Upgrading a Data Hub with the CDP CLI

You can initiate a Data Hub upgrade (either OS, Runtime, or both) with the CDP CLI. Using the same CLI command, you can also search for and validate available images to upgrade to, and generate JSON templates for specific upgrade scenarios.

Obtain image ID

If your Data Lake upgrade includes upgrading from CentOS to RHEL 8, prior to attempting an upgrade you need to obtain an ID of a target RHEL 8 image. You can obtain it from the image catalog by finding an image with your target Runtime version which has an OS Type of RHEL8.

Once you have identified the ID, you can provide it in the upgrade CLI command by using the --image-id flag.

Upgrade steps

The cdp datahub upgrade-cluster command includes the following options:

cdp datahub upgrade-cluster 

          --cluster-name <value>
          [--image-id <value>]
          [--runtime <value>]
          [--lock-components | --no-lock-components]
          [--dry-run | --no-dry-run]
          [--show-available-images | --no-show-available-images]
          [--show-available-image-per-runtime | --no-show-available-image-per-runtime]
          [--cli-input-json <value>]
          [--generate-cli-skeleton]

In order to use this command for upgrading from CentOS to RHEL, ensure to provide an image ID of a RHEL 8 image.

Table 1.
Option Description
--cluster-name (string) Required. The name or CRN of the Data Hub to upgrade.
--image-id (string) The ID of an image to upgrade to. If upgrading from CentOS to RHEL, ensure to provide a RHEL 8 image ID.
–runtime (string) The Runtime version to upgrade to. When you specify the Runtime version, the upgrade uses the latest image ID of the given Runtime version from the same image catalog used for Data Hub creation.
--lock-components | --no-lock-components (boolean) Use --lock components to perform an OS upgrade only.
--dry-run | --no-dry-run (boolean) Checks the eligibility of an image to upgrade. Can be used in conjunction with any other parameter, returning the available image (with respect to image Id, Runtime or lock-components set) without performing any actions.
--show-available-images | --no-show-available-images (boolean) Returns the list of images that are eligible to upgrade to.
--show-latest-available-image-per-runtime | --no-show-latest-available-image-per-runtime (boolean) Returns the latest image that is eligible to upgrade to, for each Runtime version with at least one available upgrade candidate.

When you run the cdp datahub upgrade-cluster command to initiate an upgrade, you have one of three options:

  1. Specify one of either --image-id, --runtime, or --lockComponents, which makes an explicit choice of the exact image, Runtime (latest OS), or latest OS (same Runtime) for upgrade.
  2. Specify both --image-id and --lockComponents,which specifies an image and ensures the image represents an OS only upgrade.
  3. Specify none of the --image-id, --runtime, or --lockComponents parameters, which initiates a Runtime/CM upgrade to the latest compatible version and OS image.

Outside of upgrade, you can use the following options:

--show-available-images/--no-show-available-images 
--show-available-images-per-runtime/--no-show-available-images-per-runtime
 --dry-run