Automatic merge from submit-queue

kubeadm

**What this PR does / why we need it**:

This PR add alpha version of `kubeadm` tool, which allows user to boostrap a cluster rather quite easily. This is the initial contribution from @kubernetes/sig-cluster-lifecycle members, who's aim is to build easy-to-use tools that help to operate a cluster throughout its lifetime.

**Which issue this PR fixes**: a leap towards kubernetes/features#11

**Special notes for your reviewer**: previously seen by many folks in #31221

**Release note**:

```release-note
`kubeadm` (alpha) provides an easy way to securely bootstrap Kubernetes on Linux, see http://kubernetes.io/docs/kubeadm/
```

Automatic merge from submit-queue

Added e2e framweork and a simple test for HA master.

Added e2e framweork for testing HA master. Added one simple e2e test for HA master that first grows and then shrinks GCE cluster.

fix https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/pr-logs/pull/33262/kubernetes-pull-verify-all/15586/

Automatic merge from submit-queue

Print a more helpful error message when failing to start rolling-updates

Hopefully this will help us track down where the 1.3 -> 1.4 upgrades are breaking down. We'll need to cherry-pick this into release-1.4 to have any effect, though.

Automatic merge from submit-queue

Removed unused var.

Automatic merge from submit-queue

Fix an 'index out of range' panic when setting capabilities.

cc @yujuhong @feiskyer

Automatic merge from submit-queue

Put loopback authn/authz first in chain

We want the loopback token auth to go first in the chain, for performance reasons, and so the loopback token isn't seen by any remote token authenticators configured.

The loopback authorizer should also go first in the chain for performance.

Automatic merge from submit-queue

Do not report error when deleting an attached volume

Persistent volume controller should not send warning events to a PV and mark the PV as failed when the volume is still attached.

This happens when a user quickly deletes a pod and associated PVC - PV is slowly detaching, while the PVC is already deleted and the PV enters Failed phase.

`Deleter.Deleter` can now return `tryAgainError`, which is sent as INFO to the PV to let the user know we did not forget to delete the PV, however the PV stays in Released state. The controller tries again in the next sync (15 seconds by default).

Fixes #31511

Automatic merge from submit-queue

Split dns healthcheck into two different urls

Attempt to fix #30633.

<s>This new kube-dns pod template creates two exechealthz processes listen on two different ports for kubedns and dnsmasq correspondingly. 

@thockin @girishkalele

Automatic merge from submit-queue

Fixed e2e_node build error on Mac.

fixes #33091 

cc @vishh , any suggestion to avoid duplicated codes?

Automatic merge from submit-queue

Node E2E: Cleanup e2e services

fixes #31765.

This PR is composed of 2 commits:
* The 1st commit split services.go into: `services.go`, `server.go` and `internal_services.go`:
  * `services.go` contains the public object `E2EServices` which is used by the test framework directly.
  * `internal_services.go` contains the internal object `e2eServices` which manages internal (statically-linked) services - apiserver, etcd and namespace_controller.
  * `server.go` is the object managing exec process, both internal_services and kubelet are running as separate processes and managed with server.go.
* The 2nd commit added `monitorParent` option in start function of `E2EServices`. This is added to fix #31765:
  * If `--stop-services=true`, `monitorParent` will be true, so that service processes will die with the parent process so as to enforce proper clean up.
  * If `--stop-services=false`, `monitorParent` will be false, so that service processes will not die with the parent process and keep running for debugging.

This PR also moved the kubelet start logic into `E2EServices` (start kubelet in the test process), so that we can use flags directly when starting kubelet. Before we had to pass them to the services process and let it start kubelet, which was quite troublesome.

@vishh 
/cc @kubernetes/sig-node

Automatic merge from submit-queue

Fix TODO: Rename podStatus to apiPodStatus, rename internalPodStatus to podStatus 

monitorParent to false when stop-services=false.

Automatic merge from submit-queue

Fix k8sm flaky UT.

fixes #25349

Currently the boostrap fails when a token is provided by the user
on `master init` and works when the token is generated. This is
because of a mismatch of how the token string in the kube-discovery
secret is encoded.

This is useful for users who are used to deploying with a flannel
overlay network.

getComponentCommand method is too hyperkube specific
to be used for etcd Pod.

Various improvements for kubeadm. Removed the user command, as it's too little time for implementing that. Now it's possible to use multiple arches.

Print user friendly error with information about required version.