Full Management Cluster Migration to Oracle Linux 8
Software release 3.0.2-b1392231 migrates the entire Private Cloud Appliance to Oracle Linux 8. It affects not only the operating system of components, but also the platform layer, container images, microservices, firmware, and RPM packages. With the new version of Oracle Linux, several improvements in the upgrade process are introduced, as well as some extra requirements.
- System backup and cluster teardown
-
The host OS is technically not upgraded but installed from scratch. First, a backup of critical data and files is created on the shared storage in the ZFS Storage Appliance, so that the system configuration can be restored at the end of the upgrade workflow. One by one, the management nodes must be removed from the clusters they belong to, so that both the nodes and the clusters can be migrated to Oracle Linux 8.
- Fresh installation from image
-
Each management node is rebooted from a new image, which allows a new disk layout and partitioning to be applied. When the OS and all other required tools are installed in the new partitions, the backups are restored and the node joins the new clusters (PCS, MySQL, Kubernetes). Checks are performed to ensure that the node can resume all its functions. At this point, all platform and infrastructure services are on Oracle Linux 8.
- Cluster reconstruction
-
To reduce complexity and risk of failure, the management nodes are migrated to Oracle Linux 8 in a fixed order: pcamn01, pcamn02, pcamn03. If the first node owns the cluster virtual IP, it is reassigned to another node. When the upgrade workflow is completed, the first node takes back the virtual IP, joins the new cluster, and waits for the other nodes to complete the new cluster setup.
The full management node cluster workflow includes these components:
-
Host operating system
-
Clustered MySQL database
-
Secret Service (including Etcd and Vault)
-
Kubernetes container orchestration packages
-
Containerized microservices
-
Oracle Cloud Infrastructure images
Caution:
When the new images have been imported, a background job is launched to ensure that running OKE clusters receive the latest available CVE fixes delivered with the new images. For more information, see Upgrading Oracle Cloud Infrastructure Images.