Upgrade with the operator

01/28/2023 Contributors

PDFs

You can easily upgrade an existing Astra Trident installation using the operator.

What you'll need

To upgrade by using the operator, the following conditions should be met:

You should have a CSI-based Astra Trident installation. To check if you are running CSI Trident, examine the pods in your Trident namespace. If they follow the trident-csi-* naming pattern, you are running CSI Trident.
You should have a CRD-based Trident installation. This represents all releases from 19.07 and later. If you have a CSI-based installation, you most likely have a CRD-based installation.
If you have uninstalled CSI Trident and the metadata from the installation persists, you can upgrade by using the operator.
Only one Astra Trident installation should exist across all the namespaces in a given Kubernetes cluster.
You should be using a Kubernetes cluster that runs a supported Kubernetes version.
If alpha snapshot CRDs are present, you should remove them with tridentctl obliviate alpha-snapshot-crd. This deletes the CRDs for the alpha snapshot spec. For existing snapshots that should be deleted/migrated, see this blog.

When upgrading Trident by using the operator on OpenShift Container Platform, you should upgrade to Trident 21.01.1 or later. The Trident operator released with 21.01.0 contains a known issue that has been fixed in 21.01.1. For more details, see the issue details on GitHub.

Upgrade a cluster-scoped operator installation

Follow these steps to upgrade from Trident 21.01 and later.

Steps

Delete the Trident operator that was used to install the current Astra Trident instance. For example, if you are upgrading from 21.01, run the following command:
```
kubectl delete -f 22.01/trident-installer/deploy/BUNDLE.YAML -n trident
```
Optionally, you can edit the TridentOrchestrator object that you created when installing Trident to modify the installation parameters. This can include changes such as modifying the custom Trident image, private image registry to pull container images from, enabling debug logs, or specifying image pull secrets.
Install Astra Trident by using the correct bundle YAML file for your environment and Astra Trident version from https://github.com/NetApp/trident/tree/stable/vXX.XX/deploy/BUNDLE.YAML where vXX.XX is the version number (for example v22.10) and BUNDLE.YAML is the bundle YAML file name.
For clusters running Kubernetes 1.24 or lower, use bundle_pre_1_25.yaml.

For clusters running Kubernetes 1.25 or higher, use bundle_post_1_25.yaml.
For example, if you are installing Astra Trident 22.10 for Kubernetes 1.25, run the following command:
```
kubectl create -f 22.10.0/trident-installer/deploy/bundle_post_1_25.yaml -n trident
```
As part of this step, the Trident operator will identify an existing Astra Trident installation and upgrade it to the same version as the operator.

Upgrade a namespace-scoped operator installation

To upgrade from an instance of Astra Trident installed using the namespace-scoped operator (versions 20.07 through 20.10), here is the set of steps to be followed:

Steps

Verify the status of the existing Trident installation. To do this, check the Status of TridentProvisioner. The status should be Installed.
```
kubectl describe tprov trident -n trident | grep Message: -A 3
Message:  Trident installed
Status:   Installed
Version:  v20.10.1
```
If status shows Updating, ensure you resolve it before proceeding. For a list of possible status values, see here.

Create the TridentOrchestrator CRD by using the manifest provided with the Trident installer.

# Download the release required [22.10.0]
mkdir 22.10.0
cd 22.10.0
wget https://github.com/NetApp/trident/releases/download/v22.10.0/trident-installer-22.10.0.tar.gz
tar -xf trident-installer-22.10.0.tar.gz
cd trident-installer
kubectl create -f deploy/crds/trident.netapp.io_tridentorchestrators_crd_post1.16.yaml

Delete the namespace-scoped operator by using its manifest. To complete this step, you need the bundle YAML file used to deploy the namespace-scoped operator from https://github.com/NetApp/trident/tree/stable/vXX.XX/deploy/BUNDLE.YAML where vXX.XX is the version number (for example v22.10) and BUNDLE.YAML is the bundle YAML file name.

You should make the necessary changes to the Trident install parameters (for example, changing the values for tridentImage, autosupportImage, private image repository, and providing imagePullSecrets) after deleting the namespace-scoped operator and before installing the cluster-scoped operator. For a complete list of parameters that can be updated, refer to the configuration options.

#Ensure you are in the right directory
pwd
/root/20.10.1/trident-installer

#Delete the namespace-scoped operator
kubectl delete -f deploy/<BUNDLE.YAML>
serviceaccount "trident-operator" deleted
clusterrole.rbac.authorization.k8s.io "trident-operator" deleted
clusterrolebinding.rbac.authorization.k8s.io "trident-operator" deleted
deployment.apps "trident-operator" deleted
podsecuritypolicy.policy "tridentoperatorpods" deleted

#Confirm the Trident operator was removed
kubectl get all -n trident
NAME                               READY   STATUS    RESTARTS   AGE
pod/trident-csi-68d979fb85-dsrmn   6/6     Running   12         99d
pod/trident-csi-8jfhf              2/2     Running   6          105d
pod/trident-csi-jtnjz              2/2     Running   6          105d
pod/trident-csi-lcxvh              2/2     Running   8          105d

NAME                  TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)              AGE
service/trident-csi   ClusterIP   10.108.174.125   <none>        34571/TCP,9220/TCP   105d

NAME                         DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR                                     AGE
daemonset.apps/trident-csi   3         3         3       3            3           kubernetes.io/arch=amd64,kubernetes.io/os=linux   105d

NAME                          READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/trident-csi   1/1     1            1           105d

NAME                                     DESIRED   CURRENT   READY   AGE
replicaset.apps/trident-csi-68d979fb85   1         1         1       105d

At this stage, the trident-operator-xxxxxxxxxx-xxxxx pod is deleted.

(Optional) If the install parameters need to be modified, update the TridentProvisioner spec. These could be changes such as modifying the private image registry to pull container images from, enabling debug logs, or specifying image pull secrets.
```
kubectl patch tprov <trident-provisioner-name> -n <trident-namespace> --type=merge -p '{"spec":{"debug":true}}'
```

Install the cluster-scoped operator.

Installing the cluster-scoped operator initiates the migration of TridentProvisioner objects to TridentOrchestrator objects, deletes TridentProvisioner objects and the tridentprovisioner CRD, and upgrades Astra Trident to the version of the cluster-scoped operator being used. In the example that follows, Trident is upgraded to 22.10.0.

Upgrading Astra Trident by using the cluster-scoped operator results in the migration of tridentProvisioner to a tridentOrchestrator object with the same name. This is automatically handled by the operator. The upgrade will also have Astra Trident installed in the same namespace as before.

#Ensure you are in the correct directory
pwd
/root/22.10.0/trident-installer

#Install the cluster-scoped operator in the **same namespace**
kubectl create -f deploy/<BUNDLE.YAML>
serviceaccount/trident-operator created
clusterrole.rbac.authorization.k8s.io/trident-operator created
clusterrolebinding.rbac.authorization.k8s.io/trident-operator created
deployment.apps/trident-operator created
podsecuritypolicy.policy/tridentoperatorpods created

#All tridentProvisioners will be removed, including the CRD itself
kubectl get tprov -n trident
Error from server (NotFound): Unable to list "trident.netapp.io/v1, Resource=tridentprovisioners": the server could not find the requested resource (get tridentprovisioners.trident.netapp.io)

#tridentProvisioners are replaced by tridentOrchestrator
kubectl get torc
NAME      AGE
trident   13s

#Examine Trident pods in the namespace
kubectl get pods -n trident
NAME                                READY   STATUS    RESTARTS   AGE
trident-csi-79df798bdc-m79dc        6/6     Running   0          1m41s
trident-csi-xrst8                   2/2     Running   0          1m41s
trident-operator-5574dbbc68-nthjv   1/1     Running   0          1m52s

#Confirm Trident has been updated to the desired version
kubectl describe torc trident | grep Message -A 3
Message:                Trident installed
Namespace:              trident
Status:                 Installed
Version:                v22.10.0

Upgrade a Helm-based operator installation

Perform the following steps to upgrade a Helm-based operator installation.

Steps

Download the latest Astra Trident release.
Use the helm upgrade command. See the following example:
```
helm upgrade <name> trident-operator-22.10.0.tgz
```
where trident-operator-22.10.0.tgz reflects the version that you want to upgrade to.
Run helm list to verify that the chart and app version have both been upgraded.

To pass configuration data during the upgrade, use --set.

For example, to change the default value of tridentDebug, run the following command:

helm upgrade <name> trident-operator-22.10.0-custom.tgz --set tridentDebug=true

If you run tridentctl logs, you can see the debug messages.

If you set any non-default options during the initial installation, ensure that the options are included in the upgrade command, or else, the values will be reset to their defaults.

Upgrade from a non-operator installation

If you have a CSI Trident instance that meets the prerequisites listed above, you can upgrade to the latest release of the Trident operator.

Steps

Download the latest Astra Trident release.

# Download the release required [22.10.0]
mkdir 22.10.0
cd 22.10.0
wget https://github.com/NetApp/trident/releases/download/v22.10.0/trident-installer-22.10.0.tar.gz
tar -xf trident-installer-22.10.0.tar.gz
cd trident-installer

Create the tridentorchestrator CRD from the manifest.

kubectl create -f deploy/crds/trident.netapp.io_tridentorchestrators_crd_post1.16.yaml

Deploy the operator.

#Install the cluster-scoped operator in the **same namespace**
kubectl create -f deploy/<BUNDLE.YAML>
serviceaccount/trident-operator created
clusterrole.rbac.authorization.k8s.io/trident-operator created
clusterrolebinding.rbac.authorization.k8s.io/trident-operator created
deployment.apps/trident-operator created
podsecuritypolicy.policy/tridentoperatorpods created

#Examine the pods in the Trident namespace
NAME                                READY   STATUS    RESTARTS   AGE
trident-csi-79df798bdc-m79dc        6/6     Running   0          150d
trident-csi-xrst8                   2/2     Running   0          150d
trident-operator-5574dbbc68-nthjv   1/1     Running   0          1m30s

Create a TridentOrchestrator CR for installing Astra Trident.

#Create a tridentOrchestrator to initate a Trident install
cat deploy/crds/tridentorchestrator_cr.yaml
apiVersion: trident.netapp.io/v1
kind: TridentOrchestrator
metadata:
  name: trident
spec:
  debug: true
  namespace: trident

kubectl create -f deploy/crds/tridentorchestrator_cr.yaml

#Examine the pods in the Trident namespace
NAME                                READY   STATUS    RESTARTS   AGE
trident-csi-79df798bdc-m79dc        6/6     Running   0          1m
trident-csi-xrst8                   2/2     Running   0          1m
trident-operator-5574dbbc68-nthjv   1/1     Running   0          5m41s

#Confirm Trident was upgraded to the desired version
kubectl describe torc trident | grep Message -A 3
Message:                Trident installed
Namespace:              trident
Status:                 Installed
Version:                v22.10.0

The existing backends and PVCs are automatically available.

Upgrade with the operator

Creating your file...

Upgrade a cluster-scoped operator installation

Upgrade a namespace-scoped operator installation

Upgrade a Helm-based operator installation

Upgrade from a non-operator installation