Usage:  "custom TCP_USER_TIMEOUT for socket buffers",
		Hidden: true,
		Value:  10 * time.Minute,
		EnvVar: "MINIO_CONN_USER_TIMEOUT",
	},
	cli.StringFlag{
		Name:   "interface",
		Usage:  "bind to right VRF device for MinIO services",
		Hidden: true,
		EnvVar: "MINIO_INTERFACE",
	},
	cli.DurationFlag{
		Name:   "dns-cache-ttl",
		Usage:  "custom DNS cache TTL",
		Hidden: true,
		Value: func() time.Duration {

- the IP allocator in v1.17.0 can return errors such as `the cluster IP <ip> for service <service-name> is not within the service CIDR <cidr>; please recreate` in the logs of the kube-apiserver. The cause is incorrect CIDR calculations if the service CIDR (`--service-cluster-ip-range`) is set to bits lower than `/16`. This is fixed in http://issue.k8s.io/86534 and will be resolved in v1.17.1.

    ask_cuda_compute_capabilities = (
        'Please specify a list of comma-separated CUDA compute capabilities '
        'you want to build with.\nYou can find the compute capability of your '
        'device at: https://developer.nvidia.com/cuda-gpus. Each capability '
        'can be specified as "x.y" or "compute_xy" to include both virtual and'
        ' binary GPU code, or as "sm_xy" to only include the binary '

			Uid:      010000000,
			ModTime:  time.Unix(0, 0),
			Typeflag: '0',
		}},
	}, {
		// USTAR archive with a regular entry with non-zero device numbers.
		file: "testdata/ustar-file-devs.tar",
		headers: []*Header{{
			Name:     "file",
			Mode:     0644,
			Typeflag: '0',
			ModTime:  time.Unix(0, 0),
			Devmajor: 1,
			Devminor: 1,

- Fixed an issue where the Service LoadBalancer controller was not correctly considering the `service.Status new IPMode` field and excluding the Ports when checking if the status was changed, resulting in the changed field potentially not to update the `service.Status` correctly. ([#125225](https://github.com/kubernetes/kubernetes/pull/125225), [@aojea](https://github.com/aojea)) [SIG Apps, Cloud...

     * Kubernetes clusters can now span zones within a cloud provider. Pods from a
service will be automatically spread across zones, enabling applications to
tolerate zone failure.
     * Simplified way to run a container on every node (DaemonSet API (Beta) in the
Extensions API group): Kubernetes can schedule a service (such as a logging
agent) that runs one, and only one, pod per node.

* Make kube-proxy service abstraction optional. ([#71355](https://github.com/kubernetes/kubernetes/pull/71355), [@bradhoekstra](https://github.com/bradhoekstra))
    * Add the 'service.kubernetes.io/service-proxy-name' label to a Service to disable the kube-proxy service proxy implementation.

    TF_OperationDescription* desc =
        TF_NewOperation(host_graph_, func_name_, func_node_name_);
    for (auto input : inputs) {
      TF_AddInput(desc, input);
    }
    // Set device to CPU because some ops inside the function might not be
    // available on GPU.
    TF_SetDevice(desc, "/cpu:0");
    *op = TF_FinishOperation(desc, s_);
    ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);

- During device plugin allocation, resources requested by the pod can only be allocated if the device plugin has registered itself to kubelet AND healthy devices are present on the node to be allocated. If these conditions are not sattisfied, the pod would fail with `UnexpectedAdmissionError` error. ([#116337](https://github.com/kube...

- Fixes service account token admission error in clusters that do not run the service account token controller ([#87029](https://github.com/kubernetes/kubernetes/pull/87029), [@liggitt](https://github.com/liggitt)) [SIG Auth]