Q: Will the Redundancy feature work properly if one of the two input domains has points that the other input domain doesn't have? If that is the case, will it throw some kind of error message?
A: Redundancy will work. If there are points in input 1 that are not in input 2 then they will appear in the output when input 1 is active, and get marked as Not Connected when input 2 becomes active. If both input 1 and input 2 become active at some point in a session then the output will be the union of input 1 and input 2, with the points that only appear in the inactive input marked as Not Connected.
If you want to use redundancy, then you need to ensure that the two input domains contain points with identical names. If the point names are not the same from the two sources then you need to create a synthetic input domain and bridge all of the points from one of the source domains into it, changing the point names as part of the bridge.
Q: Writing to the output domain only updates the currently active redundancy connection. Why not both?
A: In general there is no expectation that both pathways are always available. The inactive source should be receiving updates to its points from the underlying process, not the DataHub instance.
If both input domains are available, writing to both of them would cause two write events to the data source, which can produce strange results. Consider a boolean value in a PLC that acts as a rising edge trigger. If the PLC responds to the write by resetting the boolean then writing to both connections will cause the trigger to run twice. This is considered undesirable behaviour.
This has special implications for OPC A&E. If you are reading from two different A&E servers then you have to accept that one of the two servers will not receive an acknowledgement from the A&E client. This is because it may not be connected, or that it is connected but not active. If the inactive A&E server becomes active, the A&E client may see some acknowledged events become unacknowledged as part of the fail-over and will need to re-acknowledge them. Both A&E servers should show the A&E conditions to be in the same state (aside from ACK) as they should be receiving their state information from the same underlying process.
Q: How can I get the data hierarchy in my input domains into my output domain?
A: A hierarchy in the output domain is created when an input domain becomes valid. There are two settings in the redundancy configuration:
If neither of these options are checked (the default) the data hierarchy gets copied to the output domain whenever the input domains switch. If neither input domain becomes valid then all points in the output domain will have bad quality and will be in a flat organization.
If the first option is checked then the hierarchy will not appear in the output domain.