From d874dbfcb1529077e2889f54279b559e09a9790d Mon Sep 17 00:00:00 2001
From: Zhen Wang <zhenw@google.com>
Date: Mon, 29 Jul 2019 14:22:41 -0700
Subject: [PATCH] Bump NPD version to v0.7 for GCI

---
 cluster/gce/gci/configure-helper.sh    | 10 +++++++---
 cluster/gce/gci/configure.sh           |  4 ++--
 test/e2e/node/node_problem_detector.go | 21 +++++++++++++++++++++
 3 files changed, 30 insertions(+), 5 deletions(-)

diff --git a/cluster/gce/gci/configure-helper.sh b/cluster/gce/gci/configure-helper.sh
index b3ff4e86fb0..b64da4b6255 100644
--- a/cluster/gce/gci/configure-helper.sh
+++ b/cluster/gce/gci/configure-helper.sh
@@ -1396,13 +1396,17 @@ function start-node-problem-detector {
     local -r km_config="${KUBE_HOME}/node-problem-detector/config/kernel-monitor.json"
     # TODO(random-liu): Handle this for alternative container runtime.
     local -r dm_config="${KUBE_HOME}/node-problem-detector/config/docker-monitor.json"
+    local -r sm_config="${KUBE_HOME}/node-problem-detector/config/systemd-monitor.json"
+    local -r ssm_config="${KUBE_HOME}/node-problem-detector/config/system-stats-monitor.json"
+
     local -r custom_km_config="${KUBE_HOME}/node-problem-detector/config/kernel-monitor-counter.json"
-    local -r custom_dm_config="${KUBE_HOME}/node-problem-detector/config/docker-monitor-counter.json"
     local -r custom_sm_config="${KUBE_HOME}/node-problem-detector/config/systemd-monitor-counter.json"
+
     flags="${NPD_TEST_LOG_LEVEL:-"--v=2"} ${NPD_TEST_ARGS:-}"
     flags+=" --logtostderr"
-    flags+=" --system-log-monitors=${km_config},${dm_config}"
-    flags+=" --custom-plugin-monitors=${custom_km_config},${custom_dm_config},${custom_sm_config}"
+    flags+=" --config.system-log-monitor=${km_config},${dm_config},${sm_config}"
+    flags+=" --config.system-stats-monitor=${ssm_config}"
+    flags+=" --config.custom-plugin-monitor=${custom_km_config},${custom_sm_config}"
     local -r npd_port=${NODE_PROBLEM_DETECTOR_PORT:-20256}
     flags+=" --port=${npd_port}"
     if [[ -n "${EXTRA_NPD_ARGS:-}" ]]; then
diff --git a/cluster/gce/gci/configure.sh b/cluster/gce/gci/configure.sh
index ad5d6d6eae0..77b558e6fdd 100644
--- a/cluster/gce/gci/configure.sh
+++ b/cluster/gce/gci/configure.sh
@@ -26,8 +26,8 @@ set -o pipefail
 ### Hardcoded constants
 DEFAULT_CNI_VERSION="v0.7.5"
 DEFAULT_CNI_SHA1="52e9d2de8a5f927307d9397308735658ee44ab8d"
-DEFAULT_NPD_VERSION="v0.6.3"
-DEFAULT_NPD_SHA1="3a6ac56be6c121f1b94450bfd1a81ad28d532369"
+DEFAULT_NPD_VERSION="v0.7.1"
+DEFAULT_NPD_SHA1="a9cae965973d586bf5206ad4fe5aae07e6bfd154"
 DEFAULT_CRICTL_VERSION="v1.14.0"
 DEFAULT_CRICTL_SHA1="1f93c6183d0a4e186708efe7899da7a7bce9c736"
 DEFAULT_MOUNTER_TAR_SHA="8003b798cf33c7f91320cd6ee5cec4fa22244571"
diff --git a/test/e2e/node/node_problem_detector.go b/test/e2e/node/node_problem_detector.go
index 871018b56e6..11c26a2b79c 100644
--- a/test/e2e/node/node_problem_detector.go
+++ b/test/e2e/node/node_problem_detector.go
@@ -131,6 +131,14 @@ var _ = SIGDescribe("NodeProblemDetector [DisabledForLargeClusters]", func() {
 			gomega.Eventually(func() error {
 				return verifyEvents(f, eventListOptions, 1, "AUFSUmountHung", node.Name)
 			}, pollTimeout, pollInterval).Should(gomega.Succeed())
+
+			// Node problem detector reports kubelet start events automatically starting from NPD v0.7.0+.
+			// Since Kubelet may be restarted for a few times after node is booted. We just check the event
+			// is detected, but do not check how many times Kubelet is started.
+			ginkgo.By(fmt.Sprintf("Check node-problem-detector posted KubeletStart event on node %q", node.Name))
+			gomega.Eventually(func() error {
+				return verifyEventExists(f, eventListOptions, "KubeletStart", node.Name)
+			}, pollTimeout, pollInterval).Should(gomega.Succeed())
 		}
 
 		ginkgo.By("Gather node-problem-detector cpu and memory stats")
@@ -202,6 +210,19 @@ func verifyEvents(f *framework.Framework, options metav1.ListOptions, num int, r
 	return nil
 }
 
+func verifyEventExists(f *framework.Framework, options metav1.ListOptions, reason, nodeName string) error {
+	events, err := f.ClientSet.CoreV1().Events(metav1.NamespaceDefault).List(options)
+	if err != nil {
+		return err
+	}
+	for _, event := range events.Items {
+		if event.Reason == reason && event.Source.Host == nodeName && event.Count > 0 {
+			return nil
+		}
+	}
+	return fmt.Errorf("Event %s does not exist: %v", reason, events.Items)
+}
+
 func verifyNodeCondition(f *framework.Framework, condition v1.NodeConditionType, status v1.ConditionStatus, reason, nodeName string) error {
 	node, err := f.ClientSet.CoreV1().Nodes().Get(nodeName, metav1.GetOptions{})
 	if err != nil {