Merge pull request #416 from klihub/fixes/config/startup-error-propag…

…ation

resmgr,agent: propagate startup config error back to CR.

pkg/agent/agent.go

@@ -103,7 +103,7 @@ func TemplateConfigInterface() ConfigInterface {
 }
 
 // NotifyFn is a function to call when the effective configuration changes.
-type NotifyFn func(cfg interface{}) error
+type NotifyFn func(cfg interface{}) (bool, error)
 
 var (
 	// Our logger instance for the agent.
@@ -544,12 +544,17 @@ func (a *Agent) updateConfig(cfg metav1.Object) {
 		}
 	}
 
-	err := a.notifyFn(cfg)
+	fatal, err := a.notifyFn(cfg)
+	a.patchConfigStatus(a.currentCfg, cfg, err)
+
 	if err != nil {
-		log.Errorf("failed to apply configuration: %v", err)
+		if fatal {
+			log.Fatalf("failed to apply configuration: %v", err)
+		} else {
+			log.Errorf("failed to apply configuration: %v", err)
+		}
 	}
 
-	a.patchConfigStatus(a.currentCfg, cfg, err)
 	a.currentCfg = cfg
 }
 

pkg/resmgr/resource-manager.go

@@ -128,17 +128,17 @@ func (m *resmgr) Start() error {
 	return nil
 }
 
-func (m *resmgr) updateConfig(newCfg interface{}) error {
+func (m *resmgr) updateConfig(newCfg interface{}) (bool, error) {
 	if newCfg == nil {
-		return fmt.Errorf("can't run without effective configuration...")
+		return false, fmt.Errorf("can't run without effective configuration...")
 	}
 
 	cfg, ok := newCfg.(cfgapi.ResmgrConfig)
 	if !ok {
 		if !m.running {
-			log.Fatalf("got initial configuration of unexpected type %T", newCfg)
+			return true, fmt.Errorf("got initial configuration of unexpected type %T", newCfg)
 		} else {
-			return fmt.Errorf("got configuration of unexpected type %T", newCfg)
+			return false, fmt.Errorf("got configuration of unexpected type %T", newCfg)
 		}
 	}
 
@@ -151,17 +151,17 @@ func (m *resmgr) updateConfig(newCfg interface{}) error {
 		log.InfoBlock("  <initial config> ", "%s", dump)
 
 		if err := m.start(cfg); err != nil {
-			log.Fatalf("failed to start with initial configuration: %v", err)
+			return true, fmt.Errorf("failed to start with initial configuration: %v", err)
 		}
 
 		m.running = true
-		return nil
+		return false, nil
 	}
 
 	log.Infof("configuration update %s (generation %d):", meta.GetName(), meta.GetGeneration())
 	log.InfoBlock("  <updated config> ", "%s", dump)
 
-	return m.reconfigure(cfg)
+	return false, m.reconfigure(cfg)
 }
 
 // Start resource management once we acquired initial configuration.

test/e2e/policies.test-suite/topology-aware/n4c16/test12-config-status/code.var.sh

@@ -1,4 +1,13 @@
-helm-terminate
+CONFIG_GROUP="group.test"
+
+cleanup() {
+    vm-command "kubectl delete -n kube-system topologyawarepolicies.config.nri/default" || :
+    vm-command "kubectl delete -n kube-system topologyawarepolicies.config.nri/$CONFIG_GROUP" || :
+    vm-command "kubectl label nodes --all config.nri/group-" || :
+    helm-terminate || :
+}
+
+cleanup
 helm_config=$(instantiate helm-config.yaml) helm-launch topology-aware
 
 sleep 1
@@ -12,6 +21,7 @@ vm-command "kubectl wait -n kube-system topologyawarepolicies/default \
     error "expected initial Success status"
 }
 
+# verify propagation of errors back to source CR
 vm-put-file $(RESERVED_CPU=750x instantiate custom-config.yaml) broken-config.yaml
 vm-command "kubectl apply -f broken-config.yaml"
 
@@ -26,3 +36,19 @@ vm-command "kubectl wait -n kube-system topologyawarepolicies/default \
 }
 
 helm-terminate
+
+# verify propagation of initial configuration errors back to source CR
+vm-put-file $(CONFIG_NAME="$CONFIG_GROUP" RESERVED_CPU=750x instantiate custom-config.yaml) \
+            broken-group-config.yaml
+vm-command "kubectl apply -f broken-group-config.yaml" || \
+    error "failed to install broken group config"
+vm-command "kubectl label nodes --all config.nri/group=test" || \
+    error "failed to label nodes for group config"
+
+expect_error=1 launch_timeout=5s helm_config=$(instantiate helm-config.yaml) helm-launch topology-aware
+get-config-node-status-error topologyawarepolicies/$CONFIG_GROUP | \
+    grep "failed to parse" | grep -q 750x || {
+    error "expected initial error not found in status"
+}
+
+cleanup

test/e2e/policies.test-suite/topology-aware/n4c16/test12-config-status/custom-config.yaml.in

@@ -1,7 +1,7 @@
 apiVersion: config.nri/v1alpha1
 kind: TopologyAwarePolicy
 metadata:
-  name: default
+  name: ${CONFIG_NAME:-default}
   namespace: kube-system
 spec:
   pinCPU: true

test/e2e/run.sh

@@ -353,6 +353,7 @@ helm-launch() { # script API
     #         default: 20s
     #     cfgresource: config custom resource to wait for node status to change in
     #         default: balloonspolicies/default or topologyawarepolicies/default
+    #     expect_error: don't exit, expect availability error
     #
     # Example:
     #     helm_config=$(instantiate helm-config.yaml) helm-launch balloons
@@ -361,6 +362,7 @@ helm-launch() { # script API
     local helm_config="${helm_config:-$TEST_DIR/helm-config.yaml}"
     local ds_name="${daemonset_name:-}" ctr_name="${container_name:-nri-resource-policy-$1}"
     local helm_name="${helm_name:-test}" timeout="${launch_timeout:-20s}"
+    local expect_error="${expect_error:-0}"
     local plugin="$1" cfgresource=${cfgresource:-} cfgstatus
     local deadline
     shift
@@ -403,8 +405,15 @@ helm-launch() { # script API
 
     deadline=$(deadline-for-timeout $timeout)
     vm-command "kubectl wait -n kube-system ds/${ds_name} --timeout=$timeout \
-                    --for=jsonpath='{.status.numberAvailable}'=1" || \
-        error "Timeout while waiting daemonset/${ds_name} to be available"
+                    --for=jsonpath='{.status.numberAvailable}'=1"
+
+    if [ "$COMMAND_STATUS" != "0" ]; then
+        if [ "$expect_error" != "1" ]; then
+            error "Timeout while waiting daemonset/${ds_name} to be available"
+        else
+            return 1
+        fi
+    fi
 
     timeout=$(timeout-for-deadline $deadline)
     timeout=$timeout wait-config-node-status $cfgresource
@@ -413,7 +422,6 @@ helm-launch() { # script API
     if [ "$result" != "Success" ]; then
         reason=$(get-config-node-status-error $cfgresource)
         error "Plugin $plugin configuration failed: $reason"
-        return 1
     fi
 
     vm-start-log-collection -n kube-system ds/$ds_name -c $ctr_name
@@ -502,7 +510,7 @@ get-config-node-status-result() {
 get-config-node-status-error() {
     local resource="$1" node="$(get-hostname-for-vm)"
     vm-command-q "kubectl get -n kube-system $resource \
-                      -ojsonpath=\"{.status.nodes['$node'].error}\""
+                      -ojsonpath=\"{.status.nodes['$node'].errors}\""
 }
 
 wait-config-node-status() {