Volume snapshots

Volume snapshots allows you to initialize persistent volume claims with the contents of a preexisting volume snapshot. Use volume snapshots when working with large datasets or if you want to create development environments with real data coming from your production or staging environments.

Requirements

Self-Hosted

Follow this guide to enable Volume Snapshots in your cluster.

Use a storage class provisioned by a CSI driver that supports volume snapshots for the source persistent volume of your volume snapshots.

SaaS

If your instance is hosted by Okteto, Volume Snapshots are enabled by default.

Important: you need to use the csi-okteto storage classes for the source persistent volume of your volume snapshots.

Using Volume Snapshots in your Development Environment

Okteto enables developers to initialize persistent volume claims with the contents of a pre-existing volume snapshot. The volume snapshot is created from a persistent volume claim and it can contain database backups, big files, images, a copy of your staging data, etc.

In order to use volume snapshots with your development environment you need to:

• Create the source persistent volume
• Create a volume snapshot of the source persistent volume
• Consume the volume snapshot in a new persistent volume

Creating the Source Persistent Volume

First, create the source persistent volume. This is the data that you want to be able to clone into your development environment. In the example below we use a database, but this could be anything that uses a volume for storage: databases, ML models, images, etc...

If the default storage class of your cluster doesn't support volume snapshots, make sure you set the storage class to one that is compatible when creating your persistent volume. Check the requirements section to learn more about the available storage classes in Okteto.

Kubernetes

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: mysql-pvc
spec:
  storageClassName: csi-okteto
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 10Gi

Compose

services:
  mysql:
    image: mysql
    volumes:
      - mysql-pvc:/var/lib/mysql

volumes:
  mysql-pvc:
    driver_opts:
      class: csi-okteto

Creating a Volume Snapshot

After creating the source persistent volume, the next step is to create the Volume Snapshot. Volume snapshots are created with the content of the persistent volume at the time of creating the volume snapshot. Further updates on the persistent volume aren't reflected in the volume snapshot content.

This step is independent of whether you created your persistent volume with Kubernetes manifests or a Docker Compose file:

The following manifest creates a volume snapshot from the mysql-pvc persistent volume claim:

apiVersion: snapshot.storage.k8s.io/v1
kind: VolumeSnapshot
metadata:
  name: mysql-snapshot
spec:
  volumeSnapshotClassName: okteto-snapshot-class
  source:
    persistentVolumeClaimName: mysql-pvc

Consuming a Volume Snapshot

Finally, you can use the volume snapshot created in the previous step to populate a new persistent volume. Use the dev.okteto.com/from-snapshot-name and dev.okteto.com/from-snapshot-namespace annotations on any persistent volume claim to tell Okteto to initialize your persistent volume claim from an existing volume snapshot, as shown below:

Kubernetes

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  annotations:
    dev.okteto.com/from-snapshot-name: mysql-snapshot
    dev.okteto.com/from-snapshot-namespace: staging
  name: pvc-from-snapshot
spec:
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 10Gi

Compose

services:
  mongodb:
    image: mysql
    volumes:
      - data:/var/lib/mysql

volumes:
  data:
    labels:
      dev.okteto.com/from-snapshot-name: mysql-snapshot
      dev.okteto.com/from-snapshot-namespace: staging

If the annotation dev.okteto.com/from-snapshot-namespace is not defined, Okteto will default to the namespace of the new persistent volume claim.

Use the annotation dev.okteto.com/skip-snapshot-if-same-namespace: "true" to skip the data cloning operation if the source snapshot and the new persistent volume claim are in the same namespace.

For a full end to end sample, check the following resources:

• How to Create a Development Environment with Realistic Data in Okteto blog post
• How to Create and Use Data Clones in Okteto video tutorial