Configure the Internal Ceph Storage BackendΒΆ

This section provides steps to configure the internal Ceph storage backend. Depending on the system type, StarlingX can be configured with an internal Ceph storage backend on controller nodes or on dedicated storage nodes.

Prerequisites

Unlock all controllers before you run the following commands to configure the internal Ceph storage backend.

Procedure

  1. Run the following command:

    ~(keystone_admin)$ system storage-backend-add ceph --confirmed

    Note

    The default network used to deploy Ceph services on is the management network and this should be sufficient for most systems. The cluster-host network should only be considered for systems that are using a 1Gb management NIC.

    For cluster-host network, use the following command:

    ~(keystone_admin)]$ system storage-backend-add --network cluster-host ceph --confirmed

  2. Wait for Ceph storage to be configured. Run the following command to check if Ceph storage is configured:

    ~(keystone_admin)$ system storage-backend-list

  3. On a Standard configuration with Controller Storage, that is, where Ceph OSDs are to be configured on the controller nodes, configure the third Ceph monitor instance on a worker node, using the following command:

    ~(keystone_admin)$ system ceph-mon-add <worker_node>

    Note

    For Standard configuration with dedicated Storage, that is, where Ceph OSDs are to be configured on dedicated Storage nodes, the third Ceph monitor instance is configured by default on the first storage node.

    Note

    CephFS support requires Metadata servers (MDS) to be deployed. When CephFS is configured, an MDS is deployed automatically along with each node that has been configured to run a Ceph Monitor.

  4. Configure Ceph OSDs. For more information, see Provision Storage on a Controller or Storage Host Using Horizon.