I knew what the answer would be, so I got it "right". But "right" has to be in quotes, because it is actually wrong for application screnarios where disk utilisation is high - - the fact is that if decent disc performance is required then RAID 5 provides no protection against single disc failure, for two reasons: (i) during disc recovery the RAID's performance will not be adequate to support your workload, and (ii) disc recovery takes far longer than for RAID 10 and any disc failure during that recovery is a catastrophic failure (while for RAID 10 during recovery there is only 1 disc whose failure would be catastrophic). Every DBA who is involved in making storgae decisions should read and fully understand the material at http://www.baarf.com.
Or for a view from a RAID manufacturer who decided that RAID5 was no longer worth supporting see
That paper from Intransa was 6 years ago! And the way disc drive technology is changing - drive storage capacity going up much faster than disc throughput - the performance and security hits against raid 5 are getting bigger every year while the cost savings of raid 5 against raid 10 are getting less every year.
RAID 5 is OK if your database has a very low write rate and you can accept a disc read throughput reduction by a factor of 4 or more (depending on number of spindles in the set) during disc recovery; it's useless otherwise. It's good for holding read-only tables, provided the read rate is low enough. It's almost always bad news for transaction logs, for tables that are frequently updated, for tables that grow rapidly, and for applications very high disc throughput.