feat: improve error transparency for volume attachment failure

Longhorn 9968

Signed-off-by: Derek Su <[email protected]>

api/model.go

@@ -73,7 +73,7 @@ type Volume struct {
 	Conditions       map[string]longhorn.Condition `json:"conditions"`
 	KubernetesStatus longhorn.KubernetesStatus     `json:"kubernetesStatus"`
 	CloneStatus      longhorn.VolumeCloneStatus    `json:"cloneStatus"`
-	Ready            bool                          `json:"ready"`
+	Ready            longhorn.Condition            `json:"ready"`
 
 	AccessMode    longhorn.AccessMode        `json:"accessMode"`
 	ShareEndpoint string                     `json:"shareEndpoint"`
@@ -1373,6 +1373,35 @@ func toEmptyResource() *Empty {
 	}
 }
 
+func toVolumeReadyCondition(v *longhorn.Volume) longhorn.Condition {
+	scheduledCondition := types.GetCondition(v.Status.Conditions, longhorn.VolumeConditionTypeScheduled)
+	isCloningDesired := types.IsDataFromVolume(v.Spec.DataSource)
+	isCloningCompleted := v.Status.CloneStatus.State == longhorn.VolumeCloneStateCompleted
+
+	status := longhorn.ConditionStatusFalse
+	message := ""
+	if v.Spec.NodeID == "" && v.Status.State != longhorn.VolumeStateDetached {
+		message = "volume is not attached to any node"
+	} else if v.Status.State == longhorn.VolumeStateDetached && scheduledCondition.Status != longhorn.ConditionStatusTrue {
+		message = fmt.Sprintf("volume is detached but not scheduled with reason %v", scheduledCondition.Reason)
+	} else if v.Status.Robustness == longhorn.VolumeRobustnessFaulted {
+		message = "volume is in faulted state"
+	} else if v.Status.RestoreRequired {
+		message = "volume requires restore"
+	} else if isCloningDesired && !isCloningCompleted {
+		message = "volume is cloning"
+	} else {
+		status = longhorn.ConditionStatusTrue
+	}
+
+	return longhorn.Condition{
+		Type:    longhorn.VolumeConditionTypeReady,
+		Status:  status,
+		Reason:  "",
+		Message: message,
+	}
+}
+
 func toSettingResource(setting *longhorn.Setting) *Setting {
 	definition, _ := types.GetSettingDefinition(types.SettingName(setting.Name))
 
@@ -1563,18 +1592,6 @@ func toVolumeResource(v *longhorn.Volume, ves []*longhorn.Engine, vrs []*longhor
 	//   3. It's faulted.
 	//   4. It's restore pending.
 	//   5. It's failed to clone
-	ready := true
-	scheduledCondition := types.GetCondition(v.Status.Conditions, longhorn.VolumeConditionTypeScheduled)
-	isCloningDesired := types.IsDataFromVolume(v.Spec.DataSource)
-	isCloningCompleted := v.Status.CloneStatus.State == longhorn.VolumeCloneStateCompleted
-	if (v.Spec.NodeID == "" && v.Status.State != longhorn.VolumeStateDetached) ||
-		(v.Status.State == longhorn.VolumeStateDetached && scheduledCondition.Status != longhorn.ConditionStatusTrue) ||
-		v.Status.Robustness == longhorn.VolumeRobustnessFaulted ||
-		v.Status.RestoreRequired ||
-		(isCloningDesired && !isCloningCompleted) {
-		ready = false
-	}
-
 	r := &Volume{
 		Resource: client.Resource{
 			Id:      v.Name,
@@ -1620,7 +1637,7 @@ func toVolumeResource(v *longhorn.Volume, ves []*longhorn.Engine, vrs []*longhor
 		ReplicaZoneSoftAntiAffinity: v.Spec.ReplicaZoneSoftAntiAffinity,
 		ReplicaDiskSoftAntiAffinity: v.Spec.ReplicaDiskSoftAntiAffinity,
 		DataEngine:                  v.Spec.DataEngine,
-		Ready:                       ready,
+		Ready:                       toVolumeReadyCondition(v),
 
 		AccessMode:    v.Spec.AccessMode,
 		ShareEndpoint: v.Status.ShareEndpoint,

client/generated_volume.go

@@ -65,7 +65,7 @@ type Volume struct {
 
 	PurgeStatus []PurgeStatus `json:"purgeStatus,omitempty" yaml:"purge_status,omitempty"`
 
-	Ready bool `json:"ready,omitempty" yaml:"ready,omitempty"`
+	Ready interface{} `json:"ready,omitempty" yaml:"ready,omitempty"`
 
 	RebuildStatus []RebuildStatus `json:"rebuildStatus,omitempty" yaml:"rebuild_status,omitempty"`
 

csi/controller_server.go

@@ -487,8 +487,9 @@ func (cs *ControllerServer) ControllerPublishVolume(ctx context.Context, req *cs
 	// TODO: JM should readiness be handled by the caller?
 	//  Most of the readiness conditions are covered by the attach, except auto attachment which requires changes to the design
 	//  should be handled by the processing of the api return codes
-	if !volume.Ready {
-		return nil, status.Errorf(codes.Aborted, "volume %s is not ready for workloads", volumeID)
+	ready := volume.Ready.(longhorn.Condition)
+	if ready.Status != longhorn.ConditionStatusTrue {
+		return nil, status.Errorf(codes.Aborted, "volume %s is not ready for workloads since %s", volumeID, ready.Message)
 	}
 
 	attachmentID := generateAttachmentID(volumeID, nodeID)

csi/node_server.go

@@ -163,8 +163,9 @@ func (ns *NodeServer) NodePublishVolume(ctx context.Context, req *csi.NodePublis
 		}
 	}
 
-	if !volume.Ready {
-		return nil, status.Errorf(codes.Aborted, "volume %s is not ready for workloads", volumeID)
+	ready := volume.Ready.(longhorn.Condition)
+	if ready.Status != longhorn.ConditionStatusTrue {
+		return nil, status.Errorf(codes.Aborted, "volume %s is not ready for workloads since %s", volumeID, ready.Message)
 	}
 
 	podsStatus := ns.collectWorkloadPodsStatus(volume, log)
@@ -454,8 +455,9 @@ func (ns *NodeServer) NodeStageVolume(ctx context.Context, req *csi.NodeStageVol
 		return nil, status.Errorf(codes.InvalidArgument, "volume %s hasn't been attached yet", volumeID)
 	}
 
-	if !volume.Ready {
-		return nil, status.Errorf(codes.Aborted, "volume %s is not ready for workloads", volumeID)
+	ready := volume.Ready.(longhorn.Condition)
+	if ready.Status != longhorn.ConditionStatusTrue {
+		return nil, status.Errorf(codes.Aborted, "volume %s is not ready for workloads since %s", volumeID, ready.Message)
 	}
 
 	if requiresSharedAccess(volume, volumeCapability) && !volume.Migratable {

k8s/pkg/apis/longhorn/v1beta2/volume.go

@@ -107,6 +107,7 @@ const (
 	VolumeConditionTypeRestore             = "Restore"
 	VolumeConditionTypeTooManySnapshots    = "TooManySnapshots"
 	VolumeConditionTypeWaitForBackingImage = "WaitForBackingImage"
+	VolumeConditionTypeReady               = "Ready"
 )
 
 const (