multiple times. This is because it would produce the same values each
        time, which may not be intended.
*   API changes under `tf.experimental.dtensor`:
    *   New API for initialization of CPU/GPU/TPU in dtensor.
        `dtensor.initialize_accelerator_system` and
        `dtensor.shutdown_accelerator_system`.
    *   The following existing API will be removed:

- Bug fix for AWS NLB service when nodePort for existing servicePort changed manually. ([#89562](https://github.com/kubernetes/kubernetes/pull/89562), [@M00nF1sh](https://github.com/M00nF1sh)) [SIG Cloud Provider]
- CSINode initialization does not crash kubelet on startup when APIServer is not reachable or kubelet has not the right credentials yet. ([#89589](https://github.com/kubernetes/kubernetes/pull/89589), [@jsafrane](https://github.com/jsafrane)) [SIG Storage]