Improve nodegroup deletion error handling

- fix missing return on nodegroup deletion error causing false-positive
- ensure describe stack is called with StackId whenever possible, to avoid
  error on stacks that happen to get deleted already

pkg/cfn/manager/api.go

@@ -140,6 +140,9 @@ func (c *StackCollection) describeStack(i *Stack) (*Stack, error) {
 	input := &cloudformation.DescribeStacksInput{
 		StackName: i.StackName,
 	}
+	if i.StackId != nil && *i.StackId != "" {
+		input.StackName = i.StackId
+	}
 	if i.StackId != nil {
 		input.StackName = i.StackId
 	}

pkg/ctl/delete/cluster.go

@@ -87,7 +87,7 @@ func doDeleteCluster(p *api.ProviderConfig, cfg *api.ClusterConfig, nameArg stri
 			for _, err := range errs {
 				logger.Critical("%s\n", err.Error())
 			}
-			handleIfError(fmt.Errorf("failed to delete nodegroup(s)"), "nodegroup(s)")
+			return fmt.Errorf("failed to delete nodegroup(s)")
 		}
 		logger.Debug("all nodegroups were deleted")
 	}