Upgrading Compute Clusters

You can upgrade default and additional Compute Clusters in an environment when there is a new Kubernetes version available from Cloudera Management Console and from CDP CLI.

Required role: EnvironmentAdmin or EnvironmentOwner

Before you begin

Ensure that the Cloudera environment is running and available..

Using Cloudera Management Console

Navigate to Container Service in Cloudera Management Console.
Find the compute cluster that you want to upgrade.
You can scroll through the list of compute clusters or use the search box to filter down the list of clusters.
In case there is an upgrade available for the compute cluster, the new version is indicated next to the Kubernetes Version.
Start the upgrade process by using one of the following steps:
1. Click on the cluster that you want to upgrade. You will be redirected to the Compute Cluster Details page, and select Actions > Upgrade.
2. Click on the next to the new version.
Click Ok to confirm the upgrade.

The upgrade process will be triggered. You can track the phases of the upgrade process under the Cluster Upgrade Operation.

note
In case there is any error during the upgrade process, you can view the logs by using the Download Log button. After fixing the error that caused the upgrade process failure, you can retry the upgrade process again using next to the Download Log button.
A check mark is shown at each phase, Validate, Prepare, Execute and Finalize, when the given phase is executed successfully as shown in the following example:

Using CDP CLI

Run the following command to upgrade the Compute Cluster:

cdp compute upgrade-cluster --cluster-crn [***CLUSTER CRN ***]

The above command returns an operationId that can be used to further track the upgrade progress:

cdp compute get-operation-details --operation-id [***OPERATION ID ***]
cdp compute get-operation-status --operation-id [***OPERATION ID ***]

In case the upgrade process has failed, you can use the retry-operation command with the operationId to trigger the upgrade again:

cdp compute retry-operation --operation-id [***OPERATION ID ***]