2.5 Upgrading 1.1.9 to 1.1.12

The following instructions are specifically for a major package upgrade from Oracle Linux Container Services for use with Kubernetes 1.1.9 to version 1.1.12.

Warning

The upgrade process that is described here only applies for the stated upgrade path on existing hosts.

Oracle does not support upgrading existing clusters between smaller errata releases by using the kubeadm-uprade.sh script. Instead, you must use the kubeadm-setup.sh script that is described in Section 2.6, “Updating to Errata Releases”.

These instructions work on hosts that are booting from UEK R4, but it is recommended that hosts currently running UEK R4 are upgraded to use UEK R5 to facilitate future upgrades, where KubeDNS is deprecated.

The upgrade process requires you to first upgrade the master node in your cluster, and then update each of the worker nodes. The upgrade of the master node is scripted so that the pre-requisite checks, validation, and reconfiguration are automated. It is a good practice to make a backup file for your cluster before upgrading. This process is described in Section 2.5.1, “Upgrading the Master Node from 1.1.9 to 1.1.12”.

After the master node is upgraded, you can upgrade each worker node, as described in Section 2.5.2, “Upgrading Worker Nodes from 1.1.9 to 1.1.12”.

When the upgrade of the cluster has completed, you must restart or redeploy any applications that were running in the cluster.

Important

Oracle does not support an upgrade from a preview release to a stable and supported release.

Oracle also does not support upgrading existing single master node clusters that are built with the kubeadm-setup.sh script to High Availability clusters. You must build and manage High Availability clusters by using the kubeadm-ha-setup utility.