Setting up File Storage for Classic

File storage types

Choose the type of File Storage for Classic that you want to provision.

Bronze, silver, and gold storage classes

These storage classes provision Endurance storage. Endurance storage lets you choose the size of the storage in gigabytes at predefined IOPS tiers.

Custom storage class

This storage class provisions Performance storage. With performance storage, you have more control over the size of the storage and the IOPS.

IOPS

Choose the size and IOPS for your File Storage for Classic. The size and the number of IOPS define the total number of IOPS (input/ output operations per second) that serves as an indicator for how fast your storage is. The more total IOPS your storage has, the faster it processes read and write operations.

Bronze, silver, and gold storage classes

These storage classes come with a fixed number of IOPS per gigabyte and are provisioned on SSD hard disks. The total number of IOPS depends on the size of the storage that you choose. You can select any whole number of gigabyte within the allowed size range, such as 20 Gi, 256 Gi, or 11854 Gi. To determine the total number of IOPS, you must multiply the IOPS with the selected size. For example, if you select a 1000Gi File Storage for Classic size in the silver storage class that comes with 4 IOPS per GB, your storage has a total of 4000 IOPS.

Custom storage class

When you choose this storage class, you have more control over the size and IOPS that you want. For the size, you can select any whole number of gigabyte within the allowed size range. The size that you choose determines the IOPS range that is available to you. You can choose an IOPS that is a multiple of 100 that is in the specified range. The IOPS that you choose is static and does not scale with the size of the storage. For example, if you choose 40Gi with 100 IOPS, your total IOPS remains 100.

The IOPS to gigabyte ratio also determines the type of hard disk that is provisioned for you. For example, if you have 500Gi at 100 IOPS, your IOPS to gigabyte ratio is 0.2. Storage with a ratio of less than or equal to 0.3 is provisioned on SATA hard disks. If your ratio is greater than 0.3, then your storage is provisioned on SSD hard disks.

Reclaim policy

Choose if you want to keep your data after the cluster or the persistent volume claim (PVC) is deleted.

• If you want to keep your data, then choose a retain storage class. When you delete the PVC, only the PVC is deleted. The PV, the physical storage device in your IBM Cloud infrastructure account, and your data still exist. To reclaim the storage and use it in your cluster again, you must remove the PV and follow the steps for using existing File Storage for Classic.
• If you want the PV, the data, and your physical File Storage for Classic device to be deleted when you delete the PVC, choose a storage class without retain.

Billing type

Choose hourly or monthly. Review the pricing for more information.

By default, all File Storage for Classic devices are provisioned with an hourly billing type.

If you choose a monthly billing type, when you remove the persistent storage, you still pay the monthly charge for it, even if you used it only for a short amount of time.

Preparing your existing storage

Before you can start to mount your existing storage to an app, you must retrieve all necessary information for your PV and prepare the storage to be accessible in your cluster.

For storage that was provisioned with a retain storage class.

If you provisioned storage with a retain storage class and you remove the PVC, the PV and the physical storage device are not automatically removed. To reuse the storage in your cluster, you must remove the remaining PV first.

To use existing storage in a different cluster than the one where you provisioned it, follow the steps for storage that was created outside of the cluster to add the storage to the subnet of your worker node.

1. List existing PVs.

   kubectl get pv
   

   Look for the PV that belongs to your persistent storage. The PV is in a released state.

2. Get the details of the PV.

   kubectl describe pv <pv_name>
   

3. Note the CapacityGb, storageClass, failure-domain.beta.kubernetes.io/region, failure-domain.beta.kubernetes.io/zone, server, and path.

4. Remove the PV.

   kubectl delete pv <pv_name>
   

5. Verify that the PV is removed.

   kubectl get pv
   

For persistent storage that was provisioned outside the cluster

If you want to use existing storage that you provisioned earlier, but never used in your cluster before, you must make the storage available in the same subnet as your worker nodes.

1. From the IBM Cloud infrastructure portal, click Storage.
2. Click File Storage for Classic and from the Actions menu, select Authorize Host.
3. Select Subnets.
4. From the drop-down list, select the private VLAN subnet that your worker node is connected to. To find the subnet of your worker node, run ibmcloud ks worker ls --cluster <cluster_name> and compare the Private IP of your worker node with the subnet that you found in the drop-down list.
5. Click Submit.
6. Click the name of the File Storage for Classic.
7. Note the Mount Point, the size, and the Location field. The Mount Point field is displayed as <nfs_server>:<file_storage_path>.

Creating a persistent volume and a persistent volume claim

1. Create a storage configuration file for your PV. Include the values that you retrieved earlier.

   apiVersion: v1
   kind: PersistentVolume
   metadata:
    name: mypv
    labels:
       failure-domain.beta.kubernetes.io/region: <region>
       failure-domain.beta.kubernetes.io/zone: <zone>
   spec:
    capacity:
      storage: "<size>"
    accessModes:
      - ReadWriteMany
    nfs:
      server: "<nfs_server>"
      path: "<file_storage_path>"
   

   name

   Enter the name of the PV object to create.

   labels

   Enter the region and the zone that you retrieved earlier. You must have at least one worker node in the same region and zone.

   storage

   Enter the storage size of the existing NFS file share that you retrieved earlier. The storage size must be written in gigabytes, for example, 20Gi (20 GB) or 1000Gi (1 TB), and the size must match the size of the existing file share.

   accessMode

   Specify one of the following options.

   • ReadWriteMany: The PVC can be mounted by multiple pods. All pods can read from and write to the volume.
   • ReadOnlyMany: The PVC can be mounted by multiple pods. All pods have read-only access.
   • ReadWriteOnce: The PVC can be mounted by one pod only. This pod can read from and write to the volume.

   server

   Enter the NFS file share server ID that you retrieved earlier.

   path

   Enter the path to the NFS file share that you retrieved earlier.

2. Create the PV in your cluster.

   kubectl apply -f mypv.yaml
   

3. Verify that the PV is created.

   kubectl get pv
   

4. Create another configuration file to create your PVC. In order for the PVC to match the PV that you created earlier, you must choose the same value for storage and accessMode. The storage-class field must be an empty string. If any of these fields don't match the PV, then a new PV, and a new physical storage instance is dynamically provisioned.

   kind: PersistentVolumeClaim
   apiVersion: v1
   metadata:
     name: mypvc
   spec:
     accessModes:
      - ReadWriteMany
     resources:
       requests:
         storage: "<size>"
     storageClassName: ""
   

5. Create your PVC.

   kubectl apply -f mypvc.yaml
   

6. Verify that your PVC is created and bound to the PV.

   kubectl describe pvc mypvc
   

   Example output

   Name: mypvc
   Namespace: default
   StorageClass:    ""
   Status: Bound
   Volume: pvc-0d787071-3a67-11e7-aafc-eef80dd2dea2
   Labels: <none>
   Capacity: 20Gi
   Access Modes: RWX
   Events:
       FirstSeen LastSeen Count From        SubObjectPath Type Reason Message
       --------- -------- ----- ----        ------------- -------- ------ -------
       3m 3m 1 {ibm.io/ibmc-file 31898035-3011-11e7-a6a4-7a08779efd33 } Normal Provisioning External provisioner is provisioning volume for claim "default/my-persistent-volume-claim"
       3m 1m     10 {persistentvolume-controller } Normal ExternalProvisioning can't find provisioner "ibm.io/ibmc-file", expecting that a volume for the claim is provisioned either manually or via external software
       1m 1m 1 {ibm.io/ibmc-file 31898035-3011-11e7-a6a4-7a08779efd33 } Normal ProvisioningSucceeded    Successfully provisioned volume pvc-0d787071-3a67-11e7-aafc-eef80dd2dea2
   

You successfully created a PV and bound it to a PVC. Cluster users can now mount the PVC to their deployments and start reading from and writing to the PV object.

Creating the PVC when you create a stateful set by using dynamic provisioning

Use this option if you want to automatically create the PVC when you create the stateful set.

Before you begin: Log in to your account. If applicable, target the appropriate resource group. Set the context for your cluster.

1. Verify that all existing stateful sets in your cluster are fully deployed. If a stateful set is still being deployed, you can't start creating your stateful set. You must wait until all stateful sets in your cluster are fully deployed to avoid unexpected results. List existing stateful sets in your cluster.

   kubectl get statefulset --all-namespaces
   

   Example output

   NAME              DESIRED   CURRENT   AGE
   mystatefulset     3         3         6s
   

2. View the Pods Status of each stateful set to ensure that the deployment of the stateful set is finished.

   kubectl describe statefulset <statefulset_name>
   

   Example output

   Name:               nginx
   Namespace:          default
   CreationTimestamp:  Fri, 05 Oct 2022 13:22:41 -0400
   Selector:           app=nginx,billingType=hourly,region=us-south,zone=dal10
   Labels:             app=nginx
   billingType=hourly
   region=us-south
   zone=dal10
   Annotations:        kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"apps/v1","kind":"StatefulSet","metadata":{"annotations":{},"name":"nginx","namespace":"default"},"spec":{"podManagementPolicy":"Par..."
   Replicas:           3 desired | 3 total
   Pods Status:        0 Running / 3 Waiting / 0 Succeeded / 0 Failed
   Pod Template:
   Labels:  app=nginx
   billingType=hourly
   region=us-south
   zone=dal10
   

   A stateful set is fully deployed when the number of replicas that you find in the Replicas section of your CLI output equals the number of Running pods in the Pods Status section. If a stateful set is not fully deployed yet, wait until the deployment is finished before you proceed.

3. Create a configuration file for your stateful set and the service that you use to expose the stateful set.

   Example stateful set that specifies a zone. The following example shows how to deploy NGINX as a stateful set with 3 replicas. For each replica, a 20 gigabyte File Storage for Classic device is provisioned based on the specifications in the ibmc-file-retain-bronze storage class. All storage devices are provisioned in the dal10 zone. Because File Storage for Classic can't be accessed from other zones, all replicas of the stateful set are also deployed onto worker nodes that are located in dal10.

   apiVersion: v1
   kind: Service
   metadata:
     name: nginx
     labels:
       app: nginx
   spec:
     ports:
     - port: 80
       name: web
     clusterIP: None
     selector:
       app: nginx
   ---
   apiVersion: apps/v1
   kind: StatefulSet
   metadata:
    name: nginx
   spec:
     serviceName: "nginx"
     replicas: 3
     podManagementPolicy: Parallel
     selector:
       matchLabels:
         app: nginx
         billingType: "hourly"
         region: "us-south"
         zone: "dal10"
     template:
       metadata:
         labels:
           app: nginx
           billingType: "hourly"
           region: "us-south"
           zone: "dal10"
       spec:
         containers:
         - name: nginx
           image: k8s.gcr.io/nginx-slim:0.8
           ports:
           - containerPort: 80
             name: web
           volumeMounts:
           - name: myvol
             mountPath: /usr/share/nginx/html
       volumeClaimTemplates:
       - metadata:
          name: myvol
         spec:
          accessModes:
          - ReadWriteOnce
          resources:
            requests:
              storage: 20Gi
              iops: "300" #required only for performance storage
          storageClassName: ibmc-file-retain-bronze
   

   Example stateful set with anti-affinity rule and delayed File Storage for Classic creation. The following example shows how to deploy NGINX as a stateful set with 3 replicas. The stateful set does not specify the region and zone where the File Storage for Classic is created. Instead, the stateful set uses an anti-affinity rule to ensure that the pods are spread across worker nodes and zones. Worker node anti-affinity is achieved by defining the app: nginx label. This label instructs the Kubernetes scheduler to not schedule a pod on a worker node if a pod with the same label already runs on this worker node. The topologykey: failure-domain.beta.kubernetes.io/zone label restricts this anti-affinity rule even more and prevents the pod to be scheduled on a worker node that is located in the same zone as a worker node that already runs a pod with the app: nginx label. For each stateful set pod, two PVCs are created as defined in the volumeClaimTemplates section, but the creation of the File Storage for Classic instances is delayed until a stateful set pod that uses the storage is scheduled. This setup is referred to as topology-aware volume scheduling.

   apiVersion: storage.k8s.io/v1
   kind: StorageClass
   metadata:
     name: ibmc-file-bronze-delayed
   parameters:
     billingType: hourly
     classVersion: "2"
     iopsPerGB: "2"
     sizeRange: '[20-12000]Gi'
     type: Endurance
   provisioner: ibm.io/ibmc-file
   reclaimPolicy: Delete
   volumeBindingMode: WaitForFirstConsumer
   ---
   apiVersion: v1
   kind: Service
   metadata:
     name: nginx
     labels:
       app: nginx
   spec:
     ports:
     - port: 80
       name: web
     clusterIP: None
     selector:
       app: nginx
   ---
   apiVersion: apps/v1
   kind: StatefulSet
   metadata:
     name: web
   spec:
     serviceName: "nginx"
     replicas: 3
     podManagementPolicy: "Parallel"
     selector:
       matchLabels:
         app: nginx
     template:
       metadata:
         labels:
           app: nginx
       spec:
         affinity:
           podAntiAffinity:
             preferredDuringSchedulingIgnoredDuringExecution:
             - weight: 100
               podAffinityTerm:
                 labelSelector:
                   matchExpressions:
                   - key: app
                     operator: In
                     values:
                     - nginx
                 topologyKey: failure-domain.beta.kubernetes.io/zone
         containers:
         - name: nginx
           image: k8s.gcr.io/nginx-slim:0.8
           ports:
           - containerPort: 80
             name: web
           volumeMounts:
           - name: myvol1
             mountPath: /usr/share/nginx/html
           - name: myvol2
             mountPath: /tmp1
     volumeClaimTemplates:
     - metadata:
         name: myvol1
       spec:
         accessModes:
         - ReadWriteMany # access mode
         resources:
           requests:
             storage: 20Gi
         storageClassName: ibmc-file-bronze-delayed
     - metadata:
         name: myvol2
       spec:
         accessModes:
         - ReadWriteMany # access mode
         resources:
           requests:
             storage: 20Gi
         storageClassName: ibmc-file-bronze-delayed
   

   name

   In the metadata, enter a name for your stateful set. The name that you enter is used to create the name for your PVC in the format: <volume_name>-<statefulset_name>-<replica_number>.

   serviceName

   In the spec section, enter the name of the service that you want to use to expose your stateful set.

   replicas

   Enter the number of replicas for your stateful set.

   podManagementPolicy

   Enter the pod management policy that you want to use for your stateful set. Choose between the following options.

   • OrderedReady: With this option, stateful set replicas are deployed one after another. For example, if you specified 3 replicas, then Kubernetes creates the PVC for your first replica, waits until the PVC is bound, deploys the stateful set replica, and mounts the PVC to the replica. After the deployment is finished, the second replica is deployed. For more information about this option, see OrderedReady Pod Management.
   • Parallel: With this option, the deployment of all stateful set replicas is started at the same time. If your app supports parallel deployment of replicas, then use this option to save deployment time for your PVCs and stateful set replicas.

   matchLabels

   In the spec selector section, enter all labels that you want to include in your stateful set and your PVC. Labels that you include in the volumeClaimTemplates of your stateful set are not recognized by Kubernetes. Review the following sample labels.

   • region and zone: If you want all your stateful set replicas and PVCs to be created in one specific zone, add both labels. You can also specify the zone and region in the storage class that you use. If you don't specify a zone and region and you have a multizone cluster, the zone in which your storage is provisioned is selected on a round-robin basis to balance volume requests evenly across all zones.
   • billingType: Enter the billing type that you want to use for your PVCs. Choose between hourly or monthly. If you don't specify this label, all PVCs are created with an hourly billing type.

   labels

   In the spec template metadata section, enter the same labels that you added to the spec.selector.matchLabels section.

   affinity

   In the spec template spec affinity section, specify your anti-affinity rule to ensure that your stateful set pods are distributed across worker nodes and zones. The example shows an anti-affinity rule where the stateful set pod prefers not to be scheduled on a worker node where a pod runs that has the app: nginx label. The topologykey: failure-domain.beta.kubernetes.io/zone restricts this anti-affinity rule even more and prevents the pod to be scheduled on a worker node if the worker node is in the same zone as a pod that has the app: nginx label. By using this anti-affinity rule, you can achieve anti-affinity across worker nodes and zones.

   name

   In the spec volume claim templates metadata section, enter a name for your volume. Use the same name that you defined in the spec.containers.volumeMount.name section. The name that you enter here is used to create the name for your PVC in the format: <volume_name>-<statefulset_name>-<replica_number>.

   storage

   In the spec volume claim templates spec resources requests section, enter the size of the File Storage for Classic in gigabytes (Gi).

   iops

   In the spec volume claim templates spec resources requests section, if you want to provision performance storage, enter the number of IOPS. If you use an endurance storage class and specify a number of IOPS, the number of IOPS is ignored. Instead, the IOPS that is specified in your storage class is used.

   storageClassName

   In the spec volume claim templates spec section, enter the storage class that you want to use. To list existing storage classes, run kubectl get sc | grep file. If you don't specify a storage class, the PVC is created with the default storage class that is set in your cluster. Make sure that the default storage class uses the ibm.io/ibmc-file provisioner so that your stateful set is provisioned with File Storage for Classic.

4. Create your stateful set.

   kubectl apply -f statefulset.yaml
   

5. Wait for your stateful set to be deployed.

   kubectl describe statefulset <statefulset_name>
   

To see the current status of your PVCs, run kubectl get pvc. The name of your PVC is formatted as <volume_name>-<statefulset_name>-<replica_number>.

Static provisioning: Using an existing PVC with your stateful set

You can pre-provision your PVCs before creating your stateful set or use existing PVCs with your stateful set.

When you dynamically provision your PVCs when creating the stateful set, the name of the PVC is assigned based on the values that you used in the stateful set YAML file. In order for the stateful set to use existing PVCs, the name of your PVCs must match the name that would automatically be created when using dynamic provisioning.

Before you begin: Log in to your account. If applicable, target the appropriate resource group. Set the context for your cluster.

1. If you want to pre-provision your PVC before you create the stateful set, follow steps 1-3 in Adding File Storage for Classic to apps to create a PVC for each stateful set replica. Make sure that you create your PVC with a name that follows the following format: <volume_name>-<statefulset_name>-<replica_number>.

   <volume_name>

   Use the name that you want to specify in the spec.volumeClaimTemplates.metadata.name section of your stateful set, such as nginxvol.

   <statefulset_name>

   Use the name that you want to specify in the metadata.name section of your stateful set, such as nginx_statefulset.

   <replica_number>

   Enter the number of your replica starting with 0.

   For example, if you must create 3 stateful set replicas, create 3 PVCs with the following names: nginxvol-nginx_statefulset-0, nginxvol-nginx_statefulset-1, and nginxvol-nginx_statefulset-2.

   Looking to create a PVC and PV for an existing File Storage for Classic instance? Create your PVC and PV by using static provisioning.

2. Follow the steps in Dynamic provisioning: Creating the PVC when you create a stateful set to create your stateful set. The name of your PVC follows the format <volume_name>-<statefulset_name>-<replica_number>. Make sure to use the following values from your PVC name in the stateful set specification.

   spec.volumeClaimTemplates.metadata.name

   Enter the <volume_name> of your PVC name.

   metadata.name

   Enter the <statefulset_name> of your PVC name.

   spec.replicas

   Enter the number of replicas that you want to create for your stateful set. The number of replicas must equal the number of PVCs that you created earlier.

   If your PVCs are in different zones, don't include a region or zone label in your stateful set.

3. Verify that the PVCs are used in your stateful set replica pods by listing the pods in your cluster and identifying the pods that belong to your stateful set.

   kubectl get pods
   

4. Verify that your existing PVC is mounted to your stateful set replica. Review the ClaimName in the Volumes section of your CLI output.

   kubectl describe pod <pod_name>
   

   Example output

   Name:           nginx-0
   Namespace:      default
   Node:           10.xxx.xx.xxx/10.xxx.xx.xxx
   Start Time:     Fri, 05 Oct 2022 13:24:59 -0400
   ...
   Volumes:
   myvol:
       Type:       PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
       ClaimName:  myvol-nginx-0
   ...
   

Creating a customized storage class with a specific NFS version

1. Create a customized storage class with the NFS version that you want to provision.

2. Create the storage class in your cluster.

   kubectl apply -f nfsversion_storageclass.yaml
   

3. Verify that the customized storage class was created.

   kubectl get sc
   

4. Provision File Storage for Classic with your customized storage class.

Changing your existing PV to use a different NFS version

1. Get the PV of the File Storage for Classic where you want to change the NFS version and note the name of the PV.

   kubectl get pv
   

2. Add an annotation to your PV. Replace <version_number> with the NFS version that you want to use. For example, to change to NFS version 3.0, enter 3.

   kubectl patch pv <pv_name> -p '{"metadata": {"annotations":{"volume.beta.kubernetes.io/mount-options":"vers=<version_number>"}}}'
   

3. Delete the pod that uses the File Storage for Classic and re-create the pod.

   1. Save the pod YAML to your local machine.

      kubect get pod <pod_name> -o yaml > <filepath/pod.yaml>
      

   2. Delete the pod.

      kubectl deleted pod <pod_name>
      

   3. Re-create the pod.

      kubectl apply -f pod.yaml
      

4. Wait for the pod to deploy. The pod is fully deployed when the status changes to Running.

   kubectl get pods
   

5. Log in to your pod.

   kubectl exec -it <pod_name> sh
   

6. Verify that the File Storage for Classic was mounted with the NFS version that you specified earlier.

   mount | grep "nfs" | awk -F" |," '{ print $5, $8 }'
   

   Example output

   nfs vers=3.0
   

Setting up periodic snapshots

You can set up periodic snapshots for your File Storage for Classic, which is a read-only image that captures the state of the instance at a point in time. To store the snapshot, you must request snapshot space on your File Storage for Classic. Snapshots are stored on the existing storage instance within the same zone. You can restore data from a snapshot if a user accidentally removes important data from the volume.

Complete the following steps to create a snapshot for your volume.

1. Log in to your account. If applicable, target the appropriate resource group. Set the context for your cluster.

2. Log in to the ibmcloud sl CLI.

   ibmcloud sl init
   

3. List existing PVs in your cluster.

   kubectl get pv
   

4. Get the details for the PV for which you want to create snapshot space and note the volume ID, the size and the IOPS. The volume ID, the size and the IOPS can be found in the Labels section of your CLI output.

   kubectl describe pv <pv_name>
   

5. Create the snapshot size for your existing volume with the parameters that you retrieved in the previous step.

   ibmcloud sl file snapshot-order <volume_ID> --size <size> --tier <iops>
   

6. Wait for the snapshot size to create. The snapshot size is successfully provisioned when the **Snapshot Size (GB)**in your CLI output changes from 0 to the size that you ordered.

   ibmcloud sl file volume-detail <volume_ID>
   

7. Create the snapshot for your volume and note the ID of the snapshot that is created for you.

   ibmcloud sl file snapshot-create <volume_ID>
   

8. Verify that the snapshot is created successfully.

   ibmcloud sl file snapshot-list <volume_ID>
   

9. Set the snapshot schedule. For more information on the options available for your snapshot schedule, refer to the CLI documentation.

   ibmcloud sl block snapshot-enable VOLUME_ID <OPTIONS>
   

10. To restore data from a snapshot to an existing volume, run the following command.

    ibmcloud sl file snapshot-restore <volume_ID> <snapshot_ID>
    

Replicating snapshots to another zone

To protect your data from a zone failure, you can replicate snapshots to a File Storage for Classic instance that is set up in another zone.

Data can be replicated from the primary storage to the backup storage only. You can't mount a replicated File Storage for Classic instance to a cluster. When your primary storage fails, you can manually set your replicated backup storage to be the primary one. Then, you can mount it to your cluster. After your primary storage is restored, you can restore the data from the backup storage.

Duplicating storage

You can duplicate your File Storage for Classic instance in the same zone as the original storage instance.

A duplicate has the same data as the original storage instance at the point in time that you create the duplicate. Unlike replicas, use the duplicate as an independent storage instance from the original. To duplicate, first set up snapshots for the volume.

Backing up data to IBM Cloud® Object Storage

You can use the ibm-backup-restore Helm chart to spin up a backup and restore pod in your cluster.

This pod contains a script to run a one-time or periodic backup for any persistent volume claim (PVC) in your cluster. Data is stored in your IBM Cloud® Object Storage instance that you set up in a zone.

To make your data even more highly available and protect your app from a zone failure, set up a second IBM Cloud® Object Storage instance and replicate data across zones. If you need to restore data from your IBM Cloud® Object Storage instance, use the restore script that is provided with the Helm chart.

Copying data to and from pods and containers

You can use the kubectl cp command to copy files and directories to and from pods or specific containers in your cluster.

Before you begin: Log in to your account. If applicable, target the appropriate resource group. Set the context for your cluster. If you don't specify a container with -c, the command uses the first available container in the pod.

Copy data from your local machine to a pod in your cluster.

kubectl cp <local_filepath>/<filename> <namespace>/<pod>:<pod_filepath>

Copy data from a pod in your cluster to your local machine.

kubectl cp <namespace>/<pod>:<pod_filepath>/<filename></var> <local_filepath>/<filename>

Copy data from your local machine to a specific container that runs in a pod in your cluster.

kubectl cp <local_filepath>/<filename> <namespace>/<pod>:<pod_filepath> -c CONTAINER

Creating topology-aware storage

To use File Storage for Classic in a multizone cluster, your pod must be scheduled in the same zone as your File Storage for Classic instance so that you can read and write to the volume. Before topology-aware volume scheduling was introduced by Kubernetes, the dynamic provisioning of your storage automatically created the File Storage for Classic instance when a PVC was created. Then, when you created your pod, the Kubernetes scheduler tried to deploy the pod to a worker node in the same data center as your File Storage for Classic instance.

Creating the File Storage for Classic instance without knowing the constraints of the pod can lead to unwanted results. For example, your pod might not be able to be scheduled to the same worker node as your storage because the worker node has insufficient resources or the worker node is tainted and does not allow the pod to be scheduled. With topology-aware volume scheduling, the File Storage for Classic instance is delayed until the first pod that uses the storage is created.

The following examples show how to create storage classes that delay the creation of the File Storage for Classic instance until the first pod that uses this storage is ready to be scheduled. To delay the creation, you must include the volumeBindingMode: WaitForFirstConsumer option. If you don't include this option, the volumeBindingMode is automatically set to Immediate and the File Storage for Classic instance is created when you create the PVC.

Example for Endurance File Storage for Classic.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
    name: ibmc-file-bronze-delayed
parameters:
  billingType: hourly
  classVersion: "2"
  iopsPerGB: "2"
  sizeRange: '[20-12000]Gi'
  type: Endurance
  provisioner: ibm.io/ibmc-file
  reclaimPolicy: Delete
  volumeBindingMode: WaitForFirstConsumer

Example for Performance File Storage for Classic.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
 name: ibmc-file-performance-storageclass
 labels:
   kubernetes.io/cluster-service: "true"
provisioner: ibm.io/ibmc-file
parameters:
 billingType: "hourly"
 classVersion: "2"
 sizeIOPSRange: |-
   "[20-39]Gi:[100-1000]"
   "[40-79]Gi:[100-2000]"
   "[80-99]Gi:[100-4000]"
   "[100-499]Gi:[100-6000]"
   "[500-999]Gi:[100-10000]"
   "[1000-1999]Gi:[100-20000]"
   "[2000-2999]Gi:[200-40000]"
   "[3000-3999]Gi:[200-48000]"
   "[4000-7999]Gi:[300-48000]"
   "[8000-9999]Gi:[500-48000]"
   "[10000-12000]Gi:[1000-48000]"
 type: "Performance"
reclaimPolicy: Delete
volumeBindingMode: WaitForFirstConsumer

Specifying the zone for multizone clusters

If you want to create your File Storage for Classic in a specific zone, you can specify the zone and region in a customized storage class.

Use the customized storage class if you want to statically provision File Storage for Classic in a specific zone. In all other cases, specify the zone directly in your PVC.

When you create the customized storage class, specify the same region and zone that your cluster and worker nodes are in. To get the region of your cluster, run ibmcloud ks cluster get --cluster <cluster_name_or_ID> and look for the region prefix in the Master URL, such as eu-de in https://c2.eu-de.containers.cloud.ibm.com:11111. To get the zone of your worker node, run ibmcloud ks worker ls --cluster <cluster_name_or_ID>.

Example for Endurance File Storage for Classic.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
    name: ibmc-file-silver-mycustom-storageclass
labels:
  kubernetes.io/cluster-service: "true"
provisioner: ibm.io/ibmc-file
parameters:
  zone: "dal12"
  region: "us-south"
  type: "Endurance"
  iopsPerGB: "4"
  sizeRange: "[20-12000]Gi"
  reclaimPolicy: "Delete"
  classVersion: "2"
reclaimPolicy: Delete
volumeBindingMode: Immediate

Example for Performance File Storage for Classic.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
 name: ibmc-file-performance-storageclass
 labels:
   kubernetes.io/cluster-service: "true"
provisioner: ibm.io/ibmc-file
parameters:
  zone: "dal12"
  region: "us-south"
  billingType: "hourly"
  classVersion: "2"
  sizeIOPSRange: |-
   "[20-39]Gi:[100-1000]"
   "[40-79]Gi:[100-2000]"
   "[80-99]Gi:[100-4000]"
   "[100-499]Gi:[100-6000]"
   "[500-999]Gi:[100-10000]"
   "[1000-1999]Gi:[100-20000]"
   "[2000-2999]Gi:[200-40000]"
   "[3000-3999]Gi:[200-48000]"
   "[4000-7999]Gi:[300-48000]"
   "[8000-9999]Gi:[500-48000]"
   "[10000-12000]Gi:[1000-48000]"
  type: "Performance"
reclaimPolicy: Delete
volumeBindingMode: Immediate

Changing the default NFS version

The following customized storage class lets you define the NFS version that you want to provision. For example, to provision NFS version 3.0, replace <nfs_version> with 3.0.

Example for Endurance File Storage for Classic.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
    name: ibmc-file-mount
labels:
  kubernetes.io/cluster-service: "true"
provisioner: ibm.io/ibmc-file
parameters:
  type: "Endurance"
  iopsPerGB: "2"
  sizeRange: "[1-12000]Gi"
  reclaimPolicy: "Delete"
  classVersion: "2"
  mountOptions: nfsvers=<nfs_version>

Example for Performance File Storage for Classic.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
    name: ibmc-file-mount
labels:
  kubernetes.io/cluster-service: "true"
provisioner: ibm.io/ibmc-file
parameters:
  type: "Performance"
  classVersion: "2"
  sizeIOPSRange: |-
    "[20-39]Gi:[100-1000]"
    "[40-79]Gi:[100-2000]"
    "[80-99]Gi:[100-4000]"
    "[100-499]Gi:[100-6000]"
    "[500-999]Gi:[100-10000]"
    "[1000-1999]Gi:[100-20000]"
    "[2000-2999]Gi:[200-40000]"
    "[3000-3999]Gi:[200-48000]"
    "[4000-7999]Gi:[300-48000]"
    "[8000-9999]Gi:[500-48000]"
    "[10000-12000]Gi:[1000-48000]"
  mountOptions: nfsvers=<nfs_version>

Understanding your storage removal options

Removing persistent storage from your IBM Cloud account varies depending on how you provisioned the storage and what components you already removed.

Is my persistent storage deleted when I delete my cluster?

During cluster deletion, you have the option to remove your persistent storage. However, depending on how your storage was provisioned, the removal of your storage might not include all storage components. If you dynamically provisioned storage with a storage class that sets reclaimPolicy: Delete, your PVC, PV, and the storage instance are automatically deleted when you delete the cluster. For storage that was statically provisioned or storage that you provisioned with a storage class that sets reclaimPolicy: Retain, the PVC and the PV are removed when you delete the cluster, but your storage instance and your data remain. You are still charged for your storage instance. Also, if you deleted your cluster in an unhealthy state, the storage might still exist even if you chose to remove it.

How do I delete the storage when I want to keep my cluster?

When you dynamically provisioned the storage with a storage class that sets reclaimPolicy: Delete, you can remove the PVC to start the deletion process of your persistent storage. Your PVC, PV, and storage instance are automatically removed. For storage that was statically provisioned or storage that you provisioned with a storage class that sets reclaimPolicy: Retain, you must manually remove the PVC, PV, and the storage instance to avoid further charges.

How does the billing stop after I delete my storage?

Depending on what storage components you delete and when, the billing cycle might not stop immediately. If you delete the PVC and PV, but not the storage instance in your IBM Cloud account, that instance still exists and you are charged for it.

If you delete the PVC, PV, and the storage instance, the billing cycle stops depending on the billingType that you chose when you provisioned your storage and how you chose to delete the storage.

• When you manually cancel the persistent storage instance from the IBM Cloud console or the CLI, billing stops as follows:

  • Hourly storage: Billing stops immediately. After your storage is canceled, you might still see your storage instance in the console for up to 72 hours.
  • Monthly storage: You can choose between immediate cancellation or cancellation on the anniversary date. In both cases, you are billed until the end of the current billing cycle, and billing stops for the next billing cycle. After your storage is canceled, you might still see your storage instance in the console or the CLI for up to 72 hours.
  • Immediate cancellation: Choose this option to immediately remove your storage. Neither you nor your users can use the storage anymore or recover the data.
  • Anniversary date: Choose this option to cancel your storage on the next anniversary date. Your storage instances remain active until the next anniversary date and you can continue to use them until this date, such as to give your team time to make backups of your data.

• When you dynamically provisioned the storage with a storage class that sets reclaimPolicy: Delete and you choose to remove the PVC, the PV and the storage instance are immediately removed. For hourly billed storage, billing stops immediately. For monthly billed storage, you are still charged for the remainder of the month. After your storage is removed and billing stops, you might still see your storage instance in the console or the CLI for up to 72 hours.

What do I need to be aware of before I delete persistent storage?

When you clean up persistent storage, you delete all the data that is stored in it. If you need a copy of the data, make a backup.

I deleted my storage instance. Why can I still see my instance?

After you remove persistent storage, it can take up to 72 hours for the removal to be fully processed and for the storage to disappear from your IBM Cloud console or CLI.

Cleaning up persistent storage

Remove the PVC, PV, and the storage instance from your IBM Cloud account to avoid further charges for your persistent storage.

Before you begin:

• Make sure that you backed up any data that you want to keep.
• Log in to your account. If applicable, target the appropriate resource group. Set the context for your cluster.

To clean up persistent data:

1. List the PVCs in your cluster and note the NAME of the PVC, the STORAGECLASS, and the name of the PV that is bound to the PVC and shown as VOLUME.

   kubectl get pvc
   

   Example output

   NAME                  STATUS    VOLUME                                     CAPACITY   ACCESSMODES   STORAGECLASS            AGE
   claim1   Bound     pvc-06886b77-102b-11e8-968a-f6612bb731fb   20Gi       RWO           class       78d
   claim2     Bound     pvc-457a2b96-fafc-11e7-8ff9-b6c8f770356c   4Gi        RWX           class 105d
   claim3      Bound     pvc-1efef0ba-0c48-11e8-968a-f6612bb731fb   24Gi       RWX           class        83d
   

2. Review the ReclaimPolicy and billingType for the storage class.

   kubectl describe storageclass <storageclass_name>
   

   If the reclaim policy says Delete, your PV and the physical storage are removed when you remove the PVC. If the reclaim policy says Retain, or if you provisioned your storage without a storage class, then your PV and physical storage are not removed when you remove the PVC. You must remove the PVC, PV, and the physical storage separately.

   If your storage is charged monthly, you still get charged for the entire month, even if you remove the storage before the end of the billing cycle.

3. Remove any pods that mount the PVC. List the pods that mount the PVC. If no pod is returned in your CLI output, you don't have a pod that uses the PVC.

   kubectl get pods --all-namespaces -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.volumes[*]}{.persistentVolumeClaim.claimName}{" "}{end}{end}' | grep "<pvc_name>"
   

   Example output

   depl-12345-prz7b:    claim1
   

4. Remove the pod that uses the PVC. If the pod is part of a deployment, remove the deployment.

   kubectl delete pod <pod_name>
   

5. Verify that the pod is removed.

   kubectl get pods
   

6. Remove the PVC.

   kubectl delete pvc <pvc_name>
   

7. Review the status of your PV. Use the name of the PV that you retrieved earlier as VOLUME. When you remove the PVC, the PV that is bound to the PVC is released. Depending on how you provisioned your storage, your PV goes into a Deleting state if the PV is deleted automatically, or into a Released state, if you must manually delete the PV. Note: For PVs that are automatically deleted, the status might briefly say Released before it is deleted. Rerun the command after a few minutes to see whether the PV is removed.

   kubectl get pv <pv_name>
   

8. If your PV is not deleted, manually remove the PV.

   kubectl delete pv <pv_name>
   

9. Verify that the PV is removed.

   kubectl get pv
   

10. List the physical storage instance that your PV pointed to and note the id of the physical storage instance.

    ibmcloud sl file volume-list --columns id  --columns notes | grep <pv_name>
    

    Example output for File Storage for Classic.

    id         notes   
    12345678   {"plugin":"ibm-file-plugin-5b55b7b77b-55bb7","region":"us-south","cluster":"aa1a11a1a11b2b2bb22b22222c3c3333","type":"Endurance","ns":"default","pvc":"mypvc","pv":"pvc-d979977d-d79d-77d9-9d7d-d7d97ddd99d7","storageclass":"ibmc-file-gold"}
    

    "plugin":"ibm-file-plugin-5b55b7b77b-55bb7"

    The storage plug-in that the cluster uses.

    "region":"us-south"

    The region that your cluster is in.

    "cluster":"aa1a11a1a11b2b2bb22b22222c3c3333"

    The cluster ID that is associated with the storage instance.

    "type":"Endurance"

    The type of file or block storage, either Endurance or Performance.

    "ns":"default"

    The namespace that the storage instance is deployed to.

    "pvc":"mypvc"

    The name of the PVC that is associated with the storage instance.

    "pv":"pvc-d979977d-d79d-77d9-9d7d-d7d97ddd99d7"

    The PV that is associated with the storage instance.

    "storageclass":"ibmc-file-gold"

    The type of storage class: bronze, silver, gold, or custom.

11. Remove the physical storage instance.

    ibmcloud sl file volume-cancel <classic_file_id>
    

12. Verify that the physical storage instance is removed.

    ibmcloud sl file volume-list
    

The deletion process might take up to 72 hours to complete.