Starting and Stopping
- Jason Andrew
Running
When the Slave System Monitors are started (by a service, script or manually) they will optionally wait for the ServiceScheduling database server to become available. They will then sit in an idle state until requested to manage a component on that node on behalf of the Master System Monitor.
When the Master System Monitor is started (by a service, script or manually) it will optionally wait for the ServiceScheduling database server to become available and for the Slave System Monitors to become available. It will then immediately start the ServiceOptimizer instance. Components that are configured to run on the same node as the Master System Monitor will be controlled by the Master System Monitor. Other components will be controlled by the Slave System Monitors.
If a node fails or the Slave System Monitor is not running, then the Master System Monitor will treat all the components on that node as unavailable. The ServiceOptimizer instance will however continue to operate.
Stopping
In order to shut down a distributed ServiceOptimizer instance only the Master System Monitor needs to be stopped (by a service, script or manually). All the components will be stopped except for the Slave System Monitors which can be left running until the instance is restarted.
If it is required to shut down all the Slave System Monitors as well, then these should only be shut down once the Master System Monitor has terminated.