If you use snapshots, be very careful to work with the SAN folks so the SAN doesn't run out of whatever space it uses to store the snapshot differentials, particularly on an active database.

A snapshot means the SAN has to keep track of the state at the point in time of the snapshot, plus the actual current state of the source data. On write capable snapshots, it also has to keep track of the "changes" written to the "snapshot". For what it's worth, it appears that SANs are more likely to keep track of the old data in the snapshot storage area, while virtual hypervisors are more likely to keep track of the new data in the snapshot storage area.