]>
Commit | Line | Data |
---|---|---|
aa039b0f DM |
1 | Ceph RADOS Block Devices (RBD) |
2 | ------------------------------ | |
5f09af76 DM |
3 | ifdef::wiki[] |
4 | :pve-toplevel: | |
cb84ed18 | 5 | :title: Storage: RBD |
5f09af76 DM |
6 | endif::wiki[] |
7 | ||
aa039b0f DM |
8 | Storage pool type: `rbd` |
9 | ||
10 | http://ceph.com[Ceph] is a distributed object store and file system | |
11 | designed to provide excellent performance, reliability and | |
12 | scalability. RADOS block devices implement a feature rich block level | |
13 | storage, and you get the following advantages: | |
14 | ||
15 | * thin provisioning | |
16 | * resizable volumes | |
17 | * distributed and redundant (striped over multiple OSDs) | |
18 | * full snapshot and clone capabilities | |
19 | * self healing | |
20 | * no single point of failure | |
21 | * scalable to the exabyte level | |
5eba0743 | 22 | * kernel and user space implementation available |
aa039b0f DM |
23 | |
24 | NOTE: For smaller deployments, it is also possible to run Ceph | |
25 | services directly on your {pve} nodes. Recent hardware has plenty | |
26 | of CPU power and RAM, so running storage services and VMs on same node | |
27 | is possible. | |
28 | ||
29 | Configuration | |
30 | ~~~~~~~~~~~~~ | |
31 | ||
32 | This backend supports the common storage properties `nodes`, | |
33 | `disable`, `content`, and the following `rbd` specific properties: | |
34 | ||
35 | monhost:: | |
36 | ||
37 | List of monitor daemon IPs. | |
38 | ||
39 | pool:: | |
40 | ||
41 | Ceph pool name. | |
42 | ||
43 | username:: | |
44 | ||
45 | RBD user Id. | |
46 | ||
47 | krbd:: | |
48 | ||
49 | Access rbd through krbd kernel module. This is required if you want to | |
50 | use the storage for containers. | |
51 | ||
8c1189b6 | 52 | .Configuration Example (`/etc/pve/storage.cfg`) |
aa039b0f DM |
53 | ---- |
54 | rbd: ceph3 | |
55 | monhost 10.1.1.20 10.1.1.21 10.1.1.22 | |
56 | pool ceph3 | |
57 | content images | |
58 | username admin | |
59 | ---- | |
60 | ||
8c1189b6 | 61 | TIP: You can use the `rbd` utility to do low-level management tasks. |
aa039b0f DM |
62 | |
63 | Authentication | |
64 | ~~~~~~~~~~~~~~ | |
65 | ||
8c1189b6 | 66 | If you use `cephx` authentication, you need to copy the keyfile from |
aa039b0f DM |
67 | Ceph to Proxmox VE host. |
68 | ||
8c1189b6 | 69 | Create the directory `/etc/pve/priv/ceph` with |
aa039b0f DM |
70 | |
71 | mkdir /etc/pve/priv/ceph | |
72 | ||
73 | Then copy the keyring | |
74 | ||
75 | scp <cephserver>:/etc/ceph/ceph.client.admin.keyring /etc/pve/priv/ceph/<STORAGE_ID>.keyring | |
76 | ||
77 | The keyring must be named to match your `<STORAGE_ID>`. Copying the | |
78 | keyring generally requires root privileges. | |
79 | ||
80 | Storage Features | |
81 | ~~~~~~~~~~~~~~~~ | |
82 | ||
83 | The `rbd` backend is a block level storage, and implements full | |
84 | snapshot and clone functionality. | |
85 | ||
86 | .Storage features for backend `rbd` | |
87 | [width="100%",cols="m,m,3*d",options="header"] | |
88 | |============================================================================== | |
89 | |Content types |Image formats |Shared |Snapshots |Clones | |
90 | |images rootdir |raw |yes |yes |yes | |
91 | |============================================================================== | |
92 | ||
deb4673f DM |
93 | ifdef::wiki[] |
94 | ||
95 | See Also | |
96 | ~~~~~~~~ | |
97 | ||
f532afb7 | 98 | * link:/wiki/Storage[Storage] |
deb4673f DM |
99 | |
100 | endif::wiki[] | |
101 |