Redundant SCADA servers are used, if 100% process control and data safety is demanded, even when the server fails.
You achieve this fail safety by defining a second server, a so-called standby server, along with the project server. This standby automatically recognized a server failure Then it takes on the entire functionality of the server.
In order to avoid data loss in the time between the server failure and the recognition of the failure the standby always buffers all data. After a server failure this buffer is merged with the last data from the server and the new incoming data, so no data can be lost. So the control system guarantees seamless redundancy.
As soon as the server is online again and has copied the Runtime data, the standby automatically downgrades.
If the standby is running, when the server is started, the server copies all Runtime data from the standby. If you made any changes in the project data, while the server was offline and you have only updated them on the (not running) server, these changes will be overwritten, when the server copies the data from the standby.
In this case you have to update the standby before starting the server; or you close the standby when starting the server. As soon as it is restarted the standby then will get the new data from the (running) server.