- DRA drivers: the resource slice controller sometimes didn't react properly when kubelet or someone else deleted a recently created ResourceSlice. It incorrectly assumed that the ResourceSlice still exists and didn't recreate it. ([#132683](https://github.com/kubernetes/kubernetes/pull/132683), [@pohly](https://github.com/pohly)) [SIG Apps, Node and Testing]

* Cluster federation servers have changed the location in etcd where federated services are stored, so existing federated services must be deleted and recreated. Before upgrading, export all federated services from the federation server and delete the services. After upgrading the cluster, recreate the federated services from the exported data. ([#37770](https://github.com/kubernetes/kubernetes/pull/37770), [@enj](https://github.com/enj))

- Headless Services with no selector which were created without dual-stack enabled will be defaulted to RequireDualStack instead of PreferDualStack.  This is consistent with such Services which are created with dual-stack enabled. ([#104986](https://github.com/kubernetes/kubernetes/pull/104986), [@thockin](https://github.com/thockin)) [SIG Network]

  - Certificates created with kubeadm now have improved uniqueness of Distinguished Name fields.
- HA improvements:
  - `kubeadm join --experimental-control-plane` can now be used to join control plane instances to an existing cluster.
  - `kubeadm upgrade node experimental-control-plane` can now be used for upgrading secondary control plane instances created with `kubeadm join --experimental-control-plane`.

* Fixes an issue with kubelet-reported pod status on deleted/recreated pods. ([#86320](https://github.com/kubernetes/kubernetes/pull/86320), [@liggitt](https://github.com/liggitt))

* The `--bootstrap-kubeconfig` argument to Kubelet previously created the first bootstrap client credentials in the certificates directory as `kubelet-client.key` and `kubelet-client.crt`.  Subsequent certificates created by cert rotation were created in a combined PEM file that was atomically rotated as `kubelet-client-DATE.pem` in that directory, which meant clients relying on the `node.kubeconfig` generated...

- Failed pods associated with a job with `parallelism = 1` are recreated by the job controller honoring exponential backoff delay again. However, for jobs with `parallelism > 1`, pods might be created without exponential backoff delay. ([#115022](https://github.com/kubernetes/kubernetes/pull/115022), [@nikhita](https://github.com/nikhita)) [SIG Apps]

    ```
    * PV creation Logs:

- Fixed: during the kube-controller-manager restart, when the corresponding Endpoints resource was manually deleted and recreated, causing the endpointslice to fail to be created normally. ([#125359](https://github.com/kubernetes/kubernetes/pull/125359), [@yangjunmyfm192085](https://github.com/yangjunmyfm192085)) [SIG Apps and Network]

- Number of errors reported to the metric `storage_operation_duration_seconds_count` for emptyDir decreased significantly because previously one error was reported for each projected volume created. ([#117022](https://github.com/kubernetes/kubernetes/pull/117022), [@mpatlasov](https://github.com/mpatlasov)) [SIG Storage]