Deploying with the Trident Operator

If you are looking to deploy Trident using the Trident Operator, you are in the right place. This page contains all the steps required for getting started with the Trident Operator to install and manage Trident. You can deploy Trident Operator either manually or using Helm.


If you have not already familiarized yourself with the basic concepts, now is a great time to do that. Go ahead, we’ll be here when you get back.

To deploy Trident using the operator you need:

  • Full privileges to a supported Kubernetes cluster running Kubernetes 1.14 and above. If choosing to deploy with Helm, you will need Kubernetes 1.16 and above.
  • Helm 3 (if deploying using Helm).
  • Access to a supported NetApp storage system
  • Volume mount capability from all of the Kubernetes worker nodes
  • A Linux host with kubectl (or oc, if you’re using OpenShift) installed and configured to manage the Kubernetes cluster you want to use
  • Set the KUBECONFIG environment variable to point to your Kubernetes cluster configuration.
  • Enable the Feature Gates required by Trident
  • If you are using Kubernetes with Mirantis Container Runtime (formerly Docker Engine - Enterprise), follow their steps to enable CLI access.

Got all that? Great! Let’s get started. You can choose to either Deploy using Helm or Using the Trident Operator.

Deploy Trident Operator by using Helm

Perform the steps listed to deploy Trident Operator by using Helm. You will need the following:

  • Kubernetes 1.16 and later
  • Helm version 3

1: Download the installer bundle

Download the Trident 21.04 installer bundle from the Trident GitHub page. The installer bundle includes the Helm chart in the /helm directory.

2: Deploy the Trident operator

Use the helm install command and specify a name for your deployment. See the following example:

$ helm install <name> trident-operator-21.04.0.tgz

There are two ways to pass configuration data during the install:

  • –values (or -f): Specify a YAML file with overrides. This can be specified multiple times and the rightmost file will take precedence.
  • –set: Specify overrides on the command line.

For example, to change the default value of debug, run the following –set command:

$ helm install <name> trident-operator-21.04.0.tgz --set tridentDebug=true

The values.yaml file, which is part of the Helm chart provides the list of keys and their default values.

helm list shows you details about the Trident installation, such as name, namespace, chart, status, app version, revision number, and so on.

Deploy Trident Operator manually

Perform the steps listed to manually deploy Trident Operator.

If you are interested in upgrading an operator-based Trident install to the latest release, take a look at Upgrading Trident.

1: Qualify your Kubernetes cluster

You made sure that you have everything in hand from the previous section, right? Right.

The first thing you need to do is log into the Linux host and verify that it is managing a working, supported Kubernetes cluster that you have the necessary privileges to.


With OpenShift, you will use oc instead of kubectl in all of the examples that follow, and you need to login as system:admin first by running oc login -u system:admin or oc login -u kube-admin.

# Is your Kubernetes version greater than 1.14?
kubectl version

# Are you a Kubernetes cluster administrator?
kubectl auth can-i '*' '*' --all-namespaces

# Can you launch a pod that uses an image from Docker Hub and can reach your
# storage system over the pod network?
kubectl run -i --tty ping --image=busybox --restart=Never --rm -- \
  ping <management IP>

2: Download & setup the operator


Beginning with 21.01, the Trident Operator is cluster-scoped. Using the Trident Operator to install Trident requires creating the TridentOrchestrator Custom Resource Definition and defining other resources. You will need to perform these steps to setup the operator before you can install Trident.

Download the latest version of the Trident installer bundle from the Downloads section and extract it.

tar -xf trident-installer-21.04.0.tar.gz
cd trident-installer

Use the appropriate CRD manifest to create the TridentOrchestrator Custom Resource Definition. You will then create a TridentOrchestrator Custom Resource later on to instantiate a Trident install by the operator.

# Is your Kubernetes version < 1.16?
kubectl create -f deploy/crds/trident.netapp.io_tridentorchestrators_crd_pre1.16.yaml

# If not, your Kubernetes version must be 1.16 and above
kubectl create -f deploy/crds/trident.netapp.io_tridentorchestrators_crd_post1.16.yaml

Once the TridentOrchestrator CRD is created, you will then have to create the resources required for the operator deployment, such as:

  • a ServiceAccount for the operator.
  • a ClusterRole and ClusterRoleBinding to the ServiceAccount.
  • a dedicated PodSecurityPolicy.
  • the Operator itself.

The Trident Installer contains manifests for defining these resources. By default the operator is deployed in trident namespace, if the trident namespace does not exist use the below manifest to create one.

$ kubectl apply -f deploy/namespace.yaml

If you would like to deploy the operator in a namespace other than the default trident namespace, you will need to update the serviceaccount.yaml, clusterrolebinding.yaml and operator.yaml manifests and generate your bundle.yaml.

# Have you updated the yaml manifests? Generate your bundle.yaml
# using the kustomization.yaml
kubectl kustomize deploy/ > deploy/bundle.yaml

# Create the resources and deploy the operator
kubectl create -f deploy/bundle.yaml

You can check the status of the operator once you have deployed.

$ kubectl get deployment -n <operator-namespace>
trident-operator   1/1     1            1           3m

$ kubectl get pods -n <operator-namespace>
NAME                              READY   STATUS             RESTARTS   AGE
trident-operator-54cb664d-lnjxh   1/1     Running            0          3m

The operator deployment successfully creates a pod running on one of the worker nodes in your cluster.


There must only be one instance of the operator in a Kubernetes cluster. Do not create multiple deployments of the Trident operator.

3: Creating a TridentOrchestrator and installing Trident

You are now ready to install Trident using the operator! This will require creating a TridentOrchestrator. The Trident installer comes with example definitions for creating a TridentOrchestrator. This kicks off a Trident installation in the trident namespace.

$ kubectl create -f deploy/crds/tridentorchestrator_cr.yaml created

$  kubectl describe torc trident
Name:         trident
Labels:       <none>
Annotations:  <none>
API Version:
Kind:         TridentOrchestrator
  Debug:      true
  Namespace:  trident
  Current Installation Params:
    IPv6:                       false
    Autosupport Hostname:
    Autosupport Image:          netapp/trident-autosupport:21.04
    Autosupport Proxy:
    Autosupport Serial Number:
    Debug:                      true
    Enable Node Prep:           false
    Image Pull Secrets:
    Image Registry:
    k8sTimeout:           30
    Kubelet Dir:          /var/lib/kubelet
    Log Format:           text
    Silence Autosupport:  false
    Trident Image:        netapp/trident:21.04.0
  Message:                Trident installed
  Namespace:              trident
  Status:                 Installed
  Version:                v21.04.0
  Type    Reason      Age   From                        Message
  ----    ------      ----  ----                        -------
  Normal  Installing  74s  Installing Trident
  Normal  Installed   67s  Trident installed

Customizing your deployment

The Trident operator provides users the ability to customize the manner in which Trident is installed, using the following attributes in the TridentOrchestrator spec:

Parameter Description Default
namespace Namespace to install Trident in “default”
debug Enable debugging for Trident ‘false’
useIPv6 Install Trident over IPv6 ‘false’
k8sTimeout Timeout for Kubernetes operations 30sec
silenceAutosupport Don’t send autosupport bundles to NetApp automatically ‘false’
enableNodePrep Manage worker node dependencies automatically (BETA) ‘false’
autosupportImage The container image for Autosupport Telemetry “netapp/trident-autosupport:21.04.0”
autosupportProxy The address/port of a proxy for sending Autosupport Telemetry
uninstall A flag used to uninstall Trident ‘false’
logFormat Trident logging format to be used [text,json] “text”
tridentImage Trident image to install “netapp/trident:21.04”
imageRegistry Path to internal registry, of the format <registry FQDN>[:port][/subpath] “ (k8s 1.17+) or”
kubeletDir Path to the kubelet directory on the host “/var/lib/kubelet”
wipeout A list of resources to delete to perform a complete removal of Trident  
imagePullSecrets Secrets to pull images from an internal registry  


spec.namespace is specified in the tridentOrchestrator to signify which namespace Trident is installed in. This parameter cannot be updated after Trident is installed. Attempting to do so will cause the Status of tridentOrchestrator to change to Failed. Trident is not meant to be migrated across namespaces.


Automatic worker node prep is a beta feature meant to be used in non-production environments only.

You can use the attributes mentioned above when defining a TridentOrchestrator to customize your Trident installation. Here’s an example:

$ cat deploy/crds/tridentorchestrator_cr_imagepullsecrets.yaml
kind: TridentOrchestrator
  name: trident
  debug: true
  namespace: trident
  tridentImage: netapp/trident:21.04.0
  - thisisasecret

If you are looking to customize Trident’s installation beyond what the TridentOrchestrator’s arguments allow, you should consider using tridentctl to generate custom yaml manifests that you can modify as desired. Head on over to the deployment guide for tridentctl to learn how this works.

Observing the status of the operator

The Status of the TridentOrchestrator will indicate if the installation was successful and will display the version of Trident installed.

Status Description
Installing The operator is installing Trident using this TridentOrchestrator CR.
Installed Trident has successfully installed.
Uninstalling The operator is uninstalling Trident, since spec.uninstall=true.
Uninstalled Trident is uninstalled.
Failed The operator could not install, patch, update or uninstall Trident; the
  operator will automatically try to recover from this state. If this
  state persists you will require troubleshooting.
Updating The operator is updating an existing Trident installation.
Error The TridentOrchestrator is not used. Another one already exists.

During the installation, the status of the TridentOrchestrator will change from Installing to Installed. If you observe the Failed status and the operator is unable to recover by itself, there’s probably something wrong and you will need to check the logs of the operator by running tridentctl logs -l trident-operator.

You can also confirm if the Trident install completed by taking a look at the pods that have been created:

$ kubectl get pod -n trident
NAME                                READY   STATUS    RESTARTS   AGE
trident-csi-7d466bf5c7-v4cpw        5/5     Running   0           1m
trident-csi-mr6zc                   2/2     Running   0           1m
trident-csi-xrp7w                   2/2     Running   0           1m
trident-csi-zh2jt                   2/2     Running   0           1m
trident-operator-766f7b8658-ldzsv   1/1     Running   0           3m

You can also use tridentctl to check the version of Trident installed.

$ ./tridentctl -n trident version
| 21.04.0        | 21.04.0        |

If that’s what you see, you’re done with this step, but Trident is not yet fully configured. Go ahead and continue to the next step to create a Trident backend using tridentctl.

However, if the installer does not complete successfully or you don’t see a Running trident-csi-<generated id>, then Trident had a problem and the platform was not installed.

To understand why the installation of Trident was unsuccessful, you should first take a look at the TridentOrchestrator status.

$ kubectl describe torc trident-2
Name:         trident-2
Labels:       <none>
Annotations:  <none>
API Version:
Kind:         TridentOrchestrator
  Current Installation Params:
    Autosupport Hostname:
    Autosupport Image:
    Autosupport Proxy:
    Autosupport Serial Number:
    Enable Node Prep:
    Image Pull Secrets:         <nil>
    Image Registry:
    Kubelet Dir:
    Log Format:
    Silence Autosupport:
    Trident Image:
  Message:                      Trident is bound to another CR 'trident'
  Namespace:                    trident-2
  Status:                       Error
  Type     Reason  Age                From                        Message
  ----     ------  ----               ----                        -------
  Warning  Error   16s (x2 over 16s)  Trident is bound to another CR 'trident'

This error indicates that there already exists a TridentOrchestrator that was used to install Trident. Since each Kubernetes cluster can only have one instance of Trident, the operator ensures that at any given time there only exists one active TridentOrchestrator that it can create.

Another thing to do is to check the operator logs. Trailing the logs of the trident-operator container can point to where the problem lies.

$ tridentctl logs -l trident-operator

For example, one such issue could be the inability to pull the required container images from upstream registries in an airgapped environment. The logs from the operator can help identify this problem and fix it.

In addition, observing the status of the Trident pods can often indicate if something is not right.

$ kubectl get pods -n trident

NAME                                READY   STATUS             RESTARTS   AGE
trident-csi-4p5kq                   1/2     ImagePullBackOff   0          5m18s
trident-csi-6f45bfd8b6-vfrkw        4/5     ImagePullBackOff   0          5m19s
trident-csi-9q5xc                   1/2     ImagePullBackOff   0          5m18s
trident-csi-9v95z                   1/2     ImagePullBackOff   0          5m18s
trident-operator-766f7b8658-ldzsv   1/1     Running            0          8m17s

You can clearly see that the pods are not able to initialize completely as one or more container images were not fetched.

To address the problem, you must edit the TridentOrchestrator CR. Alternatively, you can delete the TridentOrchestrator and create a new one with the modified, accurate definition.

If you continue to have trouble, visit the troubleshooting guide for more advice.

Post-deployment steps

After you deploy Trident with the operator, you can proceed with creating a Trident backend, creating a storage class, provisioning a volume, and mounting the volume in a pod.

1: Creating a Trident backend

You can now go ahead and create a backend that will be used by Trident to provision volumes. To do this, create a backend.json file that contains the necessary parameters. Sample configuration files for different backend types can be found in the sample-input directory.

Visit the backend configuration guide for more details about how to craft the configuration file for your backend type.

cp sample-input/<backend template>.json backend.json
# Fill out the template for your backend
vi backend.json
./tridentctl -n trident create backend -f backend.json
|    NAME     | STORAGE DRIVER |                 UUID                 | STATE  | VOLUMES |
| nas-backend | ontap-nas      | 98e19b74-aec7-4a3d-8dcf-128e5033b214 | online |       0 |

If the creation fails, something was wrong with the backend configuration. You can view the logs to determine the cause by running:

./tridentctl -n trident logs

After addressing the problem, simply go back to the beginning of this step and try again. If you continue to have trouble, visit the troubleshooting guide for more advice on how to determine what went wrong.

2: Creating a Storage Class

Kubernetes users provision volumes using persistent volume claims (PVCs) that specify a storage class by name. The details are hidden from users, but a storage class identifies the provisioner that will be used for that class (in this case, Trident) and what that class means to the provisioner.

Create a storage class Kubernetes users will specify when they want a volume. The configuration of the class needs to model the backend that you created in the previous step so that Trident will use it to provision new volumes.

The simplest storage class to start with is one based on the sample-input/storage-class-csi.yaml.templ file that comes with the installer, replacing __BACKEND_TYPE__ with the storage driver name.

./tridentctl -n trident get backend
|    NAME     | STORAGE DRIVER |                 UUID                 | STATE  | VOLUMES |
| nas-backend | ontap-nas      | 98e19b74-aec7-4a3d-8dcf-128e5033b214 | online |       0 |

cp sample-input/storage-class-csi.yaml.templ sample-input/storage-class-basic-csi.yaml

# Modify __BACKEND_TYPE__ with the storage driver field above (e.g., ontap-nas)
vi sample-input/storage-class-basic-csi.yaml

This is a Kubernetes object, so you will use kubectl to create it in Kubernetes.

kubectl create -f sample-input/storage-class-basic-csi.yaml

You should now see a basic storage class in both Kubernetes and Trident, and Trident should have discovered the pools on the backend.

kubectl get sc basic-csi
NAME         PROVISIONER             AGE
basic-csi   15h

./tridentctl -n trident get storageclass basic-csi -o json
  "items": [
      "Config": {
        "version": "1",
        "name": "basic-csi",
        "attributes": {
          "backendType": "ontap-nas"
        "storagePools": null,
        "additionalStoragePools": null
      "storage": {
        "ontapnas_10.0.0.1": [

3: Provision your first volume

Now you’re ready to dynamically provision your first volume. How exciting! This is done by creating a Kubernetes persistent volume claim (PVC) object, and this is exactly how your users will do it too.

Create a persistent volume claim (PVC) for a volume that uses the storage class that you just created.

See sample-input/pvc-basic-csi.yaml for an example. Make sure the storage class name matches the one that you created in 6.

kubectl create -f sample-input/pvc-basic-csi.yaml

kubectl get pvc --watch
NAME      STATUS    VOLUME                                     CAPACITY   ACCESS MODES  STORAGECLASS   AGE
basic     Pending                                                                       basic          1s
basic     Pending   pvc-3acb0d1c-b1ae-11e9-8d9f-5254004dfdb7   0                        basic          5s
basic     Bound     pvc-3acb0d1c-b1ae-11e9-8d9f-5254004dfdb7   1Gi        RWO           basic          7s

4: Mount the volume in a pod

Now that you have a volume, let’s mount it. We’ll launch an nginx pod that mounts the PV under /usr/share/nginx/html.

cat << EOF > task-pv-pod.yaml
kind: Pod
apiVersion: v1
  name: task-pv-pod
    - name: task-pv-storage
       claimName: basic
    - name: task-pv-container
      image: nginx
        - containerPort: 80
          name: "http-server"
        - mountPath: "/usr/share/nginx/html"
          name: task-pv-storage
kubectl create -f task-pv-pod.yaml
# Wait for the pod to start
kubectl get pod --watch

# Verify that the volume is mounted on /usr/share/nginx/html
kubectl exec -it task-pv-pod -- df -h /usr/share/nginx/html
Filesystem                                                          Size  Used Avail Use% Mounted on
10.xx.xx.xx:/trident_pvc_3acb0d1c_b1ae_11e9_8d9f_5254004dfdb7       1.0G  256K  1.0G   1% /usr/share/nginx/html

# Delete the pod
kubectl delete pod task-pv-pod

At this point the pod (application) no longer exists but the volume is still there. You could use it from another pod if you wanted to.

To delete the volume, simply delete the claim:

kubectl delete pvc basic

Where do you go from here? you can do things like: