- Joining control plane nodes to a HA cluster can now be simplified by enabling the optional automatic copy of certificates from an existing control plane node.
- You can now use `kubeadm init --experimental-upload-certs` and `kubeadm join --experimental-control-plane --certificate-key`.
kubernetes/kubeadm: [#1373](https://github.com/kubernetes/kubeadm/issues/1373)

- Action-required: kubeadm: remove the deprecated feature --experimental-kustomize from kubeadm commands. The feature was replaced with --experimental-patches in 1.19. To migrate see the --help description for the --experimental-patches flag. ([#94871](https://github.com/kubernetes/kubernetes/pull/94871), [@neolit123](https://github.com/neolit123)) [SIG Cluster Lifecycle]

Consider the following changes, limitations, and guidelines before you upgrade:

* The kubelet now fails if swap is enabled on a node. To override the default and run with /proc/swaps on, set `--fail-swap-on=false`. The experimental flag `--experimental-fail-swap-on` is deprecated in this release, and will be removed in a future release.

  - The flag `--experimental-control-plane` is now deprecated in favor of  `--control-plane`. The flag `--experimental-upload-certs` is now deprecated in favor of `--upload-certs` ([...

kubelet when creating a cluster with "kubeadm init" or when joining control plane nodes with "kubeadm join".  The "kubeadm join" workflow now includes two new experimental phases called "control-plane-join-etcd" and "kubelet-wait-bootstrap" which will be used when the feature gate is enabled. This phases will be marked as non-experimental when ControlPlaneKubeletLocalMode becomes GA. During "kubeadm upgrade" commands, if the feature gate is enabled, modify the "/etc/kubernetes/kubelet.conf " to use...

    Node Allocatable. This will result in a reduction of schedulable capacity in
    clusters post upgrade where kubelet hard eviction has been turned on for
    memory. To opt-out set `--experimental-allocatable-ignore-eviction=true`.
  * More details on these feature here:
    https://kubernetes.io/docs/tasks/administer-cluster/reserve-compute-resources/

- Remove the deprecated flag `--experimental-check-node-capabilities-before-mount`. With CSI now GA, there is a better alternative. Remove any use of  `--experimental-check-node-capabilities-before-mount` from your kubelet scripts or manifests. ([#104732](https://github.com/kubernetes/kubernetes/pull/104732), [@mengjiao-liu](https://github.com/mengjiao-liu))...

- kube-controller-manager
  - The deprecated `--insecure-experimental-approve-all-kubelet-csrs-for-group` flag has been removed.
- kubelet

* Fixed 'Schedulercache is corrupted' error in kube-scheduler ([#55262](https://github.com/kubernetes/kubernetes/pull/55262), [@liggitt](https://github.com/liggitt))
* Fix hyperkube kubelet --experimental-dockershim ([#55335](https://github.com/kubernetes/kubernetes/pull/55335), [@ivan4th](https://github.com/ivan4th))

GlobalMemoryStatusEx to get total physical memory on windows nodes. ([#57124](https://github.com/kubernetes/kubernetes/pull/57124), [@JiangtianLi](https://github.com/JiangtianLi))

* Windows containers now support experimental Hyper-V isolation by setting annotation `experimental.windows.kubernetes.io/isolation-type=hyperv` and feature gates HyperVContainer. At the moment this function only supports one container per pod. ([#58751](https://github.com/kubernetes/kubernetes/pull/58751), [@feisk...