drain: adding 'USE_EXTERNAL_DRAINER' provoding an option to enable/disable SRIOV OP drain controller, in favor of using maintenance OP to drive node drain aspects

Signed-off-by: Ido Heyvi <[email protected]>

Author: heyvister1

README.md

@@ -464,7 +464,7 @@ nodes in parallel from the pool the operator can drain in parallel. maxUnavailab
 
 > **NOTE**: If a node is not part of any pool it will have a default configuration of maxUnavailable 1
 
-> **NOTE**: Internal drain controller can be disabled by exposing the following `USE_MAINTENANCE_OPERATOR_DRAINER` env variable. This means that drain operations will be done externally, utilizing [NVIDIA maintenance OP](https://github.com/Mellanox/maintenance-operator). In addition, `SriovNetworkPoolConfig` will not take any effect during drain procedure, since the maintenance operator will be in charge of [parallel node operations](https://github.com/Mellanox/maintenance-operator/blob/main/api/v1alpha1/maintenanceoperatorconfig_types.go#L38-L46).
+> **NOTE**: Internal drain controller can be disabled by exposing the following `USE_EXTERNAL_DRAINER` env variable. This means that drain operations will be done externally, for example by utilizing [NVIDIA maintenance OP](https://github.com/Mellanox/maintenance-operator). In addition, `SriovNetworkPoolConfig` will not take any effect during drain procedure, since the maintenance operator will be in charge of [parallel node operations](https://github.com/Mellanox/maintenance-operator/blob/main/api/v1alpha1/maintenanceoperatorconfig_types.go#L38-L46).
 
 #### RDMA Mode Configuration
 

deployment/sriov-network-operator-chart/templates/operator.yaml

@@ -86,8 +86,8 @@ spec:
               value: {{ .Values.operator.metricsExporter.certificates.secretName }}
             - name: METRICS_EXPORTER_KUBE_RBAC_PROXY_IMAGE
               value: {{ .Values.images.metricsExporterKubeRbacProxy }}
-            - name: USE_MAINTENANCE_OPERATOR_DRAINER
-              value: {{ .Values.operator.maintenanceOperatorDrainer.enabled | quote }}
+            - name: USE_EXTERNAL_DRAINER
+              value: {{ .Values.operator.externalDrainer.enabled | quote }}
             {{- if .Values.operator.metricsExporter.prometheusOperator.enabled }}
             - name: METRICS_EXPORTER_PROMETHEUS_OPERATOR_ENABLED
               value: {{ .Values.operator.metricsExporter.prometheusOperator.enabled | quote}}

deployment/sriov-network-operator-chart/values.yaml

@@ -40,8 +40,8 @@ operator:
       serviceAccount: "prometheus-k8s"
       namespace: "monitoring"
       deployRules: false
-  # use external drain controller, utilizing NVIDIA maintenance operator
-  maintenanceOperatorDrainer:
+  # use external drain controller, for example utilizing NVIDIA maintenance operator
+  externalDrainer:
     enabled: false
   admissionControllers:
     enabled: false

doc/design/parallel-node-config.md

@@ -60,7 +60,7 @@ Node annotation, `sriovnetwork.openshift.io/state` and SriovNetworkNodeState ann
 
 *NOTE:* In the future we are going to drop the node annotation and only use the SriovNetworkNodeState
 
-*NOTE:* Internal drain controller can be disabled by exposing the following `USE_MAINTENANCE_OPERATOR_DRAINER` env variable. This means that drain operations will be done externally, utilizing [NVIDIA maintenance OP](https://github.com/Mellanox/maintenance-operator). In addition, `SriovNetworkPoolConfig` will not take any effect during drain procedure, since the maintenance operator will be in charge of parallel node operations.
+*NOTE:* Internal drain controller can be disabled by exposing the following `USE_EXTERNAL_DRAINER` env variable. This means that drain operations will be done externally, for example by utilizing [NVIDIA maintenance OP](https://github.com/Mellanox/maintenance-operator). In addition, `SriovNetworkPoolConfig` will not take any effect during drain procedure, since the maintenance operator will be in charge of parallel node operations.
 
 Draining procedure:
 

main.go

@@ -273,7 +273,7 @@ func main() {
 
 func setupDrainController(mgr ctrl.Manager, restConfig *rest.Config,
 	platformsHelper platforms.Interface, scheme *runtime.Scheme) error {
-	if controllers.UseMaintenanceOperatorDrainer() {
+	if vars.UseExternalDrainer {
 		setupLog.Info("internal drain controller is disabled, draining will be done externally by the maintenance operator")
 		return nil
 	}

pkg/vars/vars.go

@@ -95,6 +95,10 @@ var (
 
 	// FeatureGates interface to interact with feature gates
 	FeatureGate featuregate.FeatureGate
+
+	// UseExternalDrainer controls if SRIOV operator will use an external drainer
+	// for draining nodes or its internal drain controller (default)
+	UseExternalDrainer bool
 )
 
 func init() {
@@ -117,4 +121,6 @@ func init() {
 	ResourcePrefix = os.Getenv("RESOURCE_PREFIX")
 
 	FeatureGate = featuregate.New()
+
+	UseExternalDrainer = os.Getenv("USE_EXTERNAL_DRAINER") == "true"
 }