Ceph RBD - ceph¶
Ceph is an open-source storage platform that stores its data in a storage cluster based on RADOS. It is highly scalable and, as a distributed system without a single point of failure, very reliable.
Tip
If you want to quickly set up a basic Ceph cluster, check out MicroCeph.
Ceph provides different components for block storage and for file systems.
Ceph RBD is Ceph’s block storage component that distributes data and workload across the Ceph cluster. It uses thin provisioning, which means that it is possible to over-commit resources.
Terminology¶
Ceph uses the term object for the data that it stores. The daemon that is responsible for storing and managing data is the Ceph OSD. Ceph’s storage is divided into pools, which are logical partitions for storing objects. They are also referred to as data pools, storage pools or OSD pools.
Ceph block devices are also called RBD images, and you can create snapshots and clones of these RBD images.
ceph driver in LXD¶
Note
To use the Ceph RBD driver, you must specify it as ceph.
This is slightly misleading, because it uses only Ceph RBD (block storage) functionality, not full Ceph functionality.
For storage volumes with content type filesystem (images, containers and custom file-system volumes), the ceph driver uses Ceph RBD images with a file system on top (see block.filesystem).
Alternatively, you can use the CephFS driver to create storage volumes with content type filesystem.
Unlike other storage drivers, this driver does not set up the storage system but assumes that you already have a Ceph cluster installed.
This driver also behaves differently than other drivers in that it provides remote storage. As a result and depending on the internal network, storage access might be a bit slower than for local storage. On the other hand, using remote storage has big advantages in a cluster setup, because all cluster members have access to the same storage pools with the exact same contents, without the need to synchronize storage pools.
The ceph driver in LXD uses RBD images for images, and snapshots and clones to create instances and snapshots.
LXD assumes that it has full control over the OSD storage pool. Therefore, you should never maintain any file system entities that are not owned by LXD in a LXD OSD storage pool, because LXD might delete them.
Due to the way copy-on-write works in Ceph RBD, parent RBD images can’t be removed until all children are gone.
As a result, LXD automatically renames any objects that are removed but still referenced.
Such objects are kept with a zombie_ prefix until all references are gone and the object can safely be removed.
Limitations¶
The ceph driver has the following limitations:
- Sharing custom volumes between instances
Custom storage volumes with content type
filesystemcan usually be shared between multiple instances different cluster members. However, because the Ceph RBD driver “simulates” volumes with content typefilesystemby putting a file system on top of an RBD image, custom storage volumes can only be assigned to a single instance at a time. If you need to share a custom volume with content typefilesystem, use the CephFS driver instead.- Sharing the OSD storage pool between installations
Sharing the same OSD storage pool between multiple LXD installations is not supported.
- Using an OSD pool of type “erasure”
To use a Ceph OSD pool of type “erasure”, you must create the OSD pool beforehand. You must also create a separate OSD pool of type “replicated” that will be used for storing metadata. This is required because Ceph RBD does not support
omap. To specify which pool is “erasure coded”, set theceph.osd.data_pool_nameconfiguration option to the erasure coded pool name and thesourceconfiguration option to the replicated pool name.
Configuration options¶
The following configuration options are available for storage pools that use the ceph driver and for storage volumes in these pools.
Storage pool configuration¶
| Key: | ceph.cluster_name |
| Type: | string |
| Default: |
|
| Key: | ceph.osd.pg_num |
| Type: | string |
| Default: |
|
| Key: | ceph.osd.pool_name |
| Type: | string |
| Default: | name of the pool |
| Key: | ceph.rbd.clone_copy |
| Type: | bool |
| Default: |
|
Enable this option to use RBD lightweight clones rather than full dataset copies.
| Key: | ceph.rbd.du |
| Type: | bool |
| Default: |
|
This option specifies whether to use RBD du to obtain disk usage data for stopped instances.
| Key: | ceph.rbd.features |
| Type: | string |
| Default: |
|
| Key: | ceph.user.name |
| Type: | string |
| Default: |
|
| Key: | volatile.pool.pristine |
| Type: | string |
| Default: |
|
Tip
In addition to these configurations, you can also set default values for the storage volume configurations. See Configure default values for storage volumes.
Storage volume configuration¶
| Key: | block.filesystem |
| Type: | string |
| Default: | same as |
| Condition: | block-based volume with content type |
Valid options are: btrfs, ext4, xfs
If not set, ext4 is assumed.
| Key: | block.mount_options |
| Type: | string |
| Default: | same as |
| Condition: | block-based volume with content type |
| Key: | security.shifted |
| Type: | bool |
| Default: | same as |
| Condition: | custom volume |
Enabling this option allows attaching the volume to multiple isolated instances.
| Key: | security.unmapped |
| Type: | bool |
| Default: | same as |
| Condition: | custom volume |
| Key: | size |
| Type: | string |
| Default: | same as |
| Condition: | appropriate driver |
| Key: | snapshots.expiry |
| Type: | string |
| Default: | same as |
| Condition: | custom volume |
Specify an expression like 1M 2H 3d 4w 5m 6y.
| Key: | snapshots.pattern |
| Type: | string |
| Default: | same as |
| Condition: | custom volume |
You can specify a naming template that is used for scheduled snapshots and unnamed snapshots.
The snapshots.pattern option takes a Pongo2 template string to format the snapshot name.
To add a time stamp to the snapshot name, use the Pongo2 context variable creation_date.
Make sure to format the date in your template string to avoid forbidden characters in the snapshot name.
For example, set snapshots.pattern to {{ creation_date|date:'2006-01-02_15-04-05' }} to name the snapshots after their time of creation, down to the precision of a second.
Another way to avoid name collisions is to use the placeholder %d in the pattern.
For the first snapshot, the placeholder is replaced with 0.
For subsequent snapshots, the existing snapshot names are taken into account to find the highest number at the placeholder’s position.
This number is then incremented by one for the new name.
| Key: | snapshots.schedule |
| Type: | string |
| Default: | same as |
| Condition: | custom volume |
Specify either a cron expression (<minute> <hour> <dom> <month> <dow>), a comma-separated list of schedule aliases (@hourly, @daily, @midnight, @weekly, @monthly, @annually, @yearly), or leave empty to disable automatic snapshots (the default).