* Fix a bug in kubelet hostport logic which flushes KUBE-MARK-MASQ iptables chain ([#32413](https://github.com/kubernetes/kubernetes/pull/32413), [@freehan](https://github.com/freehan))
* Stick to 2.2.1 etcd ([#32404](https://github.com/kubernetes/kubernetes/pull/32404), [@caesarxuchao](https://github.com/caesarxuchao))
* Use etcd 2.3.7 ([#32359](https://github.com/kubernetes/kubernetes/pull/32359), [@wojtek-t](https://github.com/wojtek-t))
* AWS: Change default networking for kube-up to kubenet ([#32239](https://github.com/kubernetes/kubernetes/pull/32239), [@zmerlynn](https://github.com/zmerlynn))
* Make sure finalizers prevent deletion on storage that supports graceful deletion ([#32351](https://github.com/kubernetes/kubernetes/pull/32351), [@caesarxuchao](https://github.com/caesarxuchao))
* Some components like kube-dns and kube-proxy could fail to load the service account token when started within a pod. Properly handle empty configurations to try loading the service account config. ([#31947](https://github.com/kubernetes/kubernetes/pull/31947), [@smarterclayton](https://github.com/smarterclayton))
* Use federated namespace instead of the bootstrap cluster's namespace in Ingress e2e tests. ([#32105](https://github.com/kubernetes/kubernetes/pull/32105), [@madhusudancs](https://github.com/madhusudancs))