BlockDevices are the consumable units of storage in the OpenEBS storage solution. Currently, NDM supports the discovery and management of only a complete disk.

However, users can create BlockDevices manually to support custom partitions/lvms, etc., which are not detected by NDM. To create a manual BlockDevice, follow the steps below:

  1. Download the sample block device custom resource YAML file.

2.  Edit the file and fill in the details of the blockdevice. Fields marked with optional are not mandatory and can be removed. All other fields are required and information provided will be used while claiming.

kind: BlockDevice
  name: example-blockdevice
  labels: <host name of the node in which disk/blockdevice is attached> # like gke-openebs-user-default-pool-044afcb8-bmc0 "false" # for manual disk creation put false blockdevice
  claimState: Unclaimed
  state: Active
       storage: <total capacity in bytes> #like 53687091200
    firmwareRevision: <firmware revision> #optional
    model: <model name of blockdevice> # like PersistentDisk, optional
    serial: <serial no of disk> # like google-disk-2, optional
    compliance: <compliance of disk> #like "SPC-4", optional
    vendor: <vendor of disk> #like Google, optional
  - kind: by-id
    - <link1> # like /dev/disk/by-id/scsi-0Google_PersistentDisk_disk-2
    - <link2> # like /dev/disk/by-id/google-disk-2
  - kind: by-path
    - <link1> # like /dev/disk/by-path/virtio-pci-0000:00:03.0-scsi-0:0:2:0
    nodeName: <node name> # output of `kubectl get nodes` can be used
  path: <devpath> # like /dev/sdb

3.  Apply the YAML file.

kubectl apply -f openebs_v1alpha1_blockdevice_cr.yaml

The BlockDevice CR will be created and is then used by NDM Operator for claiming, but it won’t be managed by NDM Daemon for any changes that happen on the device. However, all the Claim/Unclaim operations and cleanup operations will be performed on this BlockDevice.

Please provide your valuable feedback & comments below and let me know what I can cover in my next blog.

This blog was originally published on Oct 22nd, 2019, on the MayaData blog.