Why did we set phase Ready of modes that were not mapped?

[shirady]

Inside the reconciler of backingstore we have modeInfoMap.

noobaa-operator/pkg/backingstore/reconciler.go

Lines 48 to 67 in d821724

	func modeInfoMap() map[string]ModeInfo {
	return map[string]ModeInfo{
	"INITIALIZING": {nbv1.BackingStorePhaseCreating, corev1.EventTypeNormal},
	"DELETING": {nbv1.BackingStorePhaseReady, corev1.EventTypeNormal},
	"SCALING": {nbv1.BackingStorePhaseReady, corev1.EventTypeNormal},
	"MOST_NODES_ISSUES": {nbv1.BackingStorePhaseReady, corev1.EventTypeWarning},
	"MANY_NODES_ISSUES": {nbv1.BackingStorePhaseReady, corev1.EventTypeWarning},
	"MOST_STORAGE_ISSUES": {nbv1.BackingStorePhaseReady, corev1.EventTypeWarning},
	"MANY_STORAGE_ISSUES": {nbv1.BackingStorePhaseReady, corev1.EventTypeWarning},
	"MANY_NODES_OFFLINE": {nbv1.BackingStorePhaseReady, corev1.EventTypeWarning},
	"LOW_CAPACITY": {nbv1.BackingStorePhaseReady, corev1.EventTypeWarning},
	"OPTIMAL": {nbv1.BackingStorePhaseReady, corev1.EventTypeNormal},
	"HAS_NO_NODES": {nbv1.BackingStorePhaseRejected, corev1.EventTypeWarning},
	"ALL_NODES_OFFLINE": {nbv1.BackingStorePhaseRejected, corev1.EventTypeWarning},
	"NO_CAPACITY": {nbv1.BackingStorePhaseRejected, corev1.EventTypeWarning},
	"IO_ERRORS": {nbv1.BackingStorePhaseRejected, corev1.EventTypeWarning},
	"STORAGE_NOT_EXIST": {nbv1.BackingStorePhaseRejected, corev1.EventTypeWarning},
	"AUTH_FAILED": {nbv1.BackingStorePhaseRejected, corev1.EventTypeWarning},
	}
	}

inside the function completeReconcile we set the phase Ready to modes that were not mapped in modeInfoMap.

noobaa-operator/pkg/backingstore/reconciler.go

Lines 224 to 231 in d821724

	} else {
	// We will get here in case we do not have a mapping inside bsModeInfoMap
	// The default phase that was chosen was phase ready
	r.SetPhase(
	nbv1.BackingStorePhaseReady,
	"BackingStorePhaseReady",
	"noobaa operator completed reconcile - backing store is ready",
	)

The open questions are:

1. Why did we choose to set the phase as Ready?
2. Is there a reason that we will not map explicitly all the modes from core inside modeInfoMap?

[shirady]

We started the discussion in PR #1033 (exactly here), to separate the issues I opened this one.