This content has been marked as final. Show 3 replies
Why would you want to store the ORACLE_HOME in ASM?
Using ASM in this fashion means defining a file system in ASM that can be used via a file system driver (just like a NTFS or FAT32 file system). This is typically done in ASM on clusters for creating a shared cluster file system (called ACFS), where each server in the cluster can see the same file system and same files - and access and use the file system at the same time.
As for backups - if you use ASM then you do not need a file system. RMAN can backup directly to an ASM diskgroup.
There are also performance considerations. ASM automatically stripes a diskgroup's data across all disks in that group. If you define a diskgroup per physical drive, you will not gain any striping performance. If you have 2 physical disks in a diskgroup, there will now be 2 sets of disk controllers for reading and writing, thanks to striping. If there are 3 disks in the diskgroup, then 3 sets of disk controllers. This approach scales I/O performance.
ASM also support 2 and 3 way mirroring (or no mirroring if no redundancy in ASM need to be configured).
So if you expect to see improved I/O performance, and better and more flexible management of the storage layer, you need to look at what ASM provides and how this can be leveraged in your environment. Not look at how you can force your storage into ASM, to look exactly as the storage did prior to using ASM.
If your aim is to make the storage in ASM look exactly the same as your current layout - why then even bother using ASM?