Hello
Unless you have an edge case with very high IOPS and/or throughput requirements, I wouldn't worry too much about the random vs sequential nature. It's handled transparently by ONTAP and you don't really need to think about it like with disks attached to a server.
My opinion (I am not a DBA, nor a database expert) is that there are two important reasons for using more than one disk/LUN:
- Backup and recovery. Using SnapManager or SnapCenter it's easy to coordinate snapshots with database IO operations to protect the data. If logs and data are on the same disk/LUN, it's impossible to restore the data back to one point-in-time, then use the logs to bring the database up to a particular point-in-time.
- Specifically for LUNs, having more than one LUN means more than one queue, which is helpful in high IO situations.
Hope that helps.
Andrew