Skip to main content
Version: 3.1

Scheduled snapshots in GKE

Summary and Key concepts

Summary:​

This article provides a step-by-step guide for configuring and scheduling automatic cloud snapshots using Portworx and Stork in a Kubernetes environment. It explains how to set up cloud credentials, install the storkctl tool from the Stork container, and create schedule policies for snapshot management. It demonstrates two methods to associate snapshot schedules with storage: directly via VolumeSnapshotSchedule or through a StorageClass. The article also includes instructions for verifying the snapshot schedules and snapshots using storkctl to ensure the backups are working as expected.

Kubernetes Concepts:​

  • PersistentVolumeClaim (PVC): A request for storage by a Kubernetes user, linked to automatic snapshot scheduling in this article.
  • StorageClass: Defines storage types in Kubernetes. The article shows how to apply a snapshot schedule to all PVCs using a specific StorageClass.
  • Annotations: Used in both PVC and StorageClass definitions to specify details like snapshot type and cloud credentials.

Portworx Concepts:​

  • Stork: A tool that manages storage operations in Kubernetes, including scheduling snapshots. The storkctl command-line tool is used for operations like viewing snapshot schedules.
  • Cloud Snapshots: Backups of PVCs stored in cloud environments, requiring cloud credentials to configure and automate.

Prerequisites​

Configuring cloud secrets​

To create cloud snapshots, one needs to setup secrets with Portworx which will get used to connect and authenticate with the configured cloud provider.

Follow instructions on the create and configure credentials section to setup secrets.

Storkctl​

important

Always use the latest storkctl binary tool by downloading it from the current running Stork container.

Perform the following steps to download storkctl from the Stork pod:

  • Linux:

    STORK_POD=$(kubectl get pods -n <namespace> -l name=stork -o jsonpath='{.items[0].metadata.name}') &&
    kubectl cp -n <px-namespace> $STORK_POD:/storkctl/linux/storkctl ./storkctl
    sudo mv storkctl /usr/local/bin &&
    sudo chmod +x /usr/local/bin/storkctl

  • OS X:

    STORK_POD=$(kubectl get pods -n <namespace> -l name=stork -o jsonpath='{.items[0].metadata.name}') &&
    kubectl cp -n <px-namespace> $STORK_POD:/storkctl/darwin/storkctl ./storkctl
    sudo mv storkctl /usr/local/bin &&
    sudo chmod +x /usr/local/bin/storkctl

  • Windows:

    1. Copy storkctl.exe from the stork pod:

      STORK_POD=$(kubectl get pods -n <px-namespace> -l name=stork -o jsonpath='{.items[0].metadata.name}') &&
      kubectl cp -n <px-namespace> $STORK_POD:/storkctl/windows/storkctl.exe ./storkctl.exe

    2. Move storkctl.exe to a directory in your PATH.

Create a schedule policy​

You can use a schedule policy to specify when Portworx should trigger a specific action.

  1. Create a file named daily-policy.yaml, specifying the following fields and values:

    • apiVersion: with the version of the Stork scheduler (this example uses stork.libopenstorage.org/v1alpha1)

    • kind: with the SchedulePolicy value

    • metadata.name: with the name of the SchedulePolicy object (this example uses daily)

    • policy.daily.time: with the backup time (this example uses "10:14PM")

    • policy.retain: with the number of backups Portworx must retain (this example retains 3 backups)

      apiVersion: stork.libopenstorage.org/v1alpha1
      kind: SchedulePolicy
      metadata:
        name: daily
      policy:
        daily:
          time: "10:14PM"
          retain: 3

    For more details about how you can configure aschedule policy, see the Schedule Policy reference page.

  2. Apply the spec:

    kubectl apply -f daily-policy.yaml
    schedulepolicy.stork.libopenstorage.org/daily created

  3. You can check the status of your schedule policy by entering the storkctl get schedulepolicy command:

    storkctl get schedulepolicy
    NAME      INTERVAL-MINUTES   DAILY     WEEKLY             MONTHLY
    daily     N/A                10:14PM   N/A                N/A

Associate a schedule policy with a StorageClass or a Volume​

The following sections show how you can associate a schedule policy either with a Volume or a StorageClass.

Create a VolumeSnapshotSchedule​

Use a VolumeSnapshotSchedule to associate your schedule policy at the CRD level, and back up specific volumes according to a schedule you define.

  1. Create a file called volume-snapshot-schedule.yaml specifying the following fields and values:
  • metadata:
    • name: with the name of this VolumeSnapshotSchedule policy
    • namespace: the namespace in which this policy will exist
    • annotations:
      • portworx/snapshot-type: with the cloud or local value, depending on what environment you want store your snapshots in
      • portworx/cloud-cred-id: with your cloud environment credentials
      • stork.libopenstorage.org/snapshot-restore-namespaces: with other namespaces snapshots taken with this policy can restore to
      • The following annotations are required when PX-Security is enabled:
        • openstorage.io/auth-secret-namespace: namespace where the kubernetes secret holding the auth token resides
        • openstorage.io/auth-secret-name: name of the kubernetes secret which holds the auth token
  • spec:

    • schedulePolicyName: with the name of the schedule policy you defined in the steps above

    • suspend: with a boolean value specifying if the schedule should be in a suspended state

    • preExecRule: with the name of a rule to run before taking the snapshot

    • postExecRule: with the name of a rule to run after taking the snapshot

    • reclaimPolicy: with retain or delete, indicating what Portworx should do with the snapshots that were created using the schedule. Specifying the delete value deletes the snapshots created by this schedule when the schedule is deleted.

    • template.spec.persistentVolumeClaimName: with the PVC you want this policy to apply to

      apiVersion: stork.libopenstorage.org/v1alpha1
      kind: VolumeSnapshotSchedule
      metadata:
        name: mysql-snapshot-schedule
        namespace: mysql
        annotations:
          portworx/snapshot-type: cloud
          portworx/cloud-cred-id: <cred_id>
          stork.libopenstorage.org/snapshot-restore-namespaces: otherNamespace
          # Add the below annotations when PX-Security is enabled.
          #openstorage.io/auth-secret-namespace: <secret-namespace>
          #openstorage.io/auth-secret-name: <secret-name>
      spec:
        schedulePolicyName: testpolicy
        suspend: false
        reclaimPolicy: Delete
        preExecRule: testRule
        postExecRule: otherTestRule
        template:
          spec:
            persistentVolumeClaimName: mysql-data

  1. Apply the spec:

    kubectl apply -f volume-snapshot-schedule.yaml

Create a storage class​

Use a StorageClass to apply your schedule policy to all PVCs using that StorageClass.

  1. Create a file called sc-with-snap-schedule.yaml with the following content:

    kind: StorageClass
    apiVersion: storage.k8s.io/v1
    metadata:
      name: px-sc-with-snap-schedules
      annotations:
        # Add the below annotations when PX-Security is enabled.
        #openstorage.io/auth-secret-namespace: <secret-namespace>
        #openstorage.io/auth-secret-name: <secret-name>
    provisioner: pxd.portworx.com
    parameters:
      repl: "2"
      snapshotschedule.stork.libopenstorage.org/default-schedule: |
        schedulePolicyName: daily
        annotations:
          portworx/snapshot-type: local
      snapshotschedule.stork.libopenstorage.org/weekly-schedule: |
        schedulePolicyName: weekly
        annotations:
          portworx/snapshot-type: cloud
          portworx/cloud-cred-id: <credential-uuid>
note

This example references two schedules:

  • The default-schedule backs up volumes to the local Portworx cluster daily.
  • The weekly-schedule backs up volumes to cloud storage every week.

  1. Apply the spec:

    kubectl apply -f

Specifying the cloud credential to use​

note

Specifying the portworx/cloud-cred-id is required only if you have more than one cloud credentials configured. If you have a single one, by default, that credential is used.

Let's list all the available cloud credentials we have.

PX_POD=$(kubectl get pods -l name=portworx -n <px-namespace> -o jsonpath='{.items[0].metadata.name}')
kubectl exec $PX_POD -n <px-namespace> -- /opt/pwx/bin/pxctl credentials list

The above command will output the credentials required to authenticate/access the objectstore. Pick the one you want to use for this snapshot schedule and specify it in the portworx/cloud-cred-id annotation in the StorageClass.

Next, let's apply our newly created storage class:

kubectl apply -f sc-with-snap-schedule.yaml
storageclass.storage.k8s.io/px-sc-with-snap-schedules created

Create a PVC​

After we've created the new StorageClass, we can refer to it by name in our PVCs like this:

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: pvc-snap-schedules-demo
  annotations:
    volume.beta.kubernetes.io/storage-class: px-sc-with-snap-schedules
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 2Gi

Paste the listing from above into a file named pvc-snap-schedules-demo.yaml and run:

kubectl create -f pvc-snap-schedules-demo.yaml
persistentvolumeclaim/pvc-snap-schedules-demo created

Let's see our PVC:

kubectl get pvc
NAME                      STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS                AGE
pvc-snap-schedules-demo   Bound    pvc-xxxxxxxx-xxxx-xxxx-xxxx-080027ee1df7   2Gi        RWO            px-sc-with-snap-schedules   14s

The above output shows that a volume named pvc-xxxxxxxx-xxxx-xxxx-xxxx-080027ee1df7 was automatically created and is now bounded to our PVC.

We're all set!

Checking snapshots​

Verifying snapshot schedules​

First let's verify that the snapshot schedules are created correctly.

storkctl get volumesnapshotschedules
NAME                                       PVC                       POLICYNAME   PRE-EXEC-RULE   POST-EXEC-RULE   RECLAIM-POLICY   SUSPEND   LAST-SUCCESS-TIME
pvc-snap-schedules-demo-default-schedule   pvc-snap-schedules-demo   daily                                         Retain           false
pvc-snap-schedules-demo-weekly-schedule    pvc-snap-schedules-demo   weekly                                        Retain           false

Here we can see 2 snapshot schedules, one daily and one weekly.

Verifying snapshots​

Now that we've put everything in place, we would want to verify that our cloudsnaps are created.

Using storkctl​

Also, you can use storkctl to make sure that the snapshots are created by running:

storkctl get volumesnapshots
NAME                                                                  PVC                       STATUS    CREATED               COMPLETED             TYPE
pvc-snap-schedules-demo-default-schedule-interval-2019-03-27-015546   pvc-snap-schedules-demo   Ready     26 Mar 19 21:55 EDT   26 Mar 19 21:55 EDT   local
pvc-snap-schedules-demo-weekly-schedule-interval-2019-03-27-015546    pvc-snap-schedules-demo   Ready     26 Mar 19 21:55 EDT   26 Mar 19 21:55 EDT   cloud
Last updated on