Merge pull request #42780 from Random-Liu/handle-npd-in-cluster-upgrade

cluster/gce/config-default.sh

@@ -148,7 +148,6 @@ ENABLE_CLUSTER_UI="${KUBE_ENABLE_CLUSTER_UI:-true}"
 #   standalone     - Run node problem detector as standalone system daemon.
 if [[ "${NODE_OS_DISTRIBUTION}" == "gci" ]]; then
   # Enable standalone mode by default for gci.
-  # TODO: Consider upgrade test.
   ENABLE_NODE_PROBLEM_DETECTOR="${KUBE_ENABLE_NODE_PROBLEM_DETECTOR:-standalone}"
 else
   ENABLE_NODE_PROBLEM_DETECTOR="${KUBE_ENABLE_NODE_PROBLEM_DETECTOR:-daemonset}"

cluster/gce/config-test.sh

@@ -173,7 +173,6 @@ ENABLE_CLUSTER_UI="${KUBE_ENABLE_CLUSTER_UI:-true}"
 #   standalone     - Run node problem detector as standalone system daemon.
 if [[ "${NODE_OS_DISTRIBUTION}" == "gci" ]]; then
   # Enable standalone mode by default for gci.
-  # TODO: Consider upgrade test.
   ENABLE_NODE_PROBLEM_DETECTOR="${KUBE_ENABLE_NODE_PROBLEM_DETECTOR:-standalone}"
 else
   ENABLE_NODE_PROBLEM_DETECTOR="${KUBE_ENABLE_NODE_PROBLEM_DETECTOR:-daemonset}"

cluster/gce/upgrade.sh

@@ -88,6 +88,8 @@ function upgrade-master() {
 
   detect-master
   parse-master-env
+  upgrade-master-env
+
   backfile-kubeletauth-certs
 
   # Delete the master instance. Note that the master-pd is created
@@ -102,6 +104,15 @@ function upgrade-master() {
   wait-for-master
 }
 
+function upgrade-master-env() {
+  echo "== Upgrading master environment variables. =="
+  # Generate the node problem detector token if it isn't present on the original
+  # master.
+ if [[ "${ENABLE_NODE_PROBLEM_DETECTOR:-}" == "standalone" && "${NODE_PROBLEM_DETECTOR_TOKEN:-}" == "" ]]; then
+    NODE_PROBLEM_DETECTOR_TOKEN=$(dd if=/dev/urandom bs=128 count=1 2>/dev/null | base64 | tr -d "=+/" | dd bs=32 count=1 2>/dev/null)
+  fi
+}
+
 # TODO(mikedanese): delete when we don't support < 1.6
 function backfile-kubeletauth-certs() {
   if [[ ! -z "${KUBEAPISERVER_CERT_BASE64:-}" && ! -z "${KUBEAPISERVER_CERT_BASE64:-}" ]]; then
@@ -282,6 +293,8 @@ function prepare-node-upgrade() {
   KUBELET_CERT_BASE64=$(get-env-val "${node_env}" "KUBELET_CERT")
   KUBELET_KEY_BASE64=$(get-env-val "${node_env}" "KUBELET_KEY")
 
+  upgrade-node-env
+
   # TODO(zmerlynn): How do we ensure kube-env is written in a ${version}-
   #                 compatible way?
   write-node-env
@@ -295,6 +308,17 @@ function prepare-node-upgrade() {
   echo "== Finished preparing node upgrade (to ${KUBE_VERSION}). ==" >&2
 }
 
+function upgrade-node-env() {
+  echo "== Upgrading node environment variables. =="
+  # Get the node problem detector token from master if it isn't present on
+  # the original node.
+  if [[ "${ENABLE_NODE_PROBLEM_DETECTOR:-}" == "standalone" && "${NODE_PROBLEM_DETECTOR_TOKEN:-}" == "" ]]; then
+    detect-master
+    local master_env=$(get-master-env)
+    NODE_PROBLEM_DETECTOR_TOKEN=$(get-env-val "${master_env}" "NODE_PROBLEM_DETECTOR_TOKEN")
+  fi
+}
+
 # Prereqs:
 # - prepare-node-upgrade should have been called successfully
 function do-node-upgrade() {