Starting with Konvoy
v1.6.0, Helm v2 is no longer installed or used to deploy Konvoy Addons because Helm v2 is deprecated due to security and operational issues. Helm v3 is installed and is the primary mechanism for installation and maintenance of Konvoy Addons.
Upgrading from earlier versions of Konvoy includes an automatic migration process that requires no user input to migrate your Konvoy Addons to Helm v3. During this process, the Tiller (server) for Helm v2, and any charts deployed with Helm v2, are not removed, but remains on your cluster. We recommend you manually remove the Helm v2 Tiller due to security issues, using the process below.
Before you begin
Ensure that you have Kubectl installed and configured for each Konvoy cluster.
Remove Helm v2 tiller
Remove Helm v2 Tiller by enter the following command:
kubectl -n kube-system delete deployment tiller-deploy
This command shuts down Tiller gracefully and removes all its resources.
If you have the prometheus addon enabled, you must delete the addon before upgrading.
Before you begin the upgrade, run the following command to delete the addon:
kubectl delete addon prometheus --namespace kubeaddons --wait
If you’ve already run the upgrade and deploying the prometheus addon has failed, run this command:
helm2 delete --purge prometheus-kubeaddons
This assumes helm2 CLI has been installed. If it hasn’t, you will need to install it. If you’ve already ran the Konvoy upgrade, the Addon resource has been updated. After you’ve deleted this old Helm data from the above command and the failing helm release has been removed, kubeaddons can successfully reconcile the Addon.
If you want to verify the status of your prometheus addon state, you can monitor that with the following:
kubectl --namespace kubeaddons get addon prometheus --watch
The migration of a helm release from Helm v2 to Helm v3 adds additional metadata to the release data structure. This additional data causes the Prometheus release data to exceed the maximum size of a ConfigMap or Secret. Newer versions of the Prometheus chart reduces the amount of data held in this release structure, but the conversion will fail preventing the newer Prometheus chart installation from being reached. Removing the oversized data from Kubernetes allows the replacement to continue.
For information on related topics or procedures, refer to the following: