Upgrading SEP

First, ensure your Kubernetes context is set to the cluster you intend to upgrade. This is an important step to direct your commands to the correct cluster. Execute the command below, replacing %NAMESPACE% with your specific namespace.

If you're uncertain about your namespace, the kubectl get namespaces command will list all available namespaces for your reference.

kubectl config set-context --current --namespace=%NAMESPACE%

After executing this command, all subsequent kubectl operations will be directed towards the specified cluster.

The helm upgrade command is an important tool for applying changes to your SEP cluster, especially when you've modified settings in your custom YAML files. Each alteration in your configuration necessitates the use of helm upgrade to implement these changes effectively.

To proceed with upgrading your SEP cluster, use the following command and change %SEP_VERSION% to the version you wish to upgrade to:

helm upgrade sep-cluster starburstdata/starburst-enterprise \
  --install \
  --version %SEP_VERSION% \
  --values ./registry-access.yaml \
  --values ./sep-prod-setup.yaml \
  --values ./sep-prod-catalogs.yaml

After running the upgrade command, you can monitor the progress of the upgrade by reading the Kubernetes logs. To do this, use the following command to get all kubernetes pods:

kubectl get pods

The output will look similar to the following:

NAME                                     READY   STATUS    RESTARTS   AGE
coordinator-XXX                          1/1     Running   0          1m
worker-XXX                               1/1     Running   0          1m

To read the logs of a specific pod, use the following command, replacing %POD_NAME% with the name of the pod you wish to read. In the above example, the name of the coordinator pod would be coordinator-XXX.

This is a useful command to find an errors that may have occurred during the upgrade process. Look for the Events section at the bottom of the output to see if any errors have occurred.

Generally the coordinator pod is the most important to check for errors.

kubectl describe pods %POD_NAME%

It is also possible to get service logs from kubernetes which will provide further details such as the cluster IP address, external IP addresses and ports. To see that, run the following command:

kubectl get svc

The output will look similar to the following:

NAME       TYPE       CLUSTER-IP      EXTERNAL-IP   PORT(S)        AGE
starburst  ClusterIP  111.11.111.111  <None>        8080:8080/TCP  1m