Upgrade
PaaS upgrade workflow
Request
The update of a Kubernetes PaaS can be done in self-service on the Kubernetes cluster's resource page in ITCare
It is recommended that you upgrade your non-production environments first in order to estimate the downtime generated by the operation and to test your applications using the new engine version.
Process
The Kubernetes cluster upgrade follows these steps through a Rolling Update process :
Pods reboot one node after another : 10% of the nodes at a time
Post-upgrade validation (health checks, monitoring).
Impacts
Some API versions can be obsolete or even deleted during a Kubernetes upgrade. There is a risk of breakage if you have resources with an API version that has been removed in the new version of Kubernetes.
To avoid this issue, you can check compatibility with the tool "kubent".
Kubent checks obsolete objects on the Kubernetes cluster. You need to migrate/change the identified resources before upgrading Kubernetes.
Check compatibility before a Kubernetes upgrade
To install kubent :
sh -c "$(curl -sSL 'https://git.io/install-kubent')"To identify obsolete objects that will be deleted in the next Kubernetes version :
kubent [--context my-cluster]An output exemple :
...
__________________________________________________________________________________________
>>> Deprecated APIs removed in 1.22 <<<
------------------------------------------------------------------------------------------
KIND NAMESPACE NAME API_VERSION REPLACE_WITH (SINCE)
Ingress <undefined> toto networking.k8s.io/v1beta1 networking.k8s.io/v1 (1.19.0)
...In this tutorial, if your cluster has an upgrade planned to the Kubernetes version 1.22, you need to migrate the ingress resource named "toto" of the API version networking.k8s.io/v1beta1 to networking.k8s.io/v1 before the upgrade.
This migration can imply a change of more fields in the resource. Please check the official documentation :
Kubent could fail to report some information, for instance the ingress namespace, you can report this issue to the publisher : https://github.com/doitintl/kube-no-trouble/issues
Upgrade the Kubernetes cluster
On the top of your cluster's page, click on the Manage button, then Update. An information popup will be displayed, click on Sumit
The upgrade can take several minutes depending of the size of the cluster.
Time references
A Kubernetes cluster upgrade typically takes around 15 minutes on average. Nodes will be upgraded one after another with no more than 10% of the cluster's node at the same time
Note that these times are estimates and can vary depending on the cluster configuration
OS / Kubernetes support matrix
Distribution Linux supported by cegedim.cloud depending of the Kubernetes version :
Kubernetes 1.28
Ubuntu 22.04, 24.04
Kubernetes 1.30
Ubuntu 22.04, 24.04
Kubernetes 1.32
Ubuntu 20.04, 22.04, 24.04
Supported Kubernetes update paths
Kubernetes 1.28
✅
Kubernetes 1.28 → Kubernetes 1.30
✅
Kubernetes 1.28 → Kubernetes 1.30 → Kubernetes 1.31
✅
Kubernetes 1.28 → Kubernetes 1.30 → Kubernetes 1.31 → Kubernetes 1.32
Kubernetes 1.30
✅
Kubernetes 1.30 → Kubernetes 1.31
✅
Kubernetes 1.30 → Kubernetes 1.31 → Kubernetes 1.32
Kubernetes 1.31
✅
Kubernetes 1.31
✅
Kubernetes 1.31 → Kubernetes 1.32
Kubernetes 1.32
✅
Kubernetes 1.31 → Kubernetes 1.32
Last updated