Overview

The cinder charm deploys Cinder, the Block Storage (volume) service for OpenStack. The charm works alongside other Juju-deployed OpenStack services.

Usage

Configuration

To display all configuration option information run juju config <application>. If the application is not deployed then see the charm’s Configure tab in the Charmhub. Finally, the Juju documentation provides general guidance on configuring applications.

Deployment

The cinder application requires the following applications to be present: keystone, nova-cloud-controller, nova-compute, rabbitmq-server, and a cloud database.

The database application is determined by the series. Prior to focal percona-cluster is used, otherwise it is mysql-innodb-cluster. In the example deployment below mysql-innodb-cluster has been chosen.

Deploy Cinder itself (here, to a container on machine ‘1’), add relations to the core cloud applications, and then connect it to the cloud database:

juju deploy --to lxd:1 --config cinder.yaml cinder
juju add-relation cinder:identity-service keystone:identity-service
juju add-relation cinder:cinder-volume-service nova-cloud-controller:cinder-volume-service
juju add-relation cinder:amqp rabbitmq-server:amqp

juju deploy mysql-router cinder-mysql-router
juju add-relation cinder-mysql-router:db-router mysql-innodb-cluster:db-router
juju add-relation cinder-mysql-router:shared-db cinder:shared-db

Multiple backend storage solutions are described next.

Ceph-backed storage

Cinder can be backed by Ceph, which is the recommended storage method for production Cinder deployments. This functionality is provided by the [cinder-ceph][cinder-ceph-charm] subordinate charm.

LVM-backed storage

Cinder can be backed by storage local to the cinder unit, where local block devices are used as LVM physical volumes, and volumes are offered via iSCSI. This functionality is provided by the cinder-lvm subordinate charm.

Note: Built-in support for LVM in the cinder charm is deprecated.

NetApp-backed storage

Cinder can be backed by a NetApp appliance local to the cinder unit, where volumes are offered via iSCSI or NFS. This functionality is provided by the cinder-netapp subordinate charm.

Pure Storage-backed storage

Cinder can be backed by a Pure Storage appliance reachable by its API endpoint. This functionality is provided by the cinder-purestorage subordinate charm.

Storage Backend Secrets

Subordinate storage backends (e.g. cinder-ceph, cinder-netapp etc.) may send configuration values that contain sensitive credentials. Rather than writing these values directly in cinder.conf in plain text, the Cinder charm can resolve them at runtime via Castellan using Vault as a backend.

When a subordinate storage charm sends a Vault secret ref (vault://<secret-name>) as a config value, the cinder charm automatically configures Castellan, adds the secret reference to /etc/cinder/secret_map.conf and generates /etc/cinder/castellan.conf using the Vault backend.

The Cinder charm must be related to Vault over the secrets-storage interface and the secret should be created in the KV mountpoint created by the secrets-storage relation (e.g. charm-cinder/*).

Example (cinder-netapp):

juju integrate vault cinder:secrets-storage

# Below is an example for creating a secret in vault for the password: "password$long@#"
# Ensure the '
 character is escaped due to how oslo.config parses config files
secret_encoded=$(echo -n 'password\$long@#' | xxd -p)

vault write charm-cinder/netapp_password - <<EOF
{
  "type": "opaque",
  "value": "$secret_encoded",
  "name": "netapp_password",
  "created": $(date +%s)
}
EOF

juju config cinder-netapp netapp-password=="vault://netapp_password"

Separate Volume Service

For certain operations when an instance is not involved, the cinder application will connect directly to the storage for operations such as cloning a volume from a glance image. You can deploy a second cinder application for the volume service only where the primary cinder application cannot connect to this storage. This may be required for iSCSI connections because LXD containers cannot create iSCSI connections or where you need a physical Fibre Channel connection. This is not required for Ceph deployments which use userspace RBD tools.

Deploy cinder with enabled-services=api,scheduler
Deploy a second application of cinder named ‘cinder-volume’ with enabled-services=volume
Relate the storage subordinate (e.g. cinder-purestorage) to the cinder-volume application only (not to the ‘cinder’ application)
Keystone should be related to cinder:identity-service but cinder-volume:identity-credentials

The primary cinder application gets keystone credentials when registering a service endpoint via the identity-service relation. The cinder-volume application does not register a service, so we need to relate identity-credentials instead. The image volume cache will not work without this relation.
Both cinder and cinder-volume should otherwise have the same relations

High availability

This charm supports high availability via HAcluster.

When more than one unit is deployed with the hacluster application the charm will bring up an HA active/active cluster.

Network spaces

This charm supports the use of Juju network spaces (Juju v.2.0). This feature optionally allows specific types of the application’s network traffic to be bound to subnets that the underlying hardware is connected to.

Note: Spaces must be configured in the backing cloud prior to deployment.

API endpoints can be bound to distinct network spaces supporting the network separation of public, internal, and admin endpoints.

Access to the underlying MySQL instance can also be bound to a specific space using the shared-db relation.

For example, providing that spaces ‘public-space’, ‘internal-space’, and ‘admin-space’ exist, the deploy command above could look like this:

juju deploy --config cinder.yaml cinder \
   --bind "public=public-space internal=internal-space admin=admin-space shared-db=internal-space"

Alternatively, configuration can be provided as part of a bundle:

    cinder:
      charm: cs:cinder
      num_units: 1
      bindings:
        public: public-space
        internal: internal-space
        admin: admin-space
        shared-db: internal-space

Note: Existing cinder units configured with the os-admin-network, os-internal-network, or os-public-network options will continue to honour them. Furthermore, these options override any space bindings, if set.

Actions

This charm supports actions.

Actions allow specific operations to be performed on a per-unit basis. To display actions and their descriptions run juju actions --schema <application>. If the application is not deployed then see the charm’s Actions tab in the Charmhub.

Policy overrides

This charm supports the policy overrides feature.

Policy overrides allow an operator to override the default policy of an OpenStack service. See Policy overrides for more information on this feature.

Documentation

The OpenStack Charms project maintains two documentation guides:

OpenStack Charm Guide: the primary source of information for OpenStack charms
OpenStack Charms Deployment Guide: a step-by-step guide for deploying OpenStack with charms

Bugs

Please report bugs on Launchpad.