Ceph is a distributed storage and network file system designed to provide excellent performance, reliability, and scalability.
This charm deploys a Ceph cluster. juju
The ceph charm has two pieces of mandatory configuration for which no defaults are provided. You must set these configuration options before deployment or the charm will not work:
fsid:
uuid specific to a ceph cluster used to ensure that different
clusters don't get mixed up - use `uuid` to generate one.
monitor-secret:
a ceph generated key used by the daemons that manage to cluster
to control security. You can use the ceph-authtool command to
generate one:
ceph-authtool /dev/stdout --name=mon. --gen-key
These two pieces of configuration must NOT be changed post bootstrap; attempting to do this will cause a reconfiguration error and new service units will not join the existing ceph cluster.
The charm also supports the specification of storage devices to be used in the ceph cluster.
osd-devices:
A list of devices that the charm will attempt to detect, initialise and
activate as ceph storage.
This can be a superset of the actual storage devices presented to each
service unit and can be changed post ceph bootstrap using `juju set`.
The full path of each device must be provided, e.g. /dev/vdb.
For Ceph >= 0.56.6 (Raring or the Grizzly Cloud Archive) use of
directories instead of devices is also supported.
At a minimum you must provide a juju config file during initial deployment with the fsid and monitor-secret options (contents of cepy.yaml below):
ceph:
fsid: ecbb8960-0e21-11e2-b495-83a88f44db01
monitor-secret: AQD1P2xQiKglDhAA4NGUF5j38Mhq56qwz+45wg==
osd-devices: /dev/vdb /dev/vdc /dev/vdd /dev/vde
Specifying the osd-devices to use is also a good idea.
Boot things up by using:
juju deploy -n 3 --config ceph.yaml ceph
By default the ceph cluster will not bootstrap until 3 service units have been deployed and started; this is to ensure that a quorum is achieved prior to adding storage devices.
This charm supports pausing and resuming ceph's health functions on a cluster, for example when doing maintenance on a machine. to pause or resume, call:
juju action do --unit ceph/0 pause-health or juju action do --unit ceph/0 resume-health
You can use the Ceph OSD and Ceph Radosgw charms:
This charm supports the use of Juju Network Spaces, allowing the charm to be bound to network space configurations managed directly by Juju. This is only supported with Juju 2.0 and above.
Network traffic can be bound to specific network spaces using the public (front-side) and cluster (back-side) bindings:
juju deploy ceph --bind "public=data-space cluster=cluster-space"
alternatively these can also be provided as part of a Juju native bundle configuration:
ceph:
charm: cs:xenial/ceph
num_units: 1
bindings:
public: data-space
cluster: cluster-space
Please refer to the Ceph Network Reference for details on how using these options effects network traffic within a Ceph deployment.
NOTE: Spaces must be configured in the underlying provider prior to attempting to use them.
NOTE: Existing deployments using ceph-*-network configuration options will continue to function; these options are preferred over any network space binding provided if set.
- Paul Collins paul.collins@canonical.com,
- James Page james.page@ubuntu.com
Report bugs on Launchpad
This charm uses the new-style Ceph deployment as reverse-engineered from the Chef cookbook at https://github.com/ceph/ceph-cookbooks, although we selected a different strategy to form the monitor cluster. Since we don't know the names or addresses of the machines in advance, we use the relation-joined hook to wait for all three nodes to come up, and then write their addresses to ceph.conf in the "mon host" parameter. After we initialize the monitor cluster a quorum forms quickly, and OSD bringup proceeds.
The osds use so-called "OSD hotplugging". ceph-disk prepare is used to
create the filesystems with a special GPT partition type. udev is set up
to mount such filesystems and start the osd daemons as their storage becomes
visible to the system (or after udevadm trigger).
The Chef cookbook mentioned above performs some extra steps to generate an OSD bootstrapping key and propagate it to the other nodes in the cluster. Since all OSDs run on nodes that also run mon, we don't need this and did not implement it.
See the documentation for more information on Ceph monitor cluster deployment strategies and pitfalls.