Skip to content

Centralized leaked ENI cleanup- refactor periodic cleanup & add node termination cleaner #505

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Mar 24, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions PROJECT
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@ resources:
version: v1beta1
- api:
crdVersion: v1
controller: true
domain: k8s.aws
group: vpcresources
kind: CNINode
Expand Down
2 changes: 2 additions & 0 deletions config/rbac/role.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -60,6 +60,8 @@ rules:
- create
- get
- list
- patch
- update
- watch
- apiGroups:
- vpcresources.k8s.aws
Expand Down
42 changes: 0 additions & 42 deletions controllers/core/node_controller.go
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,6 @@ package controllers

import (
"context"
"fmt"
"net/http"
"time"

Expand All @@ -25,8 +24,6 @@ import (
"github.com/aws/amazon-vpc-resource-controller-k8s/pkg/k8s"
"github.com/aws/amazon-vpc-resource-controller-k8s/pkg/node/manager"
"github.com/google/uuid"
"github.com/prometheus/client_golang/prometheus"
"sigs.k8s.io/controller-runtime/pkg/metrics"

"github.com/go-logr/logr"
corev1 "k8s.io/api/core/v1"
Expand All @@ -36,21 +33,9 @@ import (
ctrl "sigs.k8s.io/controller-runtime"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/controller-runtime/pkg/controller"
"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
"sigs.k8s.io/controller-runtime/pkg/healthz"
)

var (
leakedCNINodeResourceCount = prometheus.NewCounter(
prometheus.CounterOpts{
Name: "orphaned_cninode_objects",
Help: "The number of leaked cninode resources",
},
)

prometheusRegistered = false
)

// MaxNodeConcurrentReconciles is the number of go routines that can invoke
// Reconcile in parallel. Since Node Reconciler, performs local operation
// on cache only a single go routine should be sufficient. Using more than
Expand Down Expand Up @@ -93,23 +78,6 @@ func (r *NodeReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.

if nodeErr := r.Client.Get(ctx, req.NamespacedName, node); nodeErr != nil {
if errors.IsNotFound(nodeErr) {
// clean up CNINode finalizer
cniNode := &v1alpha1.CNINode{}
if cninodeErr := r.Client.Get(ctx, req.NamespacedName, cniNode); cninodeErr == nil {
if yes := controllerutil.ContainsFinalizer(cniNode, NodeTerminationFinalizer); yes {
updated := cniNode.DeepCopy()
if yes = controllerutil.RemoveFinalizer(updated, NodeTerminationFinalizer); yes {
if err := r.Client.Patch(ctx, updated, client.MergeFrom(cniNode)); err != nil {
return ctrl.Result{}, err
}
r.Log.Info("removed leaked CNINode resource's finalizer", "cninode", cniNode.Name)
}
leakedCNINodeResourceCount.Inc()
}
} else if !errors.IsNotFound(cninodeErr) {
return ctrl.Result{}, fmt.Errorf("failed getting CNINode %s from cached client, %w", cniNode.Name, cninodeErr)
}

// clean up local cached nodes
_, found := r.Manager.GetNode(req.Name)
if found {
Expand Down Expand Up @@ -148,8 +116,6 @@ func (r *NodeReconciler) SetupWithManager(mgr ctrl.Manager, maxConcurrentReconci
map[string]healthz.Checker{"health-node-controller": r.Check()},
)

prometheusRegister()

return ctrl.NewControllerManagedBy(mgr).
For(&corev1.Node{}).
WithOptions(controller.Options{MaxConcurrentReconciles: maxConcurrentReconciles}).
Expand Down Expand Up @@ -193,11 +159,3 @@ func (r *NodeReconciler) Check() healthz.Checker {
return err
}
}

func prometheusRegister() {
if !prometheusRegistered {
metrics.Registry.MustRegister(leakedCNINodeResourceCount)

prometheusRegistered = true
}
}
39 changes: 0 additions & 39 deletions controllers/core/node_controller_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -26,13 +26,11 @@ import (
"github.com/golang/mock/gomock"
"github.com/stretchr/testify/assert"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/errors"
v1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/runtime"
"k8s.io/apimachinery/pkg/types"
"sigs.k8s.io/controller-runtime/pkg/client"
fakeClient "sigs.k8s.io/controller-runtime/pkg/client/fake"
"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
"sigs.k8s.io/controller-runtime/pkg/log/zap"
"sigs.k8s.io/controller-runtime/pkg/reconcile"
)
Expand Down Expand Up @@ -143,43 +141,6 @@ func TestNodeReconciler_Reconcile_DeleteNonExistentNode(t *testing.T) {
assert.Equal(t, res, reconcile.Result{})
}

func TestNodeReconciler_Reconcile_DeleteNonExistentNodesCNINode(t *testing.T) {
ctrl := gomock.NewController(t)
defer ctrl.Finish()

mock := NewNodeMock(ctrl)
cniNode := &v1alpha1.CNINode{
ObjectMeta: v1.ObjectMeta{
Name: mockNodeName,
Finalizers: []string{NodeTerminationFinalizer},
},
}
mock.Reconciler.Client = fakeClient.NewClientBuilder().WithScheme(mock.Reconciler.Scheme).WithObjects(cniNode).Build()

mock.Conditions.EXPECT().GetPodDataStoreSyncStatus().Return(true)
mock.Manager.EXPECT().GetNode(mockNodeName).Return(mock.MockNode, false)

original := &v1alpha1.CNINode{}
err := mock.Reconciler.Client.Get(context.TODO(), types.NamespacedName{Name: cniNode.Name}, original)
assert.NoError(t, err)
assert.True(t, controllerutil.ContainsFinalizer(original, NodeTerminationFinalizer), "the CNINode has finalizer")

res, err := mock.Reconciler.Reconcile(context.TODO(), reconcileRequest)
assert.NoError(t, err)
assert.Equal(t, res, reconcile.Result{})

node := &corev1.Node{}
updated := &v1alpha1.CNINode{}
err = mock.Reconciler.Client.Get(context.TODO(), types.NamespacedName{Name: cniNode.Name}, node)
assert.Error(t, err, "the node shouldn't existing")
assert.True(t, errors.IsNotFound(err))

err = mock.Reconciler.Client.Get(context.TODO(), types.NamespacedName{Name: cniNode.Name}, updated)
assert.NoError(t, err)
assert.True(t, updated.Name == mockNodeName, "the CNINode should existing and waiting for finalizer removal")
assert.False(t, controllerutil.ContainsFinalizer(updated, NodeTerminationFinalizer), "CNINode finalizer should be removed when the node is gone")
}

func TestNodeReconciler_Reconcile_DeleteNonExistentUnmanagedNode(t *testing.T) {
ctrl := gomock.NewController(t)
defer ctrl.Finish()
Expand Down
Loading