Either using cluster administrator to manually failover or having a cluster failure it's all the same. As the databases are brought online on the new node they will go through a recovery where completed transactions are rolled forward and incomplete transactions are rolled back.
No. Transaction state is "incomplete" or "uncommitted"
When the SQL Server Instance starts, it goes through recovery to redo completed/undo uncommitted transactions
All work is deferred to start up because the reasons for SQL Server shutdown can be many and varied. This isn't known at start up: so redo/undo always happens
Either using cluster administrator to manually failover or having a cluster failure it's all the same. As the databases are brought online on the new node they will go through a recovery where completed transactions are rolled forward and incomplete transactions are rolled back.
No. Transaction state is "incomplete" or "uncommitted"
When the SQL Server Instance starts, it goes through recovery to redo completed/undo uncommitted transactions
All work is deferred to start up because the reasons for SQL Server shutdown can be many and varied. This isn't known at start up: so redo/undo always happens