[overwrite of manual edits to GCP Health Checks](https://github.com/kubernetes/ingress/issues/842)
  managed by the GLBC Ingress Controller. This can cause the health checks to start failing,
  requiring you to reapply the manual edits.
  * This issue does not affect clusters that start out with Kubernetes v1.6.4 or higher.
  * This issue does not affect Health Checks that were left in the configuration

      endpoint now gets broken when that endpoint becomes non-ready
      (rather than continuing until the endpoint is fully deleted).
  
    - Traffic to a service IP now starts getting rejected (as opposed to
      merely dropped) as soon as there are no longer any *usable*
      endpoints, rather than waiting until all of the terminating

	buf, err := xattr.LGet(s.formatFile, attr)
	if err != nil {
		// We start off with '0' if we can read the attributes
		return 0
	}
	return binary.LittleEndian.Uint64(buf[:8])
}

func (s *xlStorage) getWriteAttribute() uint64 {
	attr := "user.total_writes"
	buf, err := xattr.LGet(s.formatFile, attr)
	if err != nil {
		// We start off with '0' if we can read the attributes
		return 0
	}

return document.documentElement;\n  }\n\n  // Here we make sure to give as \"start\" the element that comes first in the DOM\n  const order =\n    element1.compareDocumentPosition(element2) &\n    Node.DOCUMENT_POSITION_FOLLOWING;\n  const start = order ? element1 : element2;\n  const end = order ? element2 : element1;\n\n  // Get common ancestor container\n  const range = document.createRange();\n  range.setStart(start, 0);\n  range.setEnd(end, 0);\n  const { commonAncestorContainer } = range;\n\n...

      endpoint now gets broken when that endpoint becomes non-ready
      (rather than continuing until the endpoint is fully deleted).
  
    - Traffic to a service IP now starts getting rejected (as opposed to
      merely dropped) as soon as there are no longer any *usable*
      endpoints, rather than waiting until all of the terminating

  
  This fixes a bug where clients with requests using hand size x max seats greater than the nominal concurrency limit can starve other requests in the same priority level. ([#118601](https://github.com/kubernetes/kubernetes/pull/118601), [@andrewsykim](https://github.com/andrewsykim)) [SIG API Machinery]

## Dependencies

### Added

        the layer's dtype. If you had a model that used `float64`, it will
        probably silently use `float32` in TensorFlow2, and a warning will be
        issued that starts with Layer "layer-name" is casting an input tensor
        from dtype float64 to the layer's dtype of float32. To fix, either set
        the default dtype to float64 with

* kubeadm: use kubeconfig flag instead of kubeconfig-dir on init phase bootstrap-token ([#71803](https://github.com/kubernetes/kubernetes/pull/71803), [@yagonobre](https://github.com/yagonobre))
* On GCI, NPD starts to monitor kubelet, docker, containerd crashlooping, read-only filesystem and corrupt docker overlay2 issues. ([#71522](https://github.com/kubernetes/kubernetes/pull/71522), [@wangzhen127](https://github.com/wangzhen127))

- Reduced GCE log rotation check from 1 hour to every 5 minutes.  Rotation policy is unchanged (new day starts, log file size > 100MB). ([#76352](https://github.com/kubernetes/kubernetes/pull/76352), [@jpbetz](https://github.com/jpbetz))

- Fixed a bug where init containers may fail to start due to a temporary container runtime failure. ([#127213](https://github.com/kubernetes/kubernetes/pull/127213), [@SergeyKanzhelev](https://github.com/SergeyKanzhelev)) [SIG Node]