Known Issues
- Clusters provisioned with Kubernetes 1.22 or upgraded from 1.21 to 1.22 using KubeOne 1.3.0-alpha.1 use a metrics-server version incompatible with Kubernetes 1.22. This might cause issues with deleting Namespaces that manifests by the Namespace being stuck in the Terminating state. This can be fixed by upgrading to KubeOne 1.3.0-rc.0 and running
kubeone apply
.
Added
- Add support for Kubernetes 1.22 (#1447, #1456)
- Add support for the kubeadm v1beta3 API. The kubeadm v1beta3 API is used for all Kubernetes 1.22+ clusters. (#1457)
Changed
Fixed
- Fix adding second container to the machine-controller-webhook Deployment (#1433)
- Extend restart API server script to handle failing
crictl logs
due to missing symlink. This fixes the issue withkubeone apply
failing to restart the API server containers when provisioning or upgrading the cluster (#1448)
Updated
- Update Go to 1.16.7 (#1441)
- Update machine-controller to v1.35.1 (#1440)
- Update Hetzner CCM to v1.9.1 (#1428)
- Add
HCLOUD_LOAD_BALANCERS_USE_PRIVATE_IP=true
to the environment if the network is configured
- Add
- Update DigitalOcean CCM to v0.1.33 (#1429)
Terraform Configs
- Inherit the firmware settings from the template VM in the Terraform configs for vSphere (#1445)