MachineDeployment stuck waiting for next resync if it can't find the infrastructure template

[ncdc]

What steps did you take and what happened:

1. Create a MachineDeployment
2. Create the infrastructure template for it
3. MachineDeployment has a ReconcileError event:

machinedeployment-controller  failed to retrieve AWSMachineTemplate external object "foo"/"test1-md-0": AWSMachineTemplate.infrastructure.cluster.x-k8s.io "test1-md-0" not found

1. It takes another 10-20 minutes (resync period) for the MachineDeployment to be processed and for the MachineSet to be created

What did you expect to happen:

1. It should requeue periodically if it gets a not-found error

Anything else you would like to add:
We get the error, log/event it, and return nil, so it doesn't requeue:

cluster-api/controllers/machinedeployment_controller.go

Lines 125 to 131 in a97903f

	result, err := r.reconcile(ctx, cluster, deployment)
	if err != nil {
	logger.Error(err, "Failed to reconcile MachineDeployment")
	r.recorder.Eventf(deployment, corev1.EventTypeWarning, "ReconcileError", "%v", err)
	}
	return result, nil

Environment:

• Cluster-api version: v0.3.2

/kind bug

[vincepri]

/milestone v0.3.3