- [(No, really, you MUST read this before you upgrade)](#no-really-you-must-read-this-before-you-upgrade)
  - [Known Issues](#known-issues)
    - [Etcd v3.5.[0-2] data corruption](#etcd-v350-2-data-corruption)
  - [Changes by Kind](#changes-by-kind-17)
    - [Deprecation](#deprecation)
    - [API Change](#api-change-6)
    - [Feature](#feature-12)
    - [Documentation](#documentation)

* Fix issue in Portworx volume driver causing controller manager to crash ([#76341](https://github.com/kubernetes/kubernetes/pull/76341), [@harsh-px](https://github.com/harsh-px))
* fix azure disk list corruption issue ([#77187](https://github.com/kubernetes/kubernetes/pull/77187), [@andyzhangx](https://github.com/andyzhangx))

		userType = svcUser
	} else if cred.IsTemp() {
		userType = stsUser
	}
	ui := newUserIdentity(cred)
	// Overwrite the user identity here. As store should be
	// atomic, it shouldn't cause any corruption.
	if err := store.saveUserIdentity(ctx, cred.AccessKey, userType, ui); err != nil {
		return err
	}

	return cache.updateUserWithClaims(cred.AccessKey, ui)
}

- Fix: azure file mount timeout issue ([#88610](https://github.com/kubernetes/kubernetes/pull/88610), [@andyzhangx](https://github.com/andyzhangx)) [SIG Cloud Provider and Storage]
- Fix: corrupted mount point in csi driver ([#88569](https://github.com/kubernetes/kubernetes/pull/88569), [@andyzhangx](https://github.com/andyzhangx)) [SIG Storage]

- Fix concurrent map access causing panics when logging timed-out API calls. (#106113, @marseel) [SIG API Machinery]
- Fixed very rare volume corruption when a pod is deleted while kubelet is offline.

    val request = server.takeRequest()
    assertThat(request.headers["Accept-Encoding"]).isEqualTo("custom")
  }

  /**
   * Test a bug where gzip input streams weren't exhausting the input stream, which corrupted the
   * request that followed or prevented connection reuse. http://code.google.com/p/android/issues/detail?id=7059
   * http://code.google.com/p/android/issues/detail?id=38817
   */

		t.Fatalf("Test %s: Failed to fetch the copied object: <ERROR> %s", instanceType, err)
	}
	r.Close()
	if !bytes.Equal(buf.Bytes(), bytesData[0].byteData) {
		t.Fatalf("Test %s: returned data is not expected corruption detected:", instanceType)
	}

	globalCompressConfigMu.Lock()
	compressEnabled := globalCompressConfig.Enabled
	globalCompressConfigMu.Unlock()
	if compressEnabled && !r.ObjInfo.IsCompressed() {

- Updated `kubeadm` to use the `InitialCorruptCheck=true` etcd feature gate instead of the deprecated `--experimental-initial-corrupt-check` flag. Also replaced the use of `--experimental-watch-progress-notify-interval` with its graduated counterpart `--watch-progress-notify-interval`. ([#132838](https://github.com/kubernetes/kubernetes/pull/132838), [@AwesomePatrol](https://gi...

- In a HA env, during the period a standby scheduler lost connection to API server, if a Pod is deleted and recreated, and the standby scheduler becomes master afterwards, there could be a scheduler cache corruption. This PR fixes this issue. ([#91126](https://github.com/kubernetes/kubernetes/pull/91126), [@Huang-Wei](https://github.com/Huang-Wei)) [SIG Scheduling]

- Mount-utils: treat syscall.ENODEV as corrupted mount ([#126174](https://github.com/kubernetes/kubernetes/pull/126174), [@dobsonj](https://github.com/dobsonj)) [SIG Storage]