1 Ceph RADOS Block Devices (RBD)
2 ------------------------------
4 Storage pool type: `rbd`
6 http://ceph.com[Ceph] is a distributed object store and file system
7 designed to provide excellent performance, reliability and
8 scalability. RADOS block devices implement a feature rich block level
9 storage, and you get the following advantages:
13 * distributed and redundant (striped over multiple OSDs)
14 * full snapshot and clone capabilities
16 * no single point of failure
17 * scalable to the exabyte level
18 * kernel and unser space implementation available
20 NOTE: For smaller deployments, it is also possible to run Ceph
21 services directly on your {pve} nodes. Recent hardware has plenty
22 of CPU power and RAM, so running storage services and VMs on same node
28 This backend supports the common storage properties `nodes`,
29 `disable`, `content`, and the following `rbd` specific properties:
33 List of monitor daemon IPs.
45 Access rbd through krbd kernel module. This is required if you want to
46 use the storage for containers.
48 .Configuration Example ('/etc/pve/storage.cfg')
51 monhost 10.1.1.20 10.1.1.21 10.1.1.22
57 TIP: You can use the 'rbd' utility to do low-level management tasks.
62 If you use cephx authentication, you need to copy the keyfile from
63 Ceph to Proxmox VE host.
65 Create the directory '/etc/pve/priv/ceph' with
67 mkdir /etc/pve/priv/ceph
71 scp <cephserver>:/etc/ceph/ceph.client.admin.keyring /etc/pve/priv/ceph/<STORAGE_ID>.keyring
73 The keyring must be named to match your `<STORAGE_ID>`. Copying the
74 keyring generally requires root privileges.
79 The `rbd` backend is a block level storage, and implements full
80 snapshot and clone functionality.
82 .Storage features for backend `rbd`
83 [width="100%",cols="m,m,3*d",options="header"]
84 |==============================================================================
85 |Content types |Image formats |Shared |Snapshots |Clones
86 |images rootdir |raw |yes |yes |yes
87 |==============================================================================