- Kubeadm: make etcd pod request 100m CPU, 100Mi memory and 100Mi ephemeral_storage by default ([#94479](https://github.com/kubernetes/kubernetes/pull/94479), [@knight42](https://github.com/knight42)) [SIG Cluster Lifecycle]
- Kubeadm: make the command "kubeadm alpha kubeconfig user" accept a "--config" flag and remove the following flags:

HSPLokio/SegmentPool;-><clinit>()V
HSPLokio/SegmentPool;-><init>()V
HSPLokio/SegmentPool;->firstRef()Ljava/util/concurrent/atomic/AtomicReference;
HSPLokio/SegmentPool;->recycle(Lokio/Segment;)V
HSPLokio/SegmentPool;->take()Lokio/Segment;
HSPLokio/SocketAsyncTimeout;-><init>(Ljava/net/Socket;)V
HSPLokio/Timeout$Companion$NONE$1;-><init>()V
HSPLokio/Timeout;-><clinit>()V
HSPLokio/Timeout;-><init>()V
HSPLokio/Timeout;->throwIfReached()V

    specifically bugging me now because they make the action graph harder to
    analyze programmatically. By using aliases here, we would need to handle
    potentially aliased go_binary targets and dereference to the effective
    target.
  
    The comment references an issue with `pure = select(...)` which appears
    to be resolved considering this now builds.
  

  static <E> Queue<E> discardingQueue() {
    return (Queue) DISCARDING_QUEUE;
  }

  /*
   * Note: All of this duplicate code sucks, but it saves a lot of memory. If only Java had mixins!
   * To maintain this code, make a change for the strong reference type. Then, cut and paste, and
   * replace "Strong" with "Soft" or "Weak" within the pasted text. The primary difference is that

- This release added support for `NodeExpandSecret` for CSI driver client which enables the CSI drivers to make use of this secret while performing node expansion operation based on the user request. Previously there was no secret  provided as part of the `nodeexpansion` call, thus CSI drivers did not make use of the same while expanding the volume at the node side. ([#105963](https://github.com/kubernetes/kubernetes/pull/105963), [@zhucan](https://github.com/zhucan))...

				makeGateway(constants.DefaultNamespaceWaypoint, "fake", true, true),
			},
			expectedOutFile: "default-gateway",
		},
		{
			name: "all namespaces gateways",
			args: strings.Split("list -A", " "),
			gateways: []*gateway.Gateway{
				makeGateway(constants.DefaultNamespaceWaypoint, "default", true, true),
				makeGateway(constants.DefaultNamespaceWaypoint, "fake", true, true),
			},
			expectedOutFile: "all-gateway",
		},

	errFatal := func(err error) {
		t.Helper()
		if err != nil {
			t.Fatal(err)
		}
	}

	// We fake a local and remote server.
	remoteHost := remote.HostName()

	// 1: Echo
	serverSent := make(chan struct{})
	serverCanceled := make(chan struct{})
	register := func(manager *Manager) {
		errFatal(manager.RegisterStreamingHandler(handlerTest, StreamHandler{

  
  With this patch, the external provisioner can remove the finalizer in next reconcile loop. Unfortunately if the provious existing pv has the Failed state, this patch won't take effort. It requires users to remove finalizer. ([#126045](https://github.com/kubernetes/kubernetes/pull/126045), [@carlory](https://github.com/carlory)) [SIG Apps and Storage]

`kube.Client` just bundles these all in one place, so we have a single object to pass around anywhere we need Kubernetes access.
Additionally, it has a fake client variant, that does the same thing but with fake clients for use in unit tests.

Aside from this, there are a few conveniences and workarounds built-in to the client to make things simpler.

*All Istio Kubernetes usage should use this library and not operate on Kubernetes clients directly.*

		fakeCtx,
		pod,
		true,
	).Return(errors.New("fake error"))

	fakeClientSet := fake.NewClientset()

	fakeIPSetDeps := ipset.FakeNLDeps()
	set := ipset.IPSet{V4Name: "foo-v4", Prefix: "foo", Deps: fakeIPSetDeps}

	m := getFakeDPWithIPSet(server, fakeClientSet, set)
	expectPodRemovedFromIPSet(fakeIPSetDeps, pod.Status.PodIPs)

	// pod is not in fake client, so if this will try to remove annotation, it will fail.