### Other Deprecations
* Remove cmd/kube-discovery from the tree since it's not necessary anymore ([#42070](https://github.com/kubernetes/kubernetes/pull/42070), [@luxas](https://github.com/luxas))

## Changes to API Resources
### ABAC

* The kubelet --reconcile-cdir flag is deprecated because it has no function anymore. ([#35523](https://github.com/kubernetes/kubernetes/pull/35523), [@luxas](https://github.com/luxas))
* Notice of deprecation for recycler [#36760](https://github.com/kubernetes/kubernetes/pull/36760)

						disks = newDisks
						askDisks = 1
					} else {
						// If we did not find suitable disks, perform strict quorum listing
						// as no disk agrees on quorum anymore.
						askDisks = getListQuorum("strict", set.setDriveCount)
					}
				}

				// Special case: ask all disks if the drive count is 4
				if set.setDriveCount == 4 || askDisks > len(disks) {

- DRA: The `DynamicResourceAllocation` feature gate for the core functionality (GA in `v1.34`) has now been locked to enabled-by-default and cannot be disabled anymore. ([#134452](https://github.com/kubernetes/kubernetes/pull/134452), [@pohly](https://github.com/pohly)) [SIG Auth, Node, Scheduling and Testing]

#### Deployments
* [beta] Deployments uses a hashing collision avoidance mechanism that ensures new rollouts will not block on hashing collisions anymore. ([kubernetes/features#287](https://github.com/kubernetes/enhancements/issues/287))

#### PodDisruptionBudget

- `e2e.test` now rejects unknown `--provider` values instead of merely warning about them. An empty provider name is not accepted anymore and was replaced by `skeleton` (a provider with no special behavior). ([#73402](https://github.com/kubernetes/kubernetes/pull/73402), [@pohly](https://github.com/pohly))

  Jobs that are tracked using this mechanism get the annotation `batch.kubernetes.io/job-tracking`. This is a temporary measure. Two releases after this feature graduates to beta, the annotation won't be added to Jobs anymore. ([#98817](https://github.com/kubernetes/kubernetes/pull/98817), [@alculquicondor](https://github.com/alculquicondor))
- Added new kubelet alpha feature `SeccompDefault`. This feature enables falling back to

* Fixes volume predicate handler for equiv class ([#59335](https://github.com/kubernetes/kubernetes/pull/59335), [@resouer](https://github.com/resouer))

* Bugfix: vSphere Cloud Provider (VCP) does not need any special service account anymore. ([#59440](https://github.com/kubernetes/kubernetes/pull/59440), [@rohitjogvmw](https://github.com/rohitjogvmw))

- Kubelet: fix GetAllocatableCPUs method in cpumanager ([#113422](https://github.com/kubernetes/kubernetes/pull/113422), [@Garrybest](https://github.com/Garrybest)) [SIG Node]
- Pod logs using --timestamps are not broken up with timestamps anymore. ([#113517](https://github.com/kubernetes/kubernetes/pull/113517), [@rphillips](https://github.com/rphillips)) [SIG Node]

## Dependencies

### Added
_Nothing has changed._

### Changed

        transition canned Estimators that are warm started from
        `tf.train.Optimizers` to `tf.keras.optimizers`.
    *   Losses are scaled in canned estimator v2 and not in the optimizers
        anymore. If you are using Estimator + distribution strategy + optimikzer
        v1 then the behavior does not change. This implies that if you are using
        custom estimator with optimizer v2, you have to scale losses. We have