Set the sandbox_image for the k8s example

Kubernetes requires some changes to the default containerd
configuration, for the cgroup driver and the sandbox image.

Update documentation links, to show what is prerequisites
and container runtime configuration and what is `kubeadm`.

Signed-off-by: Anders F Björklund <[email protected]>

Author: afbjorklund

examples/k8s.yaml

@@ -32,14 +32,14 @@ mounts: []
 containerd:
   system: true
   user: false
-# See https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/
 provision:
+# See <https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/install-kubeadm/>
 - mode: system
   script: |
     #!/bin/bash
     set -eux -o pipefail
     command -v kubeadm >/dev/null 2>&1 && exit 0
-    # Installing kubeadm on your hosts
+    # Install and configure prerequisites
     cat <<EOF | sudo tee /etc/modules-load.d/containerd.conf
     overlay
     br_netfilter
@@ -52,6 +52,7 @@ provision:
     net.bridge.bridge-nf-call-ip6tables = 1
     EOF
     sysctl --system
+    # Installing kubeadm, kubelet and kubectl
     export DEBIAN_FRONTEND=noninteractive
     apt-get update
     apt-get install -y apt-transport-https ca-certificates curl
@@ -69,16 +70,19 @@ provision:
     rm -f /etc/cni/net.d/*.conf*
     apt-get install -y kubelet kubeadm kubectl && apt-mark hold kubelet kubeadm kubectl
     systemctl enable --now kubelet
+# See <https://kubernetes.io/docs/setup/production-environment/container-runtimes/>
 - mode: system
   script: |
     #!/bin/bash
     set -eux -o pipefail
     grep SystemdCgroup /etc/containerd/config.toml && exit 0
     grep "version = 2" /etc/containerd/config.toml || exit 1
-    # Configuring a cgroup driver
+    # Configuring the systemd cgroup driver
+    # Overriding the sandbox (pause) image
     cat <<EOF >>/etc/containerd/config.toml
       [plugins]
         [plugins."io.containerd.grpc.v1.cri"]
+          sandbox_image = "$(kubeadm config images list | grep pause | sort -r | head -n1)"
           [plugins."io.containerd.grpc.v1.cri".containerd]
             [plugins."io.containerd.grpc.v1.cri".containerd.runtimes]
               [plugins."io.containerd.grpc.v1.cri".containerd.runtimes.runc]
@@ -87,6 +91,7 @@ provision:
                   SystemdCgroup = true
     EOF
     systemctl restart containerd
+# See <https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/create-cluster-kubeadm/>
 - mode: system
   script: |
     #!/bin/bash