I am researching SQL Server 2016 and AlwaysOn.
One of the new features which sounds awesome on paper is "Database Health Level Detection". From my understanding if a database in a AlwaysOn Group enters a state other than
ONLINE than the databases would fail over to the secondary replica.
This got me pondering a few things and I can not find an answer yet and do not have our test lab setup yet either.
If I restore a database, the state of the database will be
RESTORING; will that trigger a failover event?
Also with AlwaysOn there is another feature called "Automatic Page Repair". If "Automatic Page Repair" fails, my understanding is that the database enters a Suspended state, which would make me expect the databases to failover as well.
So, is "Database Level Health Detection" only recommended if your cluster is only being used for one application? For example, one of our clusters is going to host about 50 different databases used by our Intranet; if one of those databases needs to be restored or has corruption, I would not want the other 49 to have a fail over event and a small outage, or would that not be an issue?