If that is the case there is no need to create journaling tasks pointing to all DAG member servers, the one task is sufficient (the journal mailbox when created as a target is linked to a task and does not move on failover).

Journaling should continue on failover of the DB. A small interruption of the journaling task could be observed if it is in the middle of processing a journal mailbox or message as it will receive an error from Exchange. This will result in a 1 minute pause of the task. After which the DB has successfully been mounted on another member of the DAG and the journal task will continue. The journal task really does not have any specific logic to continue (other than recreating MAPI sessions and re-opening the mailbox) as the Exchange Client Access Server (or array) and DAG handles the DB failover and makes sure MAPI traffic talks to the correct DAG member.