* Fix broken detection of non-root image user ID ([#78261](https://github.com/kubernetes/kubernetes/pull/78261), [@tallclair](https://github.com/tallclair))
* Active watches of custom resources now terminate properly if the CRD is modified. ([#78029](https://github.com/kubernetes/kubernetes/pull/78029), [@liggitt](https://github.com/liggitt))

* Remove stale volumes if endpoint/svc creation fails. ([#35285](https://github.com/kubernetes/kubernetes/pull/35285), [@humblec](https://github.com/humblec))
* Remove Job also from .status.active for Replace strategy ([#35420](https://github.com/kubernetes/kubernetes/pull/35420), [@soltysh](https://github.com/soltysh))

* have basic kubectl crud agnostic of registered types ([#36085](https://github.com/kubernetes/kubernetes/pull/36085), [@deads2k](https://github.com/deads2k))
* Fix how we iterate over active jobs when removing them for Replace policy ([#36161](https://github.com/kubernetes/kubernetes/pull/36161), [@soltysh](https://github.com/soltysh))

### Other (Cleanup or Flake)

* Fix incorrect conflict errors applying strategic merge patches to resources. ([#43871](https://github.com/kubernetes/kubernetes/pull/43871), [@liggitt](https://github.com/liggitt))

* Fix init container status reporting when active deadline is exceeded. ([#46305](https://github.com/kubernetes/kubernetes/pull/46305), [@sjenning](https://github.com/sjenning))

    // no notification sent with removeLoadingValue
    assertThat(map.removalNotificationQueue.isEmpty()).isTrue();
    assertThat(segment.count).isEqualTo(0);
    assertThat(table.get(0)).isNull();

    // active
    Object value = new Object();
    DummyValueReference<Object, Object> previousRef = DummyValueReference.create(value);
    valueRef = new LoadingValueReference<>(previousRef);
    entry.setValueReference(valueRef);

    // no notification sent with removeLoadingValue
    assertThat(map.removalNotificationQueue.isEmpty()).isTrue();
    assertThat(segment.count).isEqualTo(0);
    assertThat(table.get(0)).isNull();

    // active
    Object value = new Object();
    DummyValueReference<Object, Object> previousRef = DummyValueReference.create(value);
    valueRef = new LoadingValueReference<>(previousRef);
    entry.setValueReference(valueRef);

### Other (Cleanup or Flake)

			// When we are going to return error, attempt to delete success
			// on some of the drives, if we cannot we do not have to notify
			// caller this dangling object will be now scheduled to be removed
			// via active healing.
			dg.Go(func() error {
				return disks[index].DeleteVersion(context.Background(), dstBucket, dstEntry, metadata[index], false, DeleteOptions{
					UndoWrite:  true,
					OldDataDir: dataDirs[index],

for interrupting the StatefulSet Pod management is now ignored. StatefulSets with this annotation set to `true` or with no value will behave just as they did in previous versions. Dormant StatefulSets with the annotation set to `false` will become active after upgrading.

* The CronJob object is now enabled by default at `v1beta1`. CronJob `v2alpha1` is still available, but it must be explicitly enabled. We recommend that you move any current CronJob objects to `batch/v1beta1.CronJob`....