DR doesn’t have any interaction with database clients, correct? That is, after a switch the clients connected to the old primary will just see the database being locked and the cluster file on all the clients doesn’t get updated automatically to point to the new cluster.
Yes that’s correct, client processes are not notified about the switch.
Whats the reason for this choice?
It feels like clients should get notified to switch their clusterfile to the new primary database after a fdbdr switch happens. What’s the point of not having this behaviour?
If you switched your replication direction, obviously you would want your clients to point to the new primary system so that they can continue writing to the system.