Skip to content
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.

Sign up for GitHub

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

add initial grace period before triggering a failure due to missing components #273

Merged
merged 1 commit into from
Nov 27, 2024
Merged

add initial grace period before triggering a failure due to missing components #273

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
25 changes: 15 additions & 10 deletions internal/controller/appwrapper/appwrapper_controller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -268,21 +268,26 @@ func (r *AppWrapperReconciler) Reconcile(ctx context.Context, req ctrl.Request)
}

// Detect externally deleted components and transition to Failed with no GracePeriod or retry
detailMsg := fmt.Sprintf("Only found %v deployed components, but was expecting %v", compStatus.deployed, compStatus.expected)
if compStatus.deployed != compStatus.expected {
meta.SetStatusCondition(&aw.Status.Conditions, metav1.Condition{
Type: string(workloadv1beta2.Unhealthy),
Status: metav1.ConditionTrue,
Reason: "MissingComponent",
Message: detailMsg,
})
r.Recorder.Event(aw, v1.EventTypeNormal, string(workloadv1beta2.Unhealthy), "MissingComponent: "+detailMsg)
return ctrl.Result{}, r.transitionToPhase(ctx, orig, aw, workloadv1beta2.AppWrapperFailed)
// There may be a lag before created resources become visible in the cache; don't react too quickly.
whenDeployed := meta.FindStatusCondition(aw.Status.Conditions, string(workloadv1beta2.ResourcesDeployed)).LastTransitionTime
graceDuration := r.admissionGraceDuration(ctx, aw)
if time.Now().After(whenDeployed.Add(graceDuration)) {
detailMsg := fmt.Sprintf("Only found %v deployed components, but was expecting %v", compStatus.deployed, compStatus.expected)
meta.SetStatusCondition(&aw.Status.Conditions, metav1.Condition{
Type: string(workloadv1beta2.Unhealthy),
Status: metav1.ConditionTrue,
Reason: "MissingComponent",
Message: detailMsg,
})
r.Recorder.Event(aw, v1.EventTypeNormal, string(workloadv1beta2.Unhealthy), "MissingComponent: "+detailMsg)
return ctrl.Result{}, r.transitionToPhase(ctx, orig, aw, workloadv1beta2.AppWrapperFailed)
}
}

// If a component's controller has put it into a failed state, we do not need
// to allow a grace period. The situation will not self-correct.
detailMsg = fmt.Sprintf("Found %v failed components", compStatus.failed)
detailMsg := fmt.Sprintf("Found %v failed components", compStatus.failed)
if compStatus.failed > 0 {
meta.SetStatusCondition(&aw.Status.Conditions, metav1.Condition{
Type: string(workloadv1beta2.Unhealthy),
Expand Down
9 changes: 7 additions & 2 deletions test/e2e/appwrapper_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -297,8 +297,13 @@ var _ = Describe("AppWrapper E2E Test", func() {
Expect(aw.Status.Retries).Should(Equal(int32(2)))
})

It("Deleting a Running Component yields a failed AppWrapper", func() {
aw := createAppWrapper(ctx, pytorchjob(2, 500))
It("Deleting a Running Component yields a failed AppWrapper", Label("slow"), func() {
aw := toAppWrapper(pytorchjob(2, 500))
if aw.Annotations == nil {
aw.Annotations = make(map[string]string)
}
aw.Annotations[workloadv1beta2.AdmissionGracePeriodDurationAnnotation] = "5s"
Expect(getClient(ctx).Create(ctx, aw)).To(Succeed())
appwrappers = append(appwrappers, aw)
Eventually(AppWrapperPhase(ctx, aw), 60*time.Second).Should(Equal(workloadv1beta2.AppWrapperRunning))
aw = getAppWrapper(ctx, types.NamespacedName{Name: aw.Name, Namespace: aw.Namespace})
Expand Down