Scale a PVC-based Ceph cluster

For specific instructions to scale a Ceph cluster running on OpenShift 4 on Exoscale see the VSHN OpenShift 4 Exoscale how-tos to Add a storage node and Change storage node size.

This how-to assumes that you’ve setup a PVC-based Ceph cluster on an infrastructure which has a CSI driver which supports both volumeMode: Block and support for volume expansion.


  • A local setup to compile a cluster catalog, see Running Commodore for details.

  • An access token for the Lieutenant instance on which your cluster is registered.

This guide assumes that you’re familiar with making changes to a Project Syn cluster and compiling the cluster catalog to deploy those changes.

Scaling the OSD disks of a PVC-based Ceph cluster

On infrastructures which support volume expansion, scaling the OSD disks of a PVC-based Ceph cluster is very straightforward.

  1. Adjust the desired PVC size in the cluster config (in this example we’re growing the cluster from 3x100Gi OSDs to 3x150Gi OSDs):

    c-cluster-id.yml diff
           block_storage_class: ssd
    -      block_volume_size: 100Gi
    +      block_volume_size: 150Gi
           tune_fast_device_class: true
           rbd_enabled: true
           cephfs_enabled: false
  2. Compile the cluster catalog

  3. Wait until ArgoCD has updated the CephCluster resource and Rook has updated the OSD PVCs to match the CephCluster spec and restarted the OSDs.

Scaling the OSD count of a PVC-based Ceph cluster

  1. Make sure you have sufficient compute resources in your cluster to add the desired amount of OSDs.

    By default, the component configures the resource requests for each OSD as 4 CPU and 5Gi of memory.
  2. Update the cluster configuration to deploy more OSDs

    +      node_count: 4 (1)
           block_storage_class: ssd
           block_volume_size: 110Gi
           osd_portable: true
    1 Update the OSD count from 3 to 4. 3 is the default value of the parameter.
  3. Compile the cluster catalog

  4. Wait until ArgoCD has updated the CephCluster resource and Rook has deployed the new OSDs.