* Fixed detach of already detached CSI volumes. ([#63295](https://github.com/kubernetes/kubernetes/pull/63295), [@jsafrane](https://github.com/jsafrane))
* fix rbd device works at block mode not get mapped to container  ([#64555](https://github.com/kubernetes/kubernetes/pull/64555), [@wenlxie](https://github.com/wenlxie))

- Fixed an EndpointSlice Controller hashing bug that could cause EndpointSlices to incorrectly handle Pods with duplicate IP addresses. For example this could happen when a new Pod reused an IP that was also assigned to a Pod in a completed state. ([#115907](https://github.com/kubernetes/kubernetes/pull/115907), [@qinqon](https://github.com/qinqon)) [SIG Apps and Network]

- Overall, enable the feature of `PreferNominatedNode` will  improve the performance of scheduling where preemption might frequently happen, but in theory, enable the feature of `PreferNominatedNode`, the pod might not be scheduled to the best candidate node in the cluster. ([#93179](https://github.com/kubernetes/kubernetes/pull/93179), [@chendave](https://github.com/chendave)) [SIG Scheduling...

- Improved CRI handling
  - crictl is no longer required in docker-only setups.
  - Better detection of installed CRI.
  - Better output for image pull errors.
- Improved air-gapped and offline support
  - kubeadm now handles air-gapped environments by using the local client version as a fallback.
  - Some kubeadm commands are now allowed to work in a completely offline mode.
- Certificate handling improvements:

- Scheduler bugfix: Scheduler doesn't lose pod information when nodes are quickly recreated. This could happen when nodes are restarted or quickly recreated reusing a nodename. ([#93964](https://github.com/kubernetes/kubernetes/pull/93964), [@alculquicondor](https://github.com/alculquicondor)) [SIG Scheduling and Testing]

    - https://github.com/kubernetes/kubernetes/issues/92250
  
    To address this, the new debian-hyperkube-base image (v1.1.0) uses the
    debian-iptables base image (v12.1.0), which includes iptables-wrapper, a
    script used to determine the correct iptables mode to run in. ([#92494](https://github.com/kubernetes/kubernetes/pull/92494), [@justaugustus](https://github.com/justaugustus)) [SIG Cluster Lifecycle and Release]

- Pods which set `hostNetwork: true` and declare ports, get the `hostPort` field set automatically. Previously this would happen in the PodTemplate of a Deployment, DaemonSet or other workload API.  Now `hostPort` will only be set when an actual Pod is being created.  If this presents a problem, setting the feature gate "DefaultHostNetworkHostPortsInPodTemplates" to true will...

* Fix pods failing to start when subPath is a dangling symlink from kubelet point of view, which can happen if it is running inside a container ([#48555](https://github.com/kubernetes/kubernetes/pull/48555), [@redbaron](https://github.com/redbaron))

because it frontends equivalent mirrors in other clouds.  Please point your clusters to the new registry going forward. \n\nAdmission/Policy integrations that have an allowlist of registries need to include `registry.k8s.io` alongside `k8s.gcr.io`.\nAir-gapped environments and image garbage-collection configurations will need to update to pre-pull and preserve required images under `registry.k8s.io` as well as `k8s.gcr.io`. ([#109938](https://github.com/kubernetes/kubernetes/pull/109938), [@dims](https...

- Scheduler bugfix: Scheduler doesn't lose pod information when nodes are quickly recreated. This could happen when nodes are restarted or quickly recreated reusing a nodename. ([#93938](https://github.com/kubernetes/kubernetes/pull/93938), [@alculquicondor](https://github.com/alculquicondor)) [SIG Scalability, Scheduling and Testing]