When an application enters maintenance mode (MM), it affects all other applications that share the same graph. For example, if two applications are linked by a controller, then putting one in maintenance mode will cause the other to go into maintenance mode as well; which one is the parent and which one is the child does not matter in this case.
Conversely, if two applications do not share the same graph, i.e., they are not linked by one or more controllers, then one can be put into MM while the other operates under normal RMS control.
For instance, in the example presented earlier, app1 and app2 are independent. While app1 is in MM, app2 continues to operate normally and can be switched from one node to the other, as illustrated in the following figure.
Figure 7.39 Normal operation of independent application
Even though some applications may continue to operate under normal RMS control, MM still places restrictions on the overall cluster operation. In particular, note the following:
You must exit MM before starting or stopping an application or a resource.
You must exit MM everywhere in the cluster before you can shut down RMS.
