Looks like that the replication is fast enough and I didn't see it... ;-)

The migration in my test lab was running without downtime. Let's see how it works with the productive system and with 485 GB of data on it.

Manuel