Originally Posted by
Stylish 
That observation is somewhat misleading.
Two drives can be setup in a RAID setting for mirroring (RAID1, I believe). All data is mirrored on both drives, so it would take 2x Xgb to store Xgb. (the values are not added together).
If there is a failure of one drive, it is possible to replace that drive and rebuild the data.
Mirroring (RAID1) allows a single disk failure without interruption and the capability of rebuilding the entire array.
Striping (RAID0) takes N drives but allows for zero failure. In this arrangement, the drives' capacities are added together (Nx Xgb = N(X)). The draw back is tolerance - a single drive dies and the entire array is down with data loss.
I am unsure how RAID arrays behave when the drives are not the same capacity.
You seem to be mirroring my post Stylish! Anyway "that drive" does not need to be replaced. I have already stated that the drives are
preferably the same size -as otherwise the capacity of the larger drive is reduced. Which would be wasteful if one drive was 60GB and the other was 120GB. But to simplify:
The proposal is that a couple of old drives can still be used with little risk if they are set up as RAID 0 as if a fault develops the software heals the error on one of the drives whilst the drives are still in use.
I am expecting someone to suggest whether that is realistic or not. And if not why not. disk enclosures aren't cheap and what manages them? Maybe an old computer can be used as extra storage -they are cheap enough. My two redundant hard drives one of 60GB and the other of 80GB are, apart from their age - circa 2001 still serviceable, so why chuck them in the recycling bin?