(kubectl) Creating a PV from an Existing EVS Disk

Notes and Constraints

The following configuration example applies to clusters of Kubernetes 1.13 or earlier.

Procedure

Log in to the EVS console, create an EVS disk, and record the volume ID, capacity, and disk type of the EVS disk.
Use kubectl to connect to the cluster. For details, see Connecting to a Cluster Using kubectl.

Create two YAML files for creating the PersistentVolume (PV) and PersistentVolumeClaim (PVC). Assume that the file names are pv-evs-example.yaml and pvc-evs-example.yaml.

touch pv-evs-example.yaml pvc-evs-example.yaml

Kubernetes Cluster Version	Description	YAML Example
1.11.7 ≤ K8s version ≤ 1.13	Clusters from v1.11.7 to v1.13	Example YAML
1.11 ≤ K8s version < 1.11.7	Clusters from v1.11 to v1.11.7	Example YAML
K8s version = 1.9	Clusters of v1.9	Example YAML

Clusters from v1.11.7 to v1.13

Example YAML file for the PV:

apiVersion: v1 
kind: PersistentVolume 
metadata: 
  labels: 
    failure-domain.beta.kubernetes.io/region: ap-southeast-1
    failure-domain.beta.kubernetes.io/zone:  ap-southeast-1a
  annotations:
    pv.kubernetes.io/provisioned-by: flexvolume-huawei.com/fuxivol
  name: pv-evs-example 
spec: 
  accessModes: 
  - ReadWriteOnce 
  capacity: 
    storage: 10Gi 
  claimRef:
    apiVersion: v1
    kind: PersistentVolumeClaim
    name: pvc-evs-example
    namespace: default
  flexVolume: 
    driver: huawei.com/fuxivol 
    fsType: ext4 
    options:
      disk-mode: SCSI
      fsType: ext4 
      volumeID: 0992dbda-6340-470e-a74e-4f0db288ed82 
  persistentVolumeReclaimPolicy: Delete 
  storageClassName: sas

**Table 1** Key parameters
Parameter	Description
failure-domain.beta.kubernetes.io/region	Region where the cluster is located.
failure-domain.beta.kubernetes.io/zone	AZ where the EVS volume is created. It must be the same as the AZ planned for the workload.
storage	EVS volume capacity in the unit of Gi.
storageClassName	EVS disk type. Supported values: High I/O (SAS) and Ultra-high I/O (SSD)
driver	Storage driver. For EVS disks, set this parameter to huawei.com/fuxivol.
volumeID	Volume ID of the EVS disk. To obtain the volume ID, log in to the CCE console, choose Resource Management > Storage, click the PVC name in the EVS tab page, and copy the PVC ID on the PVC details page.
disk-mode	Device type of the EVS disk. The value is VBD or SCSI. For CCE clusters earlier than v1.11.7, you do not need to set this field. The value defaults to VBD. This field is mandatory for CCE clusters from v1.11.7 to v1.13 that use Linux x86. As the EVS volumes dynamically provisioned by a PVC are created from SCSI EVS disks, you are advised to choose SCSI when manually creating volumes (static PVs). Volumes in the VBD mode can still be used after cluster upgrades.
spec.claimRef.apiVersion	The value is fixed at v1.
spec.claimRef.kind	The value is fixed at PersistentVolumeClaim.
spec.claimRef.name	PVC name. The value is the same as the name of the PVC created in the next step.
spec.claimRef.namespace	Namespace of the PVC. The value is the same as the namespace of the PVC created in the next step.

Example YAML file for the PVC:

apiVersion: v1  
kind: PersistentVolumeClaim  
metadata:  
  annotations:  
    volume.beta.kubernetes.io/storage-class: sas
    volume.beta.kubernetes.io/storage-provisioner: flexvolume-huawei.com/fuxivol 
  labels: 
    failure-domain.beta.kubernetes.io/region: ap-southeast-1
    failure-domain.beta.kubernetes.io/zone: ap-southeast-1a     
  name: pvc-evs-example 
  namespace: default  
spec:  
  accessModes:  
  - ReadWriteOnce  
  resources:  
    requests:  
      storage: 10Gi
  volumeName: pv-evs-example

**Table 2** Key parameters
Parameter	Description
volume.beta.kubernetes.io/storage-class	Storage class, which must be the same as that of the existing PV.
volume.beta.kubernetes.io/storage-provisioner	The field must be set to flexvolume-huawei.com/fuxivol.
failure-domain.beta.kubernetes.io/region	Region where the cluster is located.
failure-domain.beta.kubernetes.io/zone	AZ where the EVS volume is created. It must be the same as the AZ planned for the workload.
storage	Requested capacity in the PVC, in Gi. The value must be the same as the storage size of the existing PV.
volumeName	Name of the PV.

Clusters from v1.11 to v1.11.7

Example YAML file for the PV:

apiVersion: v1 
kind: PersistentVolume 
metadata: 
  labels: 
    failure-domain.beta.kubernetes.io/region: ap-southeast-1
    failure-domain.beta.kubernetes.io/zone:  
  name: pv-evs-example 
spec: 
  accessModes: 
  - ReadWriteOnce
  capacity: 
    storage: 10Gi 
  flexVolume: 
    driver: huawei.com/fuxivol 
    fsType: ext4 
    options:
      fsType: ext4 
      volumeID: 0992dbda-6340-470e-a74e-4f0db288ed82 
  persistentVolumeReclaimPolicy: Delete 
  storageClassName: sas

**Table 3** Key parameters
Parameter	Description
failure-domain.beta.kubernetes.io/region	Region where the cluster is located.
failure-domain.beta.kubernetes.io/zone	AZ where the EVS volume is created. It must be the same as the AZ planned for the workload.
storage	EVS volume capacity in the unit of Gi.
storageClassName	EVS disk type. Supported values: High I/O (SAS) and Ultra-high I/O (SSD)
driver	Storage driver. For EVS disks, set this parameter to huawei.com/fuxivol.
volumeID	Volume ID of the EVS disk. To obtain the volume ID, log in to the CCE console, choose Resource Management > Storage, click the PVC name in the EVS tab page, and copy the PVC ID on the PVC details page.
disk-mode	Device type of the EVS disk. The value is VBD or SCSI. For CCE clusters earlier than v1.11.7, you do not need to set this field. The default value is VBD. This field is mandatory for CCE clusters from v1.11.7 to v1.13 that use Linux x86. As the EVS volumes dynamically provisioned by a PVC are created from SCSI EVS disks, you are advised to choose SCSI when manually creating volumes (static PVs). Volumes in the VBD mode can still be used after cluster upgrades.

Example YAML file for the PVC:

apiVersion: v1  
kind: PersistentVolumeClaim  
metadata:  
  annotations:  
    volume.beta.kubernetes.io/storage-class: sas
    volume.beta.kubernetes.io/storage-provisioner: flexvolume-huawei.com/fuxivol 
  labels: 
    failure-domain.beta.kubernetes.io/region: ap-southeast-1
    failure-domain.beta.kubernetes.io/zone: ap-southeast-1a     
  name: pvc-evs-example 
  namespace: default  
spec:  
  accessModes:  
  - ReadWriteOnce
  resources:  
    requests:  
      storage: 10Gi
  volumeName: pv-evs-example

**Table 4** Key parameters
Parameter	Description
volume.beta.kubernetes.io/storage-class	Storage class. The value can be sas or ssd. The value must be the same as that of the existing PV.
volume.beta.kubernetes.io/storage-provisioner	The field must be set to flexvolume-huawei.com/fuxivol.
failure-domain.beta.kubernetes.io/region	Region where the cluster is located.
failure-domain.beta.kubernetes.io/zone	AZ where the EVS volume is created. It must be the same as the AZ planned for the workload.
storage	Requested capacity in the PVC, in Gi. The value must be the same as the storage size of the existing PV.
volumeName	Name of the PV.

Clusters of v1.9

Example YAML file for the PV:

apiVersion: v1 
kind: PersistentVolume 
metadata: 
  labels: 
    failure-domain.beta.kubernetes.io/region: ap-southeast-1
    failure-domain.beta.kubernetes.io/zone:  
  name: pv-evs-example 
  namespace: default 
spec: 
  accessModes: 
  - ReadWriteOnce
  capacity: 
    storage: 10Gi 
  flexVolume: 
    driver: huawei.com/fuxivol 
    fsType: ext4 
    options: 
      fsType: ext4 
      kubernetes.io/namespace: default 
      volumeID: 0992dbda-6340-470e-a74e-4f0db288ed82 
  persistentVolumeReclaimPolicy: Delete 
  storageClassName: sas

**Table 5** Key parameters
Parameter	Description
failure-domain.beta.kubernetes.io/region	Region where the cluster is located.
failure-domain.beta.kubernetes.io/zone	AZ where the EVS volume is created. It must be the same as the AZ planned for the workload.
storage	EVS volume capacity in the unit of Gi.
storageClassName	EVS disk type. Supported values: High I/O (SAS) and Ultra-high I/O (SSD)
driver	Storage driver. For EVS disks, set this parameter to huawei.com/fuxivol.
volumeID	Volume ID of the EVS disk. To obtain the volume ID, log in to the CCE console, choose Resource Management > Storage, click the PVC name in the EVS tab page, and copy the PVC ID on the PVC details page.
disk-mode	Device type of the EVS disk. The value is VBD or SCSI. For CCE clusters earlier than v1.11.7, you do not need to set this field. The default value is VBD. This field is mandatory for CCE clusters from v1.11.7 to v1.13 that use Linux x86. As the EVS volumes dynamically provisioned by a PVC are created from SCSI EVS disks, you are advised to choose SCSI when manually creating volumes (static PVs). Volumes in the VBD mode can still be used after cluster upgrades.

Example YAML file for the PVC:

apiVersion: v1  
kind: PersistentVolumeClaim  
metadata:  
  annotations:  
    volume.beta.kubernetes.io/storage-class: sas
    volume.beta.kubernetes.io/storage-provisioner: flexvolume-huawei.com/fuxivol 
  labels: 
    failure-domain.beta.kubernetes.io/region: ap-southeast-1
    failure-domain.beta.kubernetes.io/zone: 
  name: pvc-evs-example 
  namespace: default  
spec:  
  accessModes:  
  - ReadWriteOnce 
  resources:  
    requests:  
      storage: 10Gi
  volumeName: pv-evs-example
  volumeNamespace: default

**Table 6** Key parameters
Parameter	Description
volume.beta.kubernetes.io/storage-class	Storage class, which must be the same as that of the existing PV.
volume.beta.kubernetes.io/storage-provisioner	The field must be set to flexvolume-huawei.com/fuxivol.
failure-domain.beta.kubernetes.io/region	Region where the cluster is located.
failure-domain.beta.kubernetes.io/zone	AZ where the EVS volume is created. It must be the same as the AZ planned for the workload.
storage	Requested capacity in the PVC, in Gi. The value must be the same as the storage size of the existing PV.
volumeName	Name of the PV.

Create a PV.
kubectl create -f pv-evs-example.yaml
Create a PVC.
kubectl create -f pvc-evs-example.yaml

After the operation is successful, choose Resource Management > Storage to view the created PVC. You can also view the EVS disk by name on the EVS console.

(Optional) Add the metadata associated with the cluster to ensure that the EVS disk associated with the mounted static PV is not deleted when the node or cluster is deleted.

If you skip this step in this example or when creating a static PV or PVC, ensure that the EVS disk associated with the static PV has been unbound from the node before you delete the node.

Obtain the tenant token. For details, see Obtaining a User Token.
Obtain the EVS access address EVS_ENDPOINT. For details, see Regions and Endpoints.

Add the metadata associated with the cluster to the EVS disk backing the static PV.

curl -X POST ${EVS_ENDPOINT}/v2/${project_id}/volumes/${volume_id}/metadata --insecure \
    -d '{"metadata":{"cluster_id": "${cluster_id}", "namespace": "${pvc_namespace}"}}' \
    -H 'Accept:application/json' -H 'Content-Type:application/json;charset=utf8' \
    -H 'X-Auth-Token:${TOKEN}'

**Table 7** Key parameters
Parameter	Description
EVS_ENDPOINT	EVS access address. Set this parameter to the value obtained in 6.b.
project_id	Project ID.
volume_id	ID of the associated EVS disk. Set this parameter to volume_id of the static PV to be created. You can also log in to the EVS console, click the name of the EVS disk to be imported, and obtain the ID from Summary on the disk details page, as shown in Figure 1.
cluster_id	ID of the cluster where the EVS PV is to be created. On the CCE console, choose Resource Management > Clusters. Click the name of the cluster to be associated. On the cluster details page, obtain the cluster ID, as shown in Figure 2.
pvc_namespace	Namespace where the PVC is to be bound.
TOKEN	User token. Set this parameter to the value obtained in 6.a.

Figure 1 Obtaining the disk ID

Figure 2 Obtaining the cluster ID

For example, run the following commands:

curl -X POST https://evs.ap-southeast-1.myhuaweicloud.com:443/v2/060576866680d5762f52c0150e726aa7/volumes/69c9619d-174c-4c41-837e-31b892604e14/metadata --insecure \
    -d '{"metadata":{"cluster_id": "71e8277e-80c7-11ea-925c-0255ac100442", "namespace": "default"}}' \
    -H 'Accept:application/json' -H 'Content-Type:application/json;charset=utf8' \
    -H 'X-Auth-Token:MIIPe******IsIm1ldG

After the request is executed, run the following commands to check whether the EVS disk has been associated with the metadata of the cluster:

curl -X GET ${EVS_ENDPOINT}/v2/${project_id}/volumes/${volume_id}/metadata --insecure \
    -H 'X-Auth-Token:${TOKEN}'

For example, run the following commands:

curl -X GET https://evs.ap-southeast-1.myhuaweicloud.com/v2/060576866680d5762f52c0150e726aa7/volumes/69c9619d-174c-4c41-837e-31b892604e14/metadata --insecure \
    -H 'X-Auth-Token:MIIPeAYJ***9t1c31ASaQ=='

The command output displays the current metadata of the EVS disk.

{
    "metadata": {
        "namespace": "default",
        "cluster_id": "71e8277e-80c7-11ea-925c-0255ac100442",
        "hw:passthrough": "true"
    }
}

Parent topic: Using EVS Disks as Storage Volumes

Previous topic: (kubectl) Automatically Creating an EVS Disk

Next topic: (kubectl) Creating a Pod Mounted with an EVS Volume