This information tells you what happens when you use the save-while-active function to eliminate your save-outage time.
The save-while-active function can eliminate your outage for particular save operations by not waiting for applications to end before starting the save procedure. However, you will have more complex and longer recovery procedures after restoring objects from the media.
You will have more complex recovery procedures because eliminating your save-outage time saves objects at different application boundaries. For save-while-active purposes, an application boundary is a point in time:
When you choose to eliminate your save-outage time, applications can update the objects you are saving before the objects reach a checkpoint. When this happens the server cannot determine if the images of those objects reached application boundaries when you restore those objects. Therefore at restore time, you need to define recovery procedures to bring those objects to a common application boundary. You will need these recovery procedures to bring the objects to a consistent state in relationship to each other. For this reason you should protect the objects you are saving with journaling or commitment control.
Furthermore, if you do not use commitment control, partial transactions can be saved without your knowledge. When you use commitment control, you can choose to have the save operation save all objects at transaction boundaries. However, if applications do not reach commitment boundaries within the specified time, the save operation will fail.
You should consider each of the following when you determine these recovery procedures: