Cascading schedules are always recoverable!!!
If a transaction lets say T2 makes a dirty read from another transaction T1. Transaction T3 makes another dirty read from T2 and T4 makes a dirty read from T3, Now if T1 somehow crashes before commit and then rollbacks Then T2 will also rollback so does T3 and T4. This is just like cascading effect and so does the name Cascading schedule. But since Transactions T2, T3 and T4 depends on T1 and have not committed before T1 they can also be rolled back without causing any consistencies. And hence cascading schedules are recoverable.
The only drawback with cascading schedule is they lower the throughput of the systems as crashing of one transaction leads to subsequent rollbacks of other transactions.(Only the transactions which have performed dirty read)