CVMFS CSI driver

csi-cvmfs provides read-only mounting of CVMFS volumes in CSI-enabled container orchestrators.

Building

To build the binary in ./bin:

make build

To build binaries for multiple platforms:

make build-cross

To make the docker image:

make docker

Configuration

Available command line arguments:

Option	Default value	Description
--endpoint	unix://tmp/csi.sock	CSI endpoint, must be a UNIX socket
--drivername	csi-cvmfsplugin	name of the driver (Kubernetes: provisioner field in StorageClass must correspond to this value)
--nodeid	empty	This node's ID

Available volume parameters:

Parameter	Required	Description
repository	yes	Address of the CVMFS repository
tag	no	CVMFS_REPOSITORY_TAG. Defaults to trunk
hash	no	CVMFS_REPOSITORY_HASH
proxy	no	CVMFS_HTTP_PROXY. Defaults to the value sourced from default.local. See instructions below.

By default, csi-cvmfs is distributed with default.local containing CERN defaults. You can override those at runtime by overwriting /etc/cvmfs/default.local, which is then sourced into any later CVMFS client configs used for mounting.

CVMFS configuration in Kubernetes You can use Kubernetes config map to inject a custom default.local file.

$ cat my-config
CVMFS_HTTP_PROXY="http://my-cvmfs-proxy:3128"
CVMFS_QUOTA_LIMIT=5000
$ kubectl create configmap my-configmap --from-file=cvmfs-override=./my-config

Edit the manifest for csi-cvmfsplugin: In the volumes section, add an entry for the config map:

volumes:
  - name: my-cvmfs-config
    configMap:
      name: my-configmap

Add a volume mount for the csi-cvmfsplugin container:

volumeMounts:
  - name: my-cvmfs-config
    mountPath: /etc/cvmfs/default.local
    subPath: cvmfs-override

Deployment with Kubernetes

Requires Kubernetes 1.11+

Your Kubernetes cluster must allow privileged pods (i.e. --allow-privileged flag must be set to true for both the API server and the kubelet). Moreover, as stated in the mount propagation docs, the Docker daemon of the cluster nodes must allow shared mounts.

YAML manifests are located in deploy/kubernetes.

Deploy RBACs for sidecar containers and node plugins:

$ kubectl create -f namespace.yaml
$ kubectl create -f csi-attacher-rbac.yaml
$ kubectl create -f csi-provisioner-rbac.yaml
$ kubectl create -f csi-nodeplugin-rbac.yaml

Those manifests deploy service accounts, cluster roles and cluster role bindings.

Deploy CSI sidecar containers:

$ kubectl create -f csi-cvmfsplugin-attacher.yaml
$ kubectl create -f csi-cvmfsplugin-provisioner.yaml

Deploys stateful sets for external-attacher and external-provisioner sidecar containers for CSI CernVM-FS.

Deploy the CSI CernVM-FS driver:

$ kubectl create -f csi-cvmfsplugin.yaml

Deploys a daemon set with two containers: CSI driver-registrar and the CSI CernVM-FS driver.

Verifying the deployment in Kubernetes

After successfuly completing the steps above, you should see output similar to this:

$ kubectl get all --namespace=cvmfs
NAME                                READY     STATUS    RESTARTS   AGE
pod/csi-cvmfsplugin-attacher-0      1/1       Running   0          1m
pod/csi-cvmfsplugin-bhqck           2/2       Running   0          1m
pod/csi-cvmfsplugin-provisioner-0   1/1       Running   0          1m

NAME                                  TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)     AGE
service/csi-cvmfsplugin-attacher      ClusterIP   10.104.25.190   <none>        12345/TCP   1m
service/csi-cvmfsplugin-provisioner   ClusterIP   10.101.197.42   <none>        12345/TCP   1m

NAME                             DESIRED   CURRENT   READY     UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
daemonset.apps/csi-cvmfsplugin   1         1         1         1            1           <none>          1m

NAME                                           DESIRED   CURRENT   AGE
statefulset.apps/csi-cvmfsplugin-attacher      1         1         1m
statefulset.apps/csi-cvmfsplugin-provisioner   1         1         1m

...

You can try deploying a demo pod from examples/ to test the deployment further.