docs/doc/source/dist_cloud/kubernetes/upgrade-management-overview.rst

Upgrade Management Overview

You can upgrade 's 's System Controller, and subclouds with a new release of software.

Note

Backup all yaml files that are updated using the Redfish Platform Management service. For more information, see installing-a-subcloud-using-redfish-platform-management-service. For information on back up and restore, see backing-up-starlingx-system-data.

You can use the to manage upgrades. The workflow for upgrades is as follows:

1. To upgrade the system, you must first upgrade the System Controller. See upgrading-the-systemcontroller-using-the-cli.
2. Use Upgrade Orchestration to upgrade the subclouds. See distributed-upgrade-orchestration-process-using-the-cli.
3. To handle errors during an orchestrated upgrade, see robust-error-handling-during-an-orchestrated-upgrade.

For deployment configurations, the end user container images in registry.local will be backed up during the upgrade process. This only includes images other than system and application images. These images are limited to 5 GB in total size. If the system contains more than 5 GB of these images, the upgrade start will fail. For information on back up and restore, see backing-up-starlingx-system-data.

The following prerequisites apply to a upgrade management service.

• Validate the list of new images with the target release. If you are using a private registry for installs/upgrades, you must populate your private registry with the new images prior to bootstrap and/or patch application.
• Configuration Verification: Ensure that the following configurations are verified before you proceed with the upgrade on the and subclouds:
  • Run the system application-list command to ensure that all applications are running.
  • Run the system host-list command to list the configured hosts.
  • Run the dcmanager subcloud list command to list the subclouds.
  • Run the kubectl get pods --all-namespaces command to test that the authentication token validates correctly.
  • Run the fm alarm-list command to ensure that there are no unexpected or management-affecting alarms.
  • Run the system health-query-upgrade command to check the system health to ensure that the system meets all the upgrade prerequisites.
  • Run the kubectl get host -n deployment command to ensure all nodes in the cluster have reconciled and is set to 'true'.
  • Ensure controller-0 is the active controller.
• The subclouds must use the Redfish platform management service.
• Ensure any certificates managed by cert manager will not be renewed during the upgrade process.

partner