## Changes by Kind

### Bug or Regression

- Fixed a bug that a pod may remain unscheduled if any PreFilter plugin returns nodes that do not exist. ([#124559](https://github.com/kubernetes/kubernetes/pull/124559), [@chengjoey](https://github.com/chengjoey)) [SIG Scheduling and Testing]

","expr","elements","matchesSelector","attr","attrHandle","uniqueSort","duplicates","sortStable","createPseudo","relative",">"," ","+","~","preFilter","excess","unquoted","nodeNameSelector","expectedNodeName","pattern","operator","check","result","what","_argument","simple","forward","ofType","_context","xml","outerCache","nodeIndex","start","parent","useCache","diff","firstChild","lastChild","pseudo","args","setFilters","idx","matched","not","matcher","compile","unmatched","has","lang","elemLan...

- Verify that CRD default values in OpenAPI specs are pruned, with the exceptions of values under `metadata`. ([#78829](https://github.com/kubernetes/kubernetes/pull/78829), [@sttts](https://github.com/sttts))
- Fixes a bug that when there is a "connection refused" error, the reflector's ListAndWatch func will return directly but what expected is that sleep 1 second and rewatch since the specified resourceVersion.

- Fix a bug that Pods could stuck in the unschedulable pod pool 

    finally:
        some_cleanup()
```

...that way the `finally` block is run regardless of any exception that might happen.

### Features

* The same PR [#4575](https://github.com/tiangolo/fastapi/pull/4575) from above also fixes the `contextvars` context for the code before and after `yield`. This was the main objective of that PR.

This release contains changes that address the following vulnerabilities:

### CVE-2021-25735: Validating Admission Webhook does not observe some previous fields

A security issue was discovered in kube-apiserver that could allow node
updates to bypass a Validating Admission Webhook. You are only affected
by this vulnerability if you run a Validating Admission Webhook for Nodes

The behavior of some watch calls to the server when filtering on fields was incorrect. If watching objects with a filter, when an update was made that no longer matched the filter a DELETE event was correctly sent. However, the object that was returned by that delete was not the (correct) version before the update, but instead, the newer version. That meant the new object was not matched by the filter. This was a regression from behavior between cached watches on the server side and uncached watches,...

- Fixed an issue in the kubelet that showed when writeable layers and read-only layers were at different paths within the same mount.
  Kubernetes was previously detecting that the image filesystem was split, even when that was not really the case ([#128344](https://github.com/kubernetes/kubernetes/pull/128344), [@kannon92](https://github.com/kannon92)) [SIG Node]

pkg syscall (freebsd-arm64), type IfAnnounceMsghdr struct, Type uint8
pkg syscall (freebsd-arm64), type IfAnnounceMsghdr struct, Version uint8
pkg syscall (freebsd-arm64), type IfAnnounceMsghdr struct, What uint16
pkg syscall (freebsd-arm64), type IfData struct
pkg syscall (freebsd-arm64), type IfData struct, Addrlen uint8
pkg syscall (freebsd-arm64), type IfData struct, Baudrate uint64

Traditionally, the Kubernetes scheduler was based on the assumptions that additional persistent storage is available everywhere in the cluster and has infinite capacity. Topology constraints addressed the first point, but up to now pod scheduling was still done without considering that the remaining storage capacity may not be enough to start a new pod. [Storage capacity tracking](https://docs.k8s.io/concepts/storage/storage-capacity/) addresses that by adding an API for a CSI driver to report storage capacity...