* Add orphaning finalizer logic to GC ([#25599](https://github.com/kubernetes/kubernetes/pull/25599), [@caesarxuchao](https://github.com/caesarxuchao))
* GCI-backed masters mount srv/kubernetes and srv/sshproxy in the right place ([#26238](https://github.com/kubernetes/kubernetes/pull/26238), [@ihmccreery](https://github.com/ihmccreery))

import via the `.idea/checkstyle-idea.xml` configuration file. No further
action is required.

#### Formatting

We are in the process of migrating towards automatic formatting Java file
using [spotless], backed by the Eclipse formatter. **We strongly recommend
installing using the [Eclipse Code Formatter] plugin** so that you can
apply the correct formatting directly in IntelliJ.  The configuration for

 * {@code CaffeinatedGuava} adapter class</a>, which lets you build a Guava {@code Cache} or a Guava
 * {@code LoadingCache} backed by a Guava {@code CacheLoader}.
 *
 * <p>Caffeine's API for asynchronous operations uses {@code CompletableFuture}: <a

- Kubelet metrics have been changed to buckets. 
  For example the exec/{podNamespace}/{podID}/{containerName} is now just exec. ([#87913](https://github.com/kubernetes/kubernetes/pull/87913), [@cheftako](https://github.com/cheftako)) [SIG Node]

 * {@code CaffeinatedGuava} adapter class</a>, which lets you build a Guava {@code Cache} or a Guava
 * {@code LoadingCache} backed by a Guava {@code CacheLoader}.
 *
 * <p>Caffeine's API for asynchronous operations uses {@code CompletableFuture}: <a

- Metric buckets for `pod_start_duration_seconds` were changed to `{0.5, 1, 2, 3, 4, 5, 6, 8, 10, 20, 30, 45, 60, 120, 180, 240, 300, 360, 480, 600, 900, 1200, 1800, 2700, 3600}`. ([#120680](https://github.com/kubernetes/kubernetes/pull/120680), [@ruiwen-zhao](ht...

- Kube-controller-manager...

    ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());
    ASSERT_TRUE(absl::StrContains(device_name, "GPU:0")) << device_name;

    // .backing_device of shape is CPU since the tensor is backed by CPU
    backing_device_name =
        TFE_TensorHandleBackingDeviceName(retvals[0], status.get());
    ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

}

void NoOpDeallocator(void* data, size_t, void*) {}

TEST(CAPI, MalformedTensor) {
  // See https://github.com/tensorflow/tensorflow/issues/7394
  // num_dims = 0 implies a scalar, so should be backed by at least 4 bytes of
  // data.
  TF_Tensor* t =
      TF_NewTensor(TF_FLOAT, nullptr, 0, nullptr, 0, &NoOpDeallocator, nullptr);
  ASSERT_TRUE(t == nullptr);
}

TEST(CAPI, AllocateTensor) {

* Kubernetes version command line parameter in kubeadm has been updated to drop an unnecessary redirection from ci/latest.txt to ci-cross/latest.txt. Users should know exactly where the builds are stored on Google Cloud storage buckets from now on. For example for 1.9 and 1.10, users can specify ci/latest-1.9 and ci/latest-1.10 as the CI build jobs what build images correctly updates those. The CI jobs for master update the ci-cross/latest location, so if you are looking for latest...