) block device for |
16 | storage. |
17 | |
18 | For snapshot merge the contents of the COW storage are merged back into |
19 | the origin device. |
20 | |
21 | |
22 | There are three dm targets available: |
23 | snapshot, snapshot-origin, and snapshot-merge. |
24 | |
25 | *) snapshot-origin <origin> |
26 | |
27 | which will normally have one or more snapshots based on it. |
28 | Reads will be mapped directly to the backing device. For each write, the |
29 | original data will be saved in the <COW device> of each snapshot to keep |
30 | its visible content unchanged, at least until the <COW device> fills up. |
31 | |
32 | |
33 | *) snapshot <origin> <COW device> <persistent?> <chunksize> |
34 | |
35 | A snapshot of the <origin> block device is created. Changed chunks of |
36 | <chunksize> sectors will be stored on the <COW device>. Writes will |
37 | only go to the <COW device>. Reads will come from the <COW device> or |
38 | from <origin> for unchanged data. <COW device> will often be |
39 | smaller than the origin and if it fills up the snapshot will become |
40 | useless and be disabled, returning errors. So it is important to monitor |
41 | the amount of free space and expand the <COW device> before it fills up. |
42 | |
43 | <persistent?> is P (Persistent) or N (Not persistent - will not survive |
44 | after reboot). |
45 | The difference is that for transient snapshots less metadata must be |
46 | saved on disk - they can be kept in memory by the kernel. |
47 | |
48 | |
49 | * snapshot-merge <origin> <COW device> <persistent> <chunksize> |
50 | |
51 | takes the same table arguments as the snapshot target except it only |
52 | works with persistent snapshots. This target assumes the role of the |
53 | "snapshot-origin" target and must not be loaded if the "snapshot-origin" |
54 | is still present for <origin>. |
55 | |
56 | Creates a merging snapshot that takes control of the changed chunks |
57 | stored in the <COW device> of an existing snapshot, through a handover |
58 | procedure, and merges these chunks back into the <origin>. Once merging |
59 | has started (in the background) the <origin> may be opened and the merge |
60 | will continue while I/O is flowing to it. Changes to the <origin> are |
61 | deferred until the merging snapshot's corresponding chunk(s) have been |
62 | merged. Once merging has started the snapshot device, associated with |
63 | the "snapshot" target, will return -EIO when accessed. |
64 | |
65 | |
66 | How snapshot is used by LVM2 |
67 | ============================ |
68 | When you create the first LVM2 snapshot of a volume, four dm devices are used: |
69 | |
70 | 1) a device containing the original mapping table of the source volume; |
71 | 2) a device used as the <COW device>; |
72 | 3) a "snapshot" device, combining #1 and #2, which is the visible snapshot |
73 | volume; |
74 | 4) the "original" volume (which uses the device number used by the original |
75 | source volume), whose table is replaced by a "snapshot-origin" mapping |
76 | from device #1. |
77 | |
78 | A fixed naming scheme is used, so with the following commands: |
79 | |
80 | lvcreate -L 1G -n base volumeGroup |
81 | lvcreate -L 100M --snapshot -n snap volumeGroup/base |
82 | |
83 | we'll have this situation (with volumes in above order): |
84 | |
85 | # dmsetup table|grep volumeGroup |
86 | |
87 | volumeGroup-base-real: 0 2097152 linear 8:19 384 |
88 | volumeGroup-snap-cow: 0 204800 linear 8:19 2097536 |
89 | volumeGroup-snap: 0 2097152 snapshot 254:11 254:12 P 16 |
90 | volumeGroup-base: 0 2097152 snapshot-origin 254:11 |
91 | |
92 | # ls -lL /dev/mapper/volumeGroup-* |
93 | brw------- 1 root root 254, 11 29 ago 18:15 /dev/mapper/volumeGroup-base-real |
94 | brw------- 1 root root 254, 12 29 ago 18:15 /dev/mapper/volumeGroup-snap-cow |
95 | brw------- 1 root root 254, 13 29 ago 18:15 /dev/mapper/volumeGroup-snap |
96 | brw------- 1 root root 254, 10 29 ago 18:14 /dev/mapper/volumeGroup-base |
97 | |
98 | |
99 | How snapshot-merge is used by LVM2 |
100 | ================================== |
101 | A merging snapshot assumes the role of the "snapshot-origin" while |
102 | merging. As such the "snapshot-origin" is replaced with |
103 | "snapshot-merge". The "-real" device is not changed and the "-cow" |
104 | device is renamed to <origin name>-cow to aid LVM2's cleanup of the |
105 | merging snapshot after it completes. The "snapshot" that hands over its |
106 | COW device to the "snapshot-merge" is deactivated (unless using lvchange |
107 | --refresh); but if it is left active it will simply return I/O errors. |
108 | |
109 | A snapshot will merge into its origin with the following command: |
110 | |
111 | lvconvert --merge volumeGroup/snap |
112 | |
113 | we'll now have this situation: |
114 | |
115 | # dmsetup table|grep volumeGroup |
116 | |
117 | volumeGroup-base-real: 0 2097152 linear 8:19 384 |
118 | volumeGroup-base-cow: 0 204800 linear 8:19 2097536 |
119 | volumeGroup-base: 0 2097152 snapshot-merge 254:11 254:12 P 16 |
120 | |
121 | # ls -lL /dev/mapper/volumeGroup-* |
122 | brw------- 1 root root 254, 11 29 ago 18:15 /dev/mapper/volumeGroup-base-real |
123 | brw------- 1 root root 254, 12 29 ago 18:16 /dev/mapper/volumeGroup-base-cow |
124 | brw------- 1 root root 254, 10 29 ago 18:16 /dev/mapper/volumeGroup-base |
125 | |
126 | |
127 | How to determine when a merging is complete |
128 | =========================================== |
129 | The snapshot-merge and snapshot status lines end with: |
130 | <sectors_allocated>/<total_sectors> <metadata_sectors> |
131 | |
132 | Both <sectors_allocated> and <total_sectors> include both data and metadata. |
133 | During merging, the number of sectors allocated gets smaller and |
134 | smaller. Merging has finished when the number of sectors holding data |
135 | is zero, in other words <sectors_allocated> == <metadata_sectors>. |
136 | |
137 | Here is a practical example (using a hybrid of lvm and dmsetup commands): |
138 | |
139 | # lvs |
140 | LV VG Attr LSize Origin Snap% Move Log Copy% Convert |
141 | base volumeGroup owi-a- 4.00g |
142 | snap volumeGroup swi-a- 1.00g base 18.97 |
143 | |
144 | # dmsetup status volumeGroup-snap |
145 | 0 8388608 snapshot 397896/2097152 1560 |
146 | ^^^^ metadata sectors |
147 | |
148 | # lvconvert --merge -b volumeGroup/snap |
149 | Merging of volume snap started. |
150 | |
151 | # lvs volumeGroup/snap |
152 | LV VG Attr LSize Origin Snap% Move Log Copy% Convert |
153 | base volumeGroup Owi-a- 4.00g 17.23 |
154 | |
155 | # dmsetup status volumeGroup-base |
156 | 0 8388608 snapshot-merge 281688/2097152 1104 |
157 | |
158 | # dmsetup status volumeGroup-base |
159 | 0 8388608 snapshot-merge 180480/2097152 712 |
160 | |
161 | # dmsetup status volumeGroup-base |
162 | 0 8388608 snapshot-merge 16/2097152 16 |
163 | |
164 | Merging has finished. |
165 | |
166 | # lvs |
167 | LV VG Attr LSize Origin Snap% Move Log Copy% Convert |
168 | base volumeGroup owi-a- 4.00g |
169 |
Branches:
ben-wpan
ben-wpan-stefan
javiroman/ks7010
jz-2.6.34
jz-2.6.34-rc5
jz-2.6.34-rc6
jz-2.6.34-rc7
jz-2.6.35
jz-2.6.36
jz-2.6.37
jz-2.6.38
jz-2.6.39
jz-3.0
jz-3.1
jz-3.11
jz-3.12
jz-3.13
jz-3.15
jz-3.16
jz-3.18-dt
jz-3.2
jz-3.3
jz-3.4
jz-3.5
jz-3.6
jz-3.6-rc2-pwm
jz-3.9
jz-3.9-clk
jz-3.9-rc8
jz47xx
jz47xx-2.6.38
master
Tags:
od-2011-09-04
od-2011-09-18
v2.6.34-rc5
v2.6.34-rc6
v2.6.34-rc7
v3.9