Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time



A Container Storage Interface (CSI) driver for volumes. The CSI plugin allows you to use volumes with your preferred Container Orchestrator.

The CSI plugin is mostly tested on Kubernetes. In theory, it should also work on other Container Orchestrators like Mesos or Cloud Foundry. Feel free to test it on other COs and give us feedback.


# Add a API token as secret, replace the placeholder string starting with `a05...` with your own secret
$ kubectl -n kube-system create secret generic cloudscale --from-literal=access-token=a05dd2f26b9b9ac2asdas__REPLACE_ME____123cb5d1ec17513e06da
# Add repository
$ helm repo add csi-cloudscale
# Install driver
$ helm install -n kube-system -g csi-cloudscale/csi-cloudscale

Volume parameters

This plugin supports the following volume parameters (in case of Kubernetes: parameters on the StorageClass object):

  • ssd or bulk; defaults to ssd if not set

For LUKS encryption:

  • set to the string "true" if the volume should be encrypted with LUKS
  • cipher to use; must be supported by the kernel and LUKS, we suggest aes-xts-plain64
  • key-size to use; we suggest 512 for aes-xts-plain64

For LUKS encrypted volumes, a secret that contains the LUKS key needs to be referenced through the and parameter. See the included StorageClass definitions and the examples/kubernetes/luks-encrypted-volumes folder for examples.

Pre-defined storage classes

The default deployment bundled in the deploy/kubernetes/releases folder includes the following storage classes:

  • cloudscale-volume-ssd - the default storage class; uses an ssd volume, no LUKS encryption
  • cloudscale-volume-bulk - uses a bulk volume, no LUKS encryption
  • cloudscale-volume-ssd-luks - uses an ssd volume that will be encrypted with LUKS; a luks-key must be supplied
  • cloudscale-volume-bulk-luks - uses a bulk volume that will be encrypted with LUKS; a luks-key must be supplied

To use one of the shipped LUKS storage classes, you need to create a secret named ${}-luks-key in the same namespace as the persistent volume claim. The secret must contain an element called luksKey that will be used as the LUKS encryption key.

Example: If you create a persistent volume claim with the name my-pvc, you need to create a secret my-pvc-luks-key.


The CSI plugin follows semantic versioning. The current version is: v3.5.3.

  • Bug fixes will be released as a PATCH update.
  • New features (such as CSI spec bumps) will be released as a MINOR update.
  • Significant breaking changes makes a MAJOR update.

Installing to Kubernetes

Kubernetes Compatibility

The following table describes the required driver version per Kubernetes release. We recommend using the latest CSI driver compatible with your Kubernetes release.

Kubernetes Release Minimum CSI driver Maximum CSI driver
<= 1.16 v1.3.1
1.17 v1.3.1 v3.0.0
1.18 v1.3.1 v3.3.0
1.19 v1.3.1 v3.3.0
1.20 v2.0.0 v3.5.2
1.21 v2.0.0 v3.5.2
1.22 v3.1.0 v3.5.2
1.23 v3.1.0 v3.5.2
1.24 v3.1.0 v3.5.3
1.25 v3.3.0 v3.5.3
1.26 v3.3.0 v3.5.3
1.27 v3.3.0 v3.5.3
1.28 v3.3.0 v3.5.3


  • Nodes must be able to access the metadata service at using HTTP. The required route is pushed by DHCP.
  • --allow-privileged flag must be set to true for both the API server and the kubelet
  • (if you use Docker) the Docker daemon of the cluster nodes must allow shared mounts
  • If you want to use LUKS encrypted volumes, the kernel on your nodes must have support for device mapper infrastructure with the crypt target and the appropriate cryptographic APIs

1. Create a secret with your API Access Token:

Replace the placeholder string starting with a05... with your own secret and save it as secret.yml:

apiVersion: v1
kind: Secret
  name: cloudscale
  namespace: kube-system
  access-token: "a05dd2f26b9b9ac2asdas__REPLACE_ME____123cb5d1ec17513e06da"

and create the secret using kubectl:

$ kubectl create -f ./secret.yml
secret "cloudscale" created

You should now see the cloudscale secret in the kube-system namespace along with other secrets

$ kubectl -n kube-system get secrets
NAME                  TYPE                                  DATA      AGE
default-token-jskxx   3         18h
cloudscale            Opaque                                1         18h

2. Deploy the CSI plugin and sidecars:

You can install the CSI plugin and sidecars using one of the following methods:

  • Helm (requires a Helm installation)
  • YAML Manifests (only kubectl required)

2a. Using Helm:

Before you can install the csi-cloudscale chart, you need to add the helm repository:

$ helm repo add csi-cloudscale

Then install the latest stable version:

$ helm install -n kube-system -g csi-cloudscale/csi-cloudscale

Advanced users can customize the installation by specifying custom values. The following table summarizes the most-frequently used parameters. For a complete list please refer to values.yaml

Parameter Default Description
attacher.resources {} Resource limits and requests for the attacher side-car.
cloudscale.apiUrl URL of the API. You can almost certainly use the default.
cloudscale.max_csi_volumes_per_node 125 Override max. Number of CSI Volumes per Node.
cloudscale.token.existingSecret cloudscale Name of the Kubernetes Secret which contains the API Token.
controller.resources {} Resource limits and requests for the controller container.
controller.serviceAccountName null Override the controller service account name.
driverRegistrar.resources {} Resource limits and requests for the driverRegistrar side-car.
extraDeploy [] To deploy extra objects together with the driver.
nameOverride null Override the default {{ .Release.Name }}-csi-cloudscale name pattern with a custom name.
node.resources {} Resource limits and requests for the node container.
node.serviceAccountName null Override the controller node account name.
node.tolerations [] Set tolerations on the node daemonSet.
provisioner.resources {} Resource limits and requests for the provisioner side-car.
resizer.resources {} Resource limits and requests for the resizer side-car.

Note: if you want to test a debug/dev release, you can use the following command:

$ helm install -g -n kube-system --set controller.image.tag=dev --set node.image.tag=dev --set controller.image.pullPolicy=Always --set node.image.pullPolicy=Always ./charts/csi-cloudscale

2b. Using YAML Manifests:

Before you continue, be sure to checkout to a tagged release. Always use the latest stable version For example, to use the latest stable version (v3.5.3) you can execute the following command:

$ kubectl apply -f

The storage classes cloudscale-volume-ssd and cloudscale-volume-bulk will be created. The storage class cloudscale-volume-ssd is set to "default" for dynamic provisioning. If you're using multiple storage classes you might want to remove the annotation and re-deploy it. This is based on the recommended mechanism of deploying CSI drivers on Kubernetes

3. Test and verify:

Create a PersistentVolumeClaim. This makes sure a volume is created and provisioned on your behalf:

apiVersion: v1
kind: PersistentVolumeClaim
  name: csi-pvc
  - ReadWriteOnce
      storage: 5Gi
  storageClassName: cloudscale-volume-ssd

Check that a new PersistentVolume is created based on your claim:

$ kubectl get pv
NAME                                       CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS    CLAIM             STORAGECLASS            REASON    AGE
pvc-0879b207-9558-11e8-b6b4-5218f75c62b9   5Gi        RWO            Delete           Bound     default/csi-pvc   cloudscale-volume-ssd             3m

The above output means that the CSI plugin successfully created (provisioned) a new Volume on behalf of you. You should be able to see this newly created volumes in the server detail view in the UI.

The volume is not attached to any node yet. It will only attach to a node if a workload (i.e: pod) is scheduled to a specific node. Now let us create a Pod that refers to the above volume. When the Pod is created, the volume will be attached, formatted and mounted to the specified container:

kind: Pod
apiVersion: v1
  name: my-csi-app
    - name: my-frontend
      image: busybox
      - mountPath: "/data"
        name: my-cloudscale-volume
      command: [ "sleep", "1000000" ]
    - name: my-cloudscale-volume
        claimName: csi-pvc 

Check if the pod is running successfully:

$ kubectl describe pods/my-csi-app

Write inside the app container:

$ kubectl exec -ti my-csi-app /bin/sh
/ # touch /data/hello-world
/ # exit
$ kubectl exec -ti my-csi-app /bin/sh
/ # ls /data


From csi-cloudscale v1.x to v2.x

When updating from csi-cloudscale v1.x to v2.x please note the following:

  • Ensure that all API objects of the existing v1.x installation are removed. The easiest way to achieve this is by running kubectl delete -f <old version> before installing the new driver version.
  • Prior to the installation of v2.x, existing persistent volumes (PVs) must be annotated with: "". You can use this script or any other means to set the annotation.
  • If you are using self defined storage classes: change the storage class provisioner names from "ch.cloudscale.csi" to "".

From csi-cloudscale v2.x to v3.x

When updating from csi-cloudscale v2.x to v3.x please note the following:

  • The node label region was renamed to
  • The new release adds the label to all nodes (existing ones as well as new ones added after the upgrade)
  • The region label will stay in place for existing nodes and not be added to new nodes. It can be safely removed from all nodes from a csi-cloudscale driver perspective.

Advanced Configuration

Please use the following options with care.

Max. Number of CSI Volumes per Node

In the v1.3.0 release the default CSI volumes per node limit of has been increased to 125 (previously 23). To take advantage of the higher CSI limit you must ensure that all your cluster nodes are using virtio-scsi devices (i.e. /dev/sdX devices are used). This is the default for servers created after October 1st, 2020.

If you want to use a different value, for example because one of your nodes does not use virtio-scsi, you can set the following environment variable for the csi-cloudscale-plugin container in the csi-cloudscale-node DaemonSet:

   value: '10'

Or use the cloudscale.max_csi_volumes_per_node value of the Helm chart.

Note that there are currently the following hard-limits per Node:

  • 26 volumes (including root) for virtio-blk (/dev/vdX).
  • 128 volumes (including root) for virtio-scsi (/dev/sdX).



  • Go: min v1.10.x
  • Helm

Build out the charts/ directory from the Chart.lock file:

$ cd charts/csi-cloudscale/
$ helm repo add bitnami
$ helm repo update
$ helm dependency build charts/csi-cloudscale

Install the chart from local sources:

$ helm install -n kube-system -g ./charts/csi-cloudscale

Useful commands to compare the generated helm chart to the static YAML manifests:

$ helm template csi-cloudscale --dry-run -n kube-system --set nameOverride=csi-cloudscale charts/csi-cloudscale | kubectl-slice -f - -o deploy/kubernetes/releases/generated
$ kubectl-slice -f deploy/kubernetes/releases/csi-cloudscale-v6.0.0.yaml -o deploy/kubernetes/releases/v3

After making your changes, run the unit tests:

$ make test

Note: If you want to run just a single test case, from csi-test, find the corresponding, It in the source code, and temporarly replace it with FIt, example:

- It("should work if node-expand is called after node-publish", func() {
+ FIt("should work if node-expand is called after node-publish", func() {

If you want to test your changes, create a new image with the version set to dev:

apt install
# At this point you probably need to add your user to the docker group
docker login --username=cloudscalech
$ VERSION=dev make publish

This will create a binary with version dev and docker image pushed to cloudscalech/cloudscale-csi-plugin:dev

To run the integration tests run the following:

$ export KUBECONFIG=$(pwd)/kubeconfig 
$ TESTARGS='-run TestPod_Single_SSD_Volume' make test-integration

Release a new version

To release a new version bump first the version:

$ make NEW_VERSION=vX.Y.Z bump-version
$ make NEW_CHART_VERSION=vX.Y.Z bump-chart-version

Make sure everything looks good. Verify that the Kubernetes compatibility matrix is up-to-date. Create a new branch with all changes:

$ git checkout -b new-release
$ git add .
$ git push origin

After it's merged to master, create a new Github release from master with the version v3.5.3 and then publish a new docker build:

$ git checkout master
$ make publish

This will create a binary with version v3.5.3 and docker image pushed to cloudscalech/cloudscale-csi-plugin:v3.5.3


At we value and love our community! If you have any issues or would like to contribute, feel free to open an issue/PR