Merge pull request kubernetes-sigs#74 from lreciomelero/fix/Extract_m…

…hc_from_templates

extracted machinehealthchecks from templates

pkg/cluster/internal/create/actions/cluster/templates/aws.eks.tmpl

@@ -200,21 +200,3 @@ spec:
       {{- end }}
 {{- end }}
 {{- end }}
----
-apiVersion: cluster.x-k8s.io/v1beta1
-kind: MachineHealthCheck
-metadata:
-  name: "{{ $.Descriptor.ClusterID }}-worker-node-unhealthy"
-spec:
-  clusterName: "{{ $.Descriptor.ClusterID }}"
-  nodeStartupTimeout: 300s
-  selector:
-    matchLabels:
-      keos.stratio.com/machine-role: "{{ $.Descriptor.ClusterID }}-worker-node"
-  unhealthyConditions:
-    - type: Ready
-      status: Unknown
-      timeout: 60s
-    - type: Ready
-      status: 'False'
-      timeout: 60s

pkg/cluster/internal/create/actions/cluster/templates/gcp.tmpl

@@ -277,39 +277,3 @@ spec:
             {{- end }}
 {{- end }}
 {{- end }}
----
-apiVersion: cluster.x-k8s.io/v1beta1
-kind: MachineHealthCheck
-metadata:
-  name: "{{ $.Descriptor.ClusterID }}-controlplane-node-unhealthy"
-spec:
-  clusterName: "{{ $.Descriptor.ClusterID }}"
-  nodeStartupTimeout: 300s
-  selector:
-    matchLabels:
-      keos.stratio.com/machine-role: "{{ $.Descriptor.ClusterID }}-control-plane-node"
-  unhealthyConditions:
-    - type: Ready
-      status: Unknown
-      timeout: 60s
-    - type: Ready
-      status: 'False'
-      timeout: 60s
----
-apiVersion: cluster.x-k8s.io/v1beta1
-kind: MachineHealthCheck
-metadata:
-  name: "{{ $.Descriptor.ClusterID }}-worker-node-unhealthy"
-spec:
-  clusterName: "{{ $.Descriptor.ClusterID }}"
-  nodeStartupTimeout: 300s
-  selector:
-    matchLabels:
-      keos.stratio.com/machine-role: "{{ $.Descriptor.ClusterID }}-worker-node"
-  unhealthyConditions:
-    - type: Ready
-      status: Unknown
-      timeout: 60s
-    - type: Ready
-      status: 'False'
-      timeout: 60s

pkg/cluster/internal/create/actions/createworker/createworker.go

@@ -18,8 +18,8 @@ limitations under the License.
 package createworker
 
 import (
-	_ "embed"
 	"bytes"
+	_ "embed"
 	"os"
 	"strings"
 
@@ -58,9 +58,12 @@ type Secrets struct {
 
 //go:embed files/allow-all-egress_netpol.yaml
 var allowCommonEgressNetPol string
+
 // In common with keos installer
+//
 //go:embed files/deny-all-egress-imds_gnetpol.yaml
 var denyallEgressIMDSGNetPol string
+
 //go:embed files/allow-capa-egress-imds_gnetpol.yaml
 var allowCAPAEgressIMDSGNetPol string
 
@@ -317,6 +320,17 @@ func (a *action) Execute(ctx *actions.ActionContext) error {
 				return errors.Wrap(err, "failed to create the worker Cluster")
 			}
 		}
+		ctx.Status.End(true) // End Preparing nodes in workload cluster
+
+		ctx.Status.Start("Enabling workload cluster's self-healing 🏥")
+		defer ctx.Status.End(false)
+
+		err = enableSelfHealing(node, *descriptorFile, capiClustersNamespace)
+		if err := cmd.SetStdout(&raw).Run(); err != nil {
+			return errors.Wrap(err, "failed to enable workload cluster's self-healing")
+		}
+
+		ctx.Status.End(true) // End Enabling workload cluster's self-healing
 
 		ctx.Status.Start("Installing CAPx in workload cluster 🎖️")
 		defer ctx.Status.End(false)
@@ -407,7 +421,7 @@ func (a *action) Execute(ctx *actions.ActionContext) error {
 				}
 			}
 		}
-			
+
 		ctx.Status.End(true) // End Installing Network Policy Engine in workload cluster
 
 		if descriptorFile.DeployAutoscaler {

pkg/cluster/internal/create/actions/createworker/provider.go

@@ -17,11 +17,12 @@ limitations under the License.
 package createworker
 
 import (
+	"bytes"
 	_ "embed"
 
+	"sigs.k8s.io/kind/pkg/cluster/internal/create/actions/cluster"
 	"sigs.k8s.io/kind/pkg/cluster/nodes"
 	"sigs.k8s.io/kind/pkg/errors"
-	"sigs.k8s.io/kind/pkg/cluster/internal/create/actions/cluster"
 )
 
 const (
@@ -36,6 +37,9 @@ const (
 	CalicoTemplate  = "/kind/calico-helm-values.yaml"
 )
 
+const machineHealthCheckWorkerNodePath = "/kind/manifests/machinehealthcheckworkernode.yaml"
+const machineHealthCheckControlPlaneNodePath = "/kind/manifests/machinehealthcheckcontrolplane.yaml"
+
 type PBuilder interface {
 	setCapx(managed bool)
 	setCapxEnvVars(p ProviderParams)
@@ -174,3 +178,59 @@ func (p *Provider) installCAPXLocal(node nodes.Node) error {
 
 	return nil
 }
+
+func enableSelfHealing(node nodes.Node, descriptorFile cluster.DescriptorFile, namespace string) error {
+
+	if !descriptorFile.ControlPlane.Managed {
+
+		machineRole := "-control-plane-node"
+		generateMHCManifest(node, descriptorFile.ClusterID, namespace, machineHealthCheckControlPlaneNodePath, machineRole)
+
+		raw := bytes.Buffer{}
+		cmd := node.Command("kubectl", "-n", namespace, "apply", "-f", machineHealthCheckControlPlaneNodePath)
+		if err := cmd.SetStdout(&raw).Run(); err != nil {
+			return errors.Wrap(err, "failed to apply the MachineHealthCheck manifest")
+		}
+	}
+
+	machineRole := "-worker-node"
+	generateMHCManifest(node, descriptorFile.ClusterID, namespace, machineHealthCheckWorkerNodePath, machineRole)
+
+	raw := bytes.Buffer{}
+	cmd := node.Command("kubectl", "-n", namespace, "apply", "-f", machineHealthCheckWorkerNodePath)
+	if err := cmd.SetStdout(&raw).Run(); err != nil {
+		return errors.Wrap(err, "failed to apply the MachineHealthCheck manifest")
+	}
+
+	return nil
+}
+
+func generateMHCManifest(node nodes.Node, clusterID string, namespace string, manifestPath string, machineRole string) error {
+
+	var machineHealthCheck = `
+apiVersion: cluster.x-k8s.io/v1beta1
+kind: MachineHealthCheck
+metadata:
+  name: ` + clusterID + machineRole + `-unhealthy
+  namespace: cluster-` + clusterID + `
+spec:
+  clusterName: ` + clusterID + `
+  nodeStartupTimeout: 300s
+  selector:
+    matchLabels:
+      keos.stratio.com/machine-role: ` + clusterID + machineRole + `
+  unhealthyConditions:
+    - type: Ready
+      status: Unknown
+      timeout: 60s
+    - type: Ready
+      status: 'False'
+      timeout: 60s`
+
+	raw := bytes.Buffer{}
+	cmd := node.Command("sh", "-c", "echo \""+machineHealthCheck+"\" > "+manifestPath)
+	if err := cmd.SetStdout(&raw).Run(); err != nil {
+		return errors.Wrap(err, "failed to write the MachineHealthCheck manifest")
+	}
+	return nil
+}