For Databases requiring normal redundancy I definitively recommend 11g instead of 10g, because the fast mirror resynchronization new feature that makes possible to reinstate a failed disk without requiring to rebuild it as is the case on 10g.
This Hands ON series move on a later lab the database to 11g to test this options. In this lab we still work with 10g.
Normal Redundancy on 10g provides protection against disk failures, but it does require for the failed disk to be reconfigured after failure.
On this lab we will simulate a disk failure and then we will restore the failed disk
Details on this document:
Normal Redundancy Resilience Test 10g
