health policy attach/detach action should not be successful when send notification to health manager service with messaging.MessagingTimeout

[wangyu]

When cluster do attach/detach policy action, it will send the 'res' and 'reason' to this action.
https://github.com/openstack/senlin/blob/master/senlin/engine/actions/cluster_action.py#L1065
https://github.com/openstack/senlin/blob/master/senlin/engine/actions/cluster_action.py#L1085

For health policy，the 'res' and 'reason' for attach/detach action always be True, '', no matter whether health manager register or unregister execute successfully:
https://github.com/openstack/senlin/blob/master/senlin/policies/health_policy.py#L395
https://github.com/openstack/senlin/blob/master/senlin/policies/health_policy.py#L408

So do attach/detach from cluster still successful when send notification to health manager service with messaging.MessagingTimeout. This is very bad for the cluster health manager，the following things may happen:
1. When do attach from cluster when messaging.MessagingTimeout occurred，Register cluster to health registry will be failed, and then the cluster will not do health check, even though the policy has attached.
2. When do detach from cluster when messaging.MessagingTimeout occurred，Unregister cluster to health registry will be failed, and then the cluster will still do health check, even though the policy has detached.

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/senlin 12.0.0.0rc1 release candidate.