- QueueingHint implementation for NodeAffinity is reverted because we found potential scenarios where events that make Pods schedulable could be missed. ([#122285](https://github.com/kubernetes/kubernetes/pull/122285), [@sanposhiho](https://github.com/sanposhiho)) [SIG Scheduling]

- QueueingHint implementation for NodeAffinity is reverted because we found potential scenarios where events that make Pods schedulable could be missed. ([#122327](https://github.com/kubernetes/kubernetes/pull/122327), [@sanposhiho](https://github.com/sanposhiho)) [SIG Scheduling]

### Other notable changes

* CustomResourceDefinitions: OpenAPI v3 validation schemas containing `$ref`references are no longer permitted (valid references could not be constructed previously because property ids were not permitted either). Before upgrading, ensure CRD definitions do not include those `$ref` fields. ([#58438](https://github.com/kubernetes/kubernetes/pull/58438), [@carlory](https://github.com/carlory))

#### Apps

- Hyperkube short aliases have been removed from source code, because hyperkube docker image currently creates these aliases. ([#76953](https://github.com/kubernetes/kubernetes/pull/76953), [@Rand01ph](https://github.com/Rand01ph))

#### Auth

 available, because it's incompatible with `kubectl
 apply` and strategic merge patch. You must explicitly set the spec.selector value
 in your manifest. An object with a spec.selector value that does not match the labels in
 its spec.template is invalid.

 - Selector mutation is disabled for all kinds in the
 app/v1beta2 group version, because the controllers in the workloads API do not handle

* Do not detach volume if mount in progress ([#71145](https://github.com/kubernetes/kubernetes/pull/71145), [@gnufied](https://github.com/gnufied))
* Only use the first IP address got from instance metadata. This is because Azure CNI would set up a list of IP addresses in instance metadata, while only the first one is the Node's IP. ([#71736](https://github.com/kubernetes/kubernetes/pull/71736), [@feiskyer](https://github.com/feiskyer))

* Fix a potential bug that scheduler preempts unnecessary pods. ([#70898](https://github.com/kubernetes/kubernetes/pull/70898), [@Huang-Wei](https://github.com/Huang-Wei))
* Only use the first IP address got from instance metadata. This is because Azure CNI would set up a list of IP addresses in instance metadata, while only the first one is the Node's IP. ([#71736](https://github.com/kubernetes/kubernetes/pull/71736), [@feiskyer](https://github.com/feiskyer))

- 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]

* Only use the first IP address got from instance metadata. This is because Azure CNI would set up a list of IP addresses in instance metadata, while only the first one is the Node's IP. ([#71736](https://github.com/kubernetes/kubernetes/pull/71736), [@feiskyer](https://github.com/feiskyer))

### Bug or Regression

- Cherry-pick #115769: Fix the problem Pod terminating stuck because of trying to umount not actual mounted dir. ([#119831](https://github.com/kubernetes/kubernetes/pull/119831), [@cartermckinnon](https://github.com/cartermckinnon)) [SIG Node and Storage]