Chapter 1. Overview of upgrading OpenShift AI
Red Hat OpenShift AI is automatically updated as new release or versions become available. Currently, no administrator action is necessary to trigger the process.
When an OpenShift AI upgrade occurs, you should complete the Requirements for upgrading OpenShift AI.
Notes:
- Before you can use an accelerator in OpenShift AI, your instance must have the associated accelerator profile. If your OpenShift cluster instance has an accelerator, its accelerator profile is preserved after the upgrade. For more information about accelerators, see Working with accelerators.
- Notebook images are integrated into the image stream during the upgrade and subsequently appear in the OpenShift AI dashboard. Notebook images are constructed externally; they are prebuilt images that undergo quarterly changes and they do not change with every OpenShift AI upgrade.
After you upgrade to OpenShift AI with data science pipelines 2.0, pipelines created with data science pipelines 1.0 continue to run, but are inaccessible from the OpenShift AI dashboard. If you are a current data science pipelines user, do not upgrade to OpenShift AI with data science pipelines 2.0 until you are ready to migrate to the new pipelines solution.
Data science pipelines 2.0 contains an installation of Argo Workflows. OpenShift AI does not support direct customer usage of this installation of Argo Workflows. To upgrade to OpenShift AI with data science pipelines 2.0, ensure that no separate installation of Argo Workflows exists on your cluster.
