Ceph CSI

Learn about configurations >

• ceph-ext4-storage-class-name-formatter | string

  Default: ceph-ext4

  Formatter for the ceph ext4 storage class name

  The formatter is a string that can contain the following placeholders:

  • {app} - the name of the juju application
  • {namespace} - the charm configured namespace

  Example: juju config ceph-csi ceph-ext4-storage-class-name-formatter="ceph-ext4-{app}"

  The default is to use the storage class name

• ceph-ext4-storage-class-parameters | string

  Default: imageFeatures=layering

  Parameters to be used when creating the the ceph-ext4 storage class. Changes are only applied to the storage class if it does not exist.

  Declare additional/replacement parameters in key=value format, separated by spaces. Declare removed parameters in the key- format, separated by spaces.

  Optional parameters can be found in the ceph-csi documentation: https://github.com/ceph/ceph-csi/blob/devel/examples/rbd/storageclass.yaml

• ceph-rbac-name-formatter | string

  Default: {name}

  Formatter for colliding kubernetes rbac resources

  The formatter is used to create a unique name for the resource when the name collides with an existing resource. This is useful when deploying multiple ceph-csi charms.

  This is a list of non-namespaced resources affected by this formatter:

  • ClusterRole/cephfs-csi-nodeplugin
  • ClusterRoleBinding/cephfs-csi-nodeplugin
  • ClusterRole/cephfs-external-provisioner-runner
  • ClusterRoleBinding/cephfs-external-provisioner-runner
  • ClusterRole/rbd-csi-nodeplugin
  • ClusterRoleBinding/rbd-csi-nodeplugin
  • ClusterRole/rbd-external-provisioner-runner
  • ClusterRoleBinding/rbd-csi-provisioner-role

  The formatter is a string that can contain the following placeholders:

  • {app} - the name of the juju application
  • {name} - the name of the kubernetes resource
  • {namespace} - the charm configured namespace

  Example: juju config ceph-csi ceph-rbac-name-formatter="{name}-{app}"

  The default is to use the {name} of the resource

• ceph-rbd-enable | boolean

  Default: True

  Whether or not ceph-rbd manifests should be installed. This controls both storage classes xfs and ext4

• ceph-rbd-tolerations | string

  Tolerations to be used when creating the rbd pods for Daemonsets and Deployments.

  Declare tolerations in key=value,operator,effect format, separating each by spaces.

  Optional tolerations can be found in the kubernetes documentation: https://kubernetes.io/docs/concepts/scheduling-eviction/taint-and-toleration/

• ceph-xfs-storage-class-name-formatter | string

  Default: ceph-xfs

  Formatter for the ceph xfs storage class name

  The formatter is a string that can contain the following placeholders:

  • {app} - the name of the juju application
  • {namespace} - the charm configured namespace

  Example: juju config ceph-csi ceph-xfs-storage-class-name-formatter="ceph-xfs-{app}"

  The default is to use the storage class name

• ceph-xfs-storage-class-parameters | string

  Default: imageFeatures=layering

  Parameters to be used when creating the ceph-xfs storage class. Changes are only applied to the storage class if it does not exist.

  Declare additional/replacement parameters in key=value format, separated by spaces. Declare removed parameters in the key- format, separated by spaces.

  Optional parameters can be found in the ceph-csi documentation: https://github.com/ceph/ceph-csi/blob/devel/examples/rbd/storageclass.yaml

• cephfs-enable | boolean

  Whether or not cephfs manifests should be installed

  Only enable if ceph-fs is deployed and related to ceph-mon

• cephfs-mounter | string

  The driver can use either ceph-fuse (fuse) or ceph kernelclient (kernel). If omitted, default volume mounter will be used - this is determined by probing for ceph-fuse and mount.ceph

• cephfs-storage-class-name-formatter | string

  Default: cephfs

  Formatter for the cephfs storage class name

  The formatter is a string that can contain the following placeholders:

  • {app} - the name of the juju application
  • {namespace} - the charm configured namespace
  • {name} - the name of the filesystem
  • {pool} - the name of the data-pool
  • {pool-id} - the id of the data-pool

  Example: juju config ceph-csi cephfs-storage-class-name-formatter="cephfs-{namespace}-{pool}"

  The default is to use the storage class name

• cephfs-storage-class-parameters | string

  Parameters to be used when creating the cephfs storage classes. Changes are only applied to the storage classes if they do not exist.

  Declare additional/replacement parameters in key=value format, separated by spaces. Declare removed parameters in the key- format, separated by spaces.

  Optional parameters can be found in the ceph-csi documentation: https://github.com/ceph/ceph-csi/blob/devel/examples/rbd/storageclass.yaml

• cephfs-tolerations | string

  Default: $csi-cephfsplugin-legacy$

  Tolerations to be used when creating the cephfs pods for Daemonsets and Deployments.

  Declare tolerations in key=value,operator,effect format, separating each by spaces.

  Optional tolerations can be found in the kubernetes documentation: https://kubernetes.io/docs/concepts/scheduling-eviction/taint-and-toleration/

  Note: The default value "$csi-cephfsplugin-legacy$" is a special token which applies a "=,Exists" toleration only to pods associated with csi-cephfsplugin.

• create-namespace | boolean

  Whether or not the charm should create the namespace specified in the namespace configuration option.

  If set to true, the charm will create the namespace if it does not exist. If set to false, the charm will not create the namespace and will expect it to be created by the user.

  NOTE: Can only be specified on deployment since some attributes of kubernetes resources are non-modifiable.

• csidriver-name-formatter | string

  Default: {name}

  Formatter for the 2 CSIDrivers managed by this charm:

  • rbd.csi.ceph.com
  • cephfs.csi.ceph.com

  The formatter is a string that can contain the following placeholders:

  • {name} - original name from the manifest
  • {app} - the name of the juju application
  • {namespace} - the charm configured namespace

  Example: juju config ceph-csi csidriver-name-formatter="{app}.{name}"

  The default is to use the built-in storage classname

  NOTE: Can only be specified on deployment since some attributes of kubernetes resources cannot be modified once installed.

• default-storage | string

  Default: ceph-xfs

  Default storage class to use. Allowed values: 'ceph-ext4', 'ceph-xfs', 'cephfs'

• enable-host-networking | boolean

  Whether or not csi-*plugin-provisioner deployments use host-networking

• image-registry | string

  Default: rocks.canonical.com/cdk

  Image registry for all Ceph CSI container images.

  This value replaces the image registry in image URLs from the release manifests. For example, an image URL like quay.io/cephcsi/cephcsi:v3.8.0 becomes rocks.canonical.com/cdk/cephcsi/cephcsi:v3.8.0 if set to rocks.canonical.com/cdk.

  Example: juju config ceph-csi image-registry="rocks.canonical.com/cdk"

• metrics-port-cephfsplugin | int

  Default: -1

  Port for csi-cephfsplugin liveness-prometheus metrics

  If set to -1, the metrics service will not be created

• metrics-port-cephfsplugin-provisioner | int

  Default: -1

  Port for csi-cephfsplugin-provisioner liveness-prometheus metrics

  If set to -1, the metrics service will not be created

• metrics-port-rbdplugin | int

  Default: -1

  Port for csi-rbdplugin liveness-prometheus metrics

  If set to -1, the metrics service will not be created

• metrics-port-rbdplugin-provisioner | int

  Default: -1

  Port for csi-rbdplugin-provisioner liveness-prometheus metrics

  If set to -1, the metrics service will not be created

• namespace | string

  Kubernetes namespace in which to install the rbd and cephfs components including deployments, daemonsets, secrets, configmaps.

  if unspecified, "default" is assumed

  NOTE: Can only be specified on deployment since some attributes of kubernetes resources are non-modifiable. The admin is responsible for creating the namespace.

• provisioner-replicas | int

  Default: 3

  Number of replicas of any csi-*plugin-provisioner deployment

• release | string

  Default: v3.9.0

  Specify the version of ceph-csi as defined by the release tags of https://github.com/ceph/ceph-csi

  example) juju config ceph-csi release='v3.7.2'

  A list of supported versions is available through the action: juju run ceph-csi/leader list-versions

  To reset by to the latest supported by the charm use: juju config ceph-csi --reset release

  The current release deployed is available by viewing juju status ceph-csi