Use containerd cri plugin from init

[detiber]

This PR attempts to address https://github.com/linuxkit/kubernetes/issues/74

The kernel, init, and runc image updates are a bit extraneous for this changeset, but the containerd update is required.

kubelet changes:

• update to latest stable k8s release
• update to latest cri tools
• onboot hack to create /var/lib/cni/{bin,conf}
  • removing the cri-containerd image caused the directories to not exist and containerd appears to be processing the runtime mounts before the mkdir entries.

I also introduced files necessary to deploy calico networking rather than weave, since weave currently does not work with these changes. The issue appears to be with the way weave cni plugin uses nsenter.

[detiber]

Currently all of the cri rtf tests are failing with flakiness starting the static pod manifests, there appears to be some flakiness around the interaction between the kubelet and containerd for pods that fail to come up and should be restarted, I'm seeing errors similar to:

I0504 14:43:30.680271     638 kuberuntime_manager.go:513] Container {Name:kube-scheduler Image:k8s.gcr.io/kube-scheduler-amd64:v1.10.2 Command:[kube-scheduler --address=127.0.0.1 --leader-elect=true --kubeconfig=/etc/kubernetes/scheduler.conf] Args:[] WorkingDir: Ports:[] EnvFrom:[] Env:[] Resources:{Limits:map[] Requests:map[cpu:{i:{value:100 scale:-3} d:{Dec:<nil>} s:100m Format:DecimalSI}]} VolumeMounts:[{Name:kubeconfig ReadOnly:true MountPath:/etc/kubernetes/scheduler.conf SubPath: MountPropagation:<nil>}] VolumeDevices:[] LivenessProbe:&Probe{Handler:Handler{Exec:nil,HTTPGet:&HTTPGetAction{Path:/healthz,Port:10251,Host:127.0.0.1,Scheme:HTTP,HTTPHeaders:[],},TCPSocket:nil,},InitialDelaySeconds:15,TimeoutSeconds:15,PeriodSeconds:10,SuccessThreshold:1,FailureThreshold:8,} ReadinessProbe:nil Lifecycle:nil TerminationMessagePath:/dev/termination-log TerminationMessagePolicy:File ImagePullPolicy:IfNotPresent SecurityContext:nil Stdin:false StdinOnce:false TTY:false} is dead, but RestartPolicy says that we should restart it.
linuxkit-b2a001ae29c0:/# Connection to localhost closed by remote host.

I suspect this might be similar to the flakes mentioned in https://github.com/linuxkit/kubernetes/issues/64

A bit of digging, and I suspect that https://github.com/containerd/cri/issues/733 might be coming into play.