kovaleff - Fotolia
RAID 1 vs RAID 5 is mostly a question of what is more important to you in terms of performance and cost.
RAID 1 is a mirrored pair of disk drives. When a write is carried out to the mirrored pair of drives, the data is written to both disks simultaneously. When a read is carried out, the data is taken from one of the two drives. A RAID 1 drive pair can therefore handle twice the amount of reads than a standalone disk drive, and is why RAID 1 has been the RAID level of choice for high-end applications.
Due to the larger data capacities required by modern environments, RAID 10 is now the most prevalent among the RAID technologies for high-end applications, combining the mirroring of RAID 1 with the striping of RAID 0.
The main drawback of RAID 1 (and, for that matter, RAID 10) is that for the required disk space, the RAID overhead will double the amount of capacity you actually have to buy. For example, if you needed to store 300GB of data, you would have to buy 600GB of capacity.
With RAID 5 – assuming again that you buy five disks – four-fifths of the capacity you buy is usable because the data is shared across four of the drives while one of the disks in the array acts as a parity disk that knows where all that data is located.
If a disk fails, the array can work out what is supposed to be on the volume by checking against the parity disk. In the past, there was an issue with speed, but disks today are pretty quick. Normally a RAID 5 array would not be considered for a high read/write environment, because every transaction must also write to the parity disk, which creates a bottleneck.
RAID 5 volumes do take a while to rebuild, so that is another sacrifice you have to make when you choose this type of RAID level. You can also put an OS on a RAID 5 array, but there will be a performance issue to contend with, which is simply not worth it. Again, dividing a RAID slice into multiple partitions could be possible, but I would not recommend it.